K6JRF Home Pics


K6JRF's Page
formerly W6FZC

FM-DX Tuner and FM Yagi Design

Photos by K6JRF
All pages on this site are best viewed with 1024 x 768 or greater resolution
(Update: May 1, 2018)

Quick Select Menu

The Sony XDR-F1HD analysis and modification section has grown exponentially so I've added a quick select menu so you can navigate directly to the section that interests you.

- Stellar Labs Yagi. . . click here
- Sony DN1010 Rec. . . click here
- Sony XDR Overview. . . click here
5/1/18   - KUSC-FM Coverage. . . click here
- Heat Problems. . . click here
- Memory Mods. . . click here
- Forced Analog. . . click here
- Treble Rolloff. . . click here
- Treble Rolloff Cause. . . click here
- Total Harmonic Distortion. . . click here
- Total Harmonic Distortion in 'DB'. . . click here
- Negative FeedBack [NFB] Installation. . . click here
- Freq Response and THD. . . click here
- Spectrum Analysis. . . click here
- K6JRF External Preamp. . . click here
- K6JRF Yagi Design. . . click here
- Over-The-Air [OTA] Mods. . . click here

Click on any image to enlarge
FM Tuner and Stellar Labs FM Yagi
I've always enjoyed receiving programming direct esp FM classical music stereo programming. However, in San Diego, there aren't a lot of choices. Basically there is one (1) station that broadcasts analog monaural* classical music; XLNC1 from Tecate Mexico. Since it's about 40 miles, a well tuned yagi should do the job when put up at 30 ft or so.

However, if you have a modern FM tuner, then you discover that there are two (2) other FM stations that feature classical music. They are KPBS FM-2 and KUSC FM-1. Both of these stations use the newer HD format and KUSC uses both "analog" and HD format. So things are looking up!

I purchased a yagi from MCM Electronics made by Stellar Labs called "Four Element Directional Outdoor FM Antenna" for $20. Hard to go wrong at that price!

Check it out here. It features four (4) elements on a 66" boom. The driven element is a folded dipole with a built-in matching xformer, 300 to 75 ohms. Info states 7db gain with 15 db F/B ratio.

I wrote a review of the Stellar Labs antenna and it's posted here. I believe it's a great antenna that works as is but can be easily modified to other configurations as you will see here.

The small yagi below the Stellar Labs 6el yagi is my Radio Shack Corner Reflector antenna used for OTA reception. Check out the section at the bottom of the page here for the detailed design and how-to-modifiy details.

What I especially like about the SL 4el yagi is the driven element built-in matching network. It matches 300 feed impedance to 75 ohm via a internal xformer attached to the feed point of the driven element. It's sealed to prevent water entry which means that it will last a long time in So Cal weather. Most antennas require an external balun as an add-on which raises concerns about the balun's quality plus how well it will weather in So Cal.

Recommend that you use ferrite chokes at the feed point as shown in the picture. There are three (3) snap-on ferrite chokes to ensure that there is minimal, if any, outside shield current to modify/change the yagi's pattern. The impedance of the three (3) chokes is apx 930 ohms at 100Mhz.

Good selection of snap-on ferrite chokes is available at Palomar Engineers here. Their "FSB31-1/4" is reasonably priced @ $2.75 each and have 310 ohms impedance @ 100Mhz for a single turn.

* June 2016: XLNC suddenly stopped broadcasting in STEREO, going to MONO operation. About 2 weeks after continued MONO broadcasting I called the USA number and talked to a station representative who said ". . . I wasn't aware that we are not broadcasting in STEREO". They took my email address and said that I'd receive an reply from the station "techs". Nothing was received. About 1 month later, I inquired again via their web site's direct contact form but again no reply. About 3 weeks ago the music programming went silent (full carrier but no "modulation" for apx 5 hours) until somebody realized that there was no programming. It went down for apx 1 hour to return to STEREO broadcasting. About 24 hours later, STEREO ceased and replaced with MONO operation. Today, XLNC1 is still in MONO. (3/1/18)Last week they announced that FM OTA broadcasts will cease on Mar 1, 2018.
Antenna Optimizer
I used Antenna Optimzer [AO] to analyze the yagi and then optimize it for XLNC1 @ 104.9Mhz. The 'stock' design showed poor performance above 95Mhz to the top of the band at 108Mhz. Check the design section below for details.

The "beam" heading from my house is 121 degs and the picture shows the modified yagi pointing in that direction. Unfortunately, Mount Miguel is directly in front of XLNC1's broadcast tower so it's partially "shaded" at my house. Usually that means if you swing the yagi around, you MIGHT find a spot where the signal bounces off a tall bldg or another hill/mountain. I didn't do that. The signal is adequate when beamed directly at Tecate, MX.
Antenna Mount
The pictures shows the yagi-pair mounted in-use.

Many years ago, I had a 10.5 ft Radex dish that was mounted on the top of the 14 ft pole. I kept the pole when the dish was sold so other antennas could be mounted on it. It has proved its worth over the years. The top antenna is close to 30 ft over ground.

The left picture shows the XLNC redesigned FM yagi mounted on 15 ft steel pole along with the modified Radio Shack UHF corner-reflector underneath at apx 2.5 ft below. I use it to receive all UHF/VHF OTA signals. Check out the Yagi Design section at the bottom of this page for details.

This first picture on the right shows the XLNC yagi-pair from the front of my home beamed toward Tecate, MX (121degs). The last picture shows the KUSC 6el yagi-pair beamed at KUSC xmitter (335 degs). Pretty unobtrusive IMO but I'm very biased!
Sony STR-DN1010 Receiver/Amplifier
My home audio system consists mainly of "one" component, a Sony STR-DN1010 7.1-Channel A/V Receiver. The details can be seen here.

The info states that you can "enjoy 3D images while hearing sound closer to the way it was captured, plus improve image quality of non-HD video sources to HD using one convenient cable. Offering plenty of connectivity options for your expanding lifestyle, this A/V receiver features 4 HDMI inputs, 3 component inputs and is SIRIUS Satellite Radio-ready. You can also share your music in multiple rooms with wireless 2nd zone technology".

I do enjoy its overall performance but when my "FM-quest" started, I thought (wrongfully) that the FM tuner in the STR-DN1010 would perform OK! I was very wrong . . . but it took some time to find out.

Listening to the FM analog stereo @ 104.9 Mhz showed distortion, poor channel separation and overall "noises". At one point it appeared that it was signal overload, so adding some attenuation should improve the overall performance. After adding 40db in line to the 75 ohm input connector, the sound would, at times, be almost "perfect"! The reason was that it would drop stereo reception and switch to "monaural" operation.

Then conditions would change and the received FM audio quality would again be extremely poor. Changing the attenuation to 20db, the receiver noise would increase indicating some kind of rcvr desense. I started to realize that it basically was mass produced FM tuner with IF selectivity problems.

To test this theory, move off the main frequency on either side of center frequency. If the IF selectivity is adequate, the channel should go away. For example, 104.9 was changed to 104.8 but the channel was clear, now in MONO. Even at 200khz away, the channel was still somewhat audible . . . and that shouldn't be. In other words, the FM tuner in the STR - DN1010 is crap!

There are some interesting sites that discuss the various tuner problems with references to knowledgeable sites to fix/improve the tuner's performance. Check out "The Tuner Information Center" here. Take the time to read the complete discussion and you'll be much more knowledgeable about FM tuners.

I'm sure that my STR-DN1010 FM tuner could use a gang filter section to improve the awful IF selectivity. That's a task for another day.
Sony XDR-F1HD FM Tuner
So how about a "new" tuner?? While FM tuner shopping, I read about the Sony XDR FM tuner that can receive analog AM and FM, as well as digital AM and FM stations that have adapted HD Radio technology. So maybe I can still find one??

Many have abandoned AM and FM in favor of Internet and satellite radio but the sound quality doesn't compare to clean FM signal. [I've recently made some spectrum analysis plots of KUSC FM HD1, KPBS HD1 and HD2. The spectrum is full to beyond 15Khz, so my initial observation wasn't correct!]

However, getting a clean OTA signal can be problematic. FM is too often plagued with static, noise, and distortion and that's where the XDR-F1HD HD Radio is so impressive. If you have a good directional yagi, it's mostly immune to those forms of interference. The Sony tuner pulls in hard-to-receive stations better than any high-end tuner that experts have ever used. Some audiophiles believe the XDR-F1HD may be the best tuner available, regardless of cost, and many agree.

I've always wanted a Sony XDR-F1HD tuner but today, it is hard to find. But after some searching, I found one that was in stock condition, so I bought it. At the same time, after reading the fame of this tuner, I consulted K6STI's site here for information on the Sony XDR tuner. He has tested it in detail although some of it doesn't check-out with my findings. You should read the info contained there and follow the recommendations except for the treble rolloff 'cap' mod.

Another prominent Sony XDR tech is Dave known as "XDRguy". His site here details the many aspects of his repair and upgrade of the Sony XDR tuner.

The left picture shows a completed, upgraded Sony XDR tuner with a custom audio preamp (lower noise/less loss), upgraded power supply with new caps, two (2) added cooling fans, Memory Backup mod, FA/HD switch w/ LED and gold plated L + R output connectors. Unit is also realigned for improved S/N ratio ensuring great reception for a low level RF input signal.

I especially like his bench setup including a Sony XDR Tuner and Sangean HDT-1X. The HP8903 Audio Analyzer and the Panasonic VP-8191A FM Signal generator are key instruments to properly align and test a Sony XDR tuner.

My Sony XDR Arrives . .
After unboxing it, visually inspecting it (appeared to be perfect!), I connected the 75 ohm coax as well as the L - R audio output cable to the STR DN1010 amplifier. After setting the clock and other housekeeping task, a "SCAN" was performed.

To say that the sound quality is amazing is an huge understatement! The sound (even with its deficiencies) is absolutely amazing! And the channel separation places you in the center of the concert hall. The sound field settings on the STR DN1010 are quite good and turn the line-level L - R tuner output into room filling stereo sound. Needless to say I spent a lot of time listening and tuning for FM stations, both analog and digital.

Since I've had the SonyXDR tuner, I've NEVER heard any 'noise' (background, hiss, etc) when listening to an FM HD) or analog stereo broadcast. Well today, I did!

We had a band wipeout similar to the satellite summer and winter equinox; the signal started to drop . . . from 3 bars and solid HD) to HD) gone away, then 2 bars, then 1 bar and finally "0" bars! At that instant, a flute passage was airing and with it some 'noise'!

I'm sure that the DSP was having a hard time distinguishing between the flute and 'noise'. It wasn't much but there was some!

This tuner has been around since 2009 and it was an instant hit back then. Click here to read a 2009 article on the tuner.
KUSC FM @ 91.5Mhz Coverage
One of the best So Cal FM stations is KUSC @ 91.5Mhz from Los Angeles. During a "SCAN", I found that it was received at my house with typically "one" bar (out of three) indicating that I MAY be able to get a decent signal with my yagi turned towards Los Angeles!

Today, I beamed the antenna toward KUSC xmitter site [34 12 48 N; 118 03 41 W] to see the signal level possible. Maybe three (3) bars is in my future??

The beam heading from my house [32 49 6 N; 117 10 27 W] to KUSC is 332 degs. The distance and heading can be calculated from a program here. It calculates that my house to the KUSC transmitter is 109 miles.

Looking at the chart it clearly shows, I'm in the far-fringe area. The good news is the station employs 39KW and the antenna is at 2923 ft!  Click here for KUSC FM transmitter data.

[Email from KB7NIE - tried to reply to email address but it was denied!]
Question about your Stellar Labs Yagi setup for broadcast FM reception: Have you compared reception using vertical vs horizontal polarization? Broadcast FM transmitters are vertically polarized, while your receive antenna is horizontal. I didn't see mention of one versus the other in your website. Thanks! 73,Henry

Not sure where you got the idea that commercial FM stations broadcast with vertical polarization only. Certainly not b/c the tower is "vertical" doesn't mean that its signal is vertically polarized.

If you examine the FCC license info for commercial FM stations, you will see that they require the same (identical) ERP for both horizontal and vertical polarizations. As a result many FM stations use circular polarized [CP] antennas. They are lossy and expensive but they are regurlarly used to satisfy the FCC license requirement. Shown in the left picture is from ERI; 1180 series CP FM broadcast antenna capable of 140KW power broadcasting an omnidirectional signal! In the right picture: PCP FM Broadcast Antenna- 4 bay w/ 3dB gain broadcasting a directional signal.

Click the link here to read a 'treatise' on the use of CP vs VP. Towards the end of the article, the author states:
"Once CP is at the same ERP as VP, you should have some improved reception to vehicles that may be experiencing multi-path distortion. In urban areas with an increased possibility of reflections, you may notice CP has less momentary signal fluctuations as you drive through these reflections. They are typically very brief, lasting a fraction of a second and do not cause the listener to change channels. In these cases CP may provide a signal that is a bit more consistent but not with increased range or distance."
So the author concludes that CP actually performs better than VP independent of all REQUIREMENTS (FCC, legal, etc). CP, despite its high cost, actually provides a better signal to all users including those in vehicles.

The left side picture shows that the orientation of the SL 4el yagi is NOT 335 degs!! The reason is obvious when you look at the pattern. It's tuned for XLNC @ 104.9Mhz, so its F/G and F/B ratio @ 91.5Mhz is "awful"! The F/G is 4.8db and F/B = 8db so it has more gain off the back side than the front! So when re-designed for KUSC @ 91.5Mhz, it should give me reliable year round reception. See the yagi section below for the KUSC results.

Wow ... three bars and the HD)) symbol!
[Unfortunately the HD) blinks/goes off and
the 3 bar level becomes as low as 1 bar!]

As an aside, the received FM audio produced by the Sony XDR tuner is outstanding after the ckt mods (described below) are made. There are differences in audio quality the three (3) FM stations. Ranking the three (3) stations . . .
KUSC - - - #1 by a small margin using HD1 - no compression*, great clarity and high presence!
KPBS - - - #2 almost equal to #1 using HD2 - lower presence and some compression* due to HD2 bw restrictions.
XLNC - - - #3 lowest of the three; only analog monaural FM. High THD at times and low S/N ratio; can hear noise about -50db down **

      *   = obviously, some recordings MUST be compressed.
      ** = Mar 1, 2018, they will cease FM OTA broadcasting and use Internet only.
Sony Tuner Heat Problems
After using the tuner all evening, the case temperature is a bit alarming! It measured 128F. Resting your hand on top was a bit uncomfortable. The picture shows the detailed temperature measurements made with an infrared sensor. As many know, extended temperature will shorten the life of the components, esp the capacitors even though the PS Board uses 65C hi-temp capacitors. The internal temperature is higher than I would like it to be so something has to been done.

The 10.5 volt produced on the PS Board, is half-wave rectified. Some have added another diode to make it a full-wave rectified voltage. Adding the diode will slightly RAISE the internal temperature b/c of the higher average current value of the rectified voltage produced by the added rectifier. It's not needed since the 10.5V is again "regulated" to 8.5V (in IC902) with added filtering provided by R302, C302 (LPF to the Audio Preamp ckt) and C10 (to the DSP Tuner Unit).

Cooling Fan Added
I removed the cover and added a fan as shown in the picture. I usually use a glue-gun but this time I used a double-sided tape strip since the fan's weight is not a problem for the tape's adhesive strength. To ensure that it won't droop due the high temperature in the case, I added two (2) support posts super-glued on each side of the fan. They are visible as "white-spots" in the picture and rest on top of each capacitor under the fan.

My rationale for the site chosen is that it appears to be optimum b/c hot air is drawn from the transformer, power supply components from the bottom of the case and then forced out the back. The fan is an "Evercool", 12Vdc fan with apx 1W power draw.

I soldered it to the 5.2V tap so it would run at low speed and be noiseless. When the XDR tuner is "OFF" the voltage rises to 6.8 V and moves even more hot air from the inside. Did not "gate" it with a control signal b/c it should run continuously to cool the cabinet even when the main power is "OFF". When measured after sitting for 8 hours, the top cabinet temperature was 78F with the room temp at 75F. Since it can't be heard, why not take advantage of as much cooling as possible.

This fan placement is not without a consequence; the case inside "lip" must be cut as shown in the picture and the center post must be removed. This is not a problem since the four screws adequately/securely hold the case together.

In order to raise the unit and not make it look too obvious, I added felt pads, apx 3/16" thick to raise the bottom above the cabinet top so that there is less restriction on input air flow.
Heat Mods Results
With the cover installed and running, the picture shows the measured temperatures on a 90F outside temp day. Inside the temp was 79F with the cabinet top at 83.4F. This is mainly due to the STR DN1010 amplifier power dissipation as heat BION.

The XDR cabinet top was reduced to 101.5 F, a decrease of apx 25+F degs. The air exit temp is 99.5F showing that the fan IS doing its job. All things considered, a worthwhile and necessary modification.

The weather has cooled (66F outside) and the SonyXDR tuner has a new home (on a shelf), so I took temperature measurements to see what it is now with a 70F house temperature. Very nice!
Memory Backup Mod
As fine as the unit is, there is no memory backup ckts in the Sony XDR tuner, BION. You'd think that it would have been a major design function. But it wasn't. So the user looses all stored parameters (clock, LCD display and all memory presets) are wiped clean when main power (120Vac) is lost. Another serious deficiency that must be fixed.

There are two (2) main ways to provide memory backup; use a "large" cap such as a 1.5 Farad or a battery. Of the two, I prefer the battery method since it can potentially last at least 2-3 years or more! So I chose the battery method.

I have a LIR 2450 3.6V LiOn button battery with leads, so it's the natural choice. Also using the battery method, you don't have to remove the main PCB, all mods can be made from the top side.

The LIR 2450 specs are here. Max current draw is 120ma so there's a lot of headroom. In charging the battery, you need to limit the charge voltage to 4.2V max. Charging from a 4.7V ckt, could present problems for the battery over a long time.

After making measurements with the Sony tuner operating, the voltage can NOT exceed 4.2V b/c of the D912 forward voltage drop. But the voltage comes very close, so I studied the ckt awhile and the fix was obvious.

Resistor Method
The measured D912 forward voltage drop is only 0.6V (4.7 - 4.1 = 0.6V) so that means there is very little current flowing through it. The solution is obvious . . cause more current flow through D912 in order to raise the FV drop to 0.7V or 0.8V by adding a resistor to ground. That will lower the max charging voltage to 4.1V. Perfect!

Rephrasing the above; currently, the FV drop through D912 is apx 0.6V making the terminal charging voltage of 4.1V. Adding the resistor (Rbc = 27K) causes the FV drop to increase to 0.74 reducing the terminal charge voltage to 4.0V. You can increase Rbc value but more current is 'wasted' (in the OFF state, the battery must supply this current) further lowering the available charging voltage.

Note that D914 is not effected b/c the current flow to the Memory is low (90ua) so it's FV drop is still 0.6V.

The right side chart shows what happens by using a Spice transient analysis where the 4.7V regulator is turned on for 20sec and then off for 60sec to 100sec total. The analysis shows the battery at 4.0V which is close to the max voltage point, 4.02V before draining some charge forcing the battery to recharge.

Using Spice, a resistor of apx 27K ohms will cause enough extra current to flow that raises the D912 FV drop to 0.74 volts. The added resistor (Rbc) does the job and ensures that there will never be excess voltage applied to the Li-ion backup battery.

What happens is this; as the battery slowly charges, it reaches a point where the battery's charge voltage equals the available charging voltage. At that point the ckt reaches equilibrium; ie, the battery charging current is "zero". So now the battery will start to slowly discharge. As that happens, the battery voltage drops to a point where now it can be charged. The equilibrium point is apx 4.025 Volts of battery charge voltage.

The chart on the right shows the a plot of Li-ion battery (B1) voltage plotted against the charging current (TPi2). At 4.02V, the charging current is barely positive (10ua). At 4.03V, the battery charging current is negative meaning that the battery is reached its max terminal voltage and is now discharging. Thus the battery can't charge above 4.02V. Neat solution IMO.

Without the 27k Rbc resistor, my measurements showed 4.19V @ D912 cathode, tuner ON. With the tuner OFF, this results in apx 35ua current draw with battery = 4.17V. At that current discharge rate it will provide memory backup for a very long time! With the Rbc (27K) resistor added, the values change as shown in the above Spice chart.

The picture shows the details with the cooling fan removed for clarity. Solder at the points shown; JW20 for "ground" and D912 cathode for the 220 ohm resistor going to the LIR 2450 "+" side. The Rbc resistor is soldered to the junction of D912 cathode as shown. This is a very simple and straightforward method which I believe is the best method. So far, I've tested the backup system for three (3) months with no memory loss. Even the clock time setting is preserved.

Zener Diode Method
Another way to accomplish the 4.0V "clamp" action on the Li-ion battery is to use a zener diode. With todays technology, almost all voltage variations are available. One such zener from NXP Semiconductors is their 4.1V zener with 2% tolerance. With that tight tolerance, it's possible to use it b/c UPPER limit = 4.11V, the LOWER = 4.02V. Mouser PN: 771-NZX4V3A-133. Perfect for this application.

There isn't a Spice model for the NXP zener diode so I used a 3.9V, 5% zener's characteristics to get the same UPPPER voltage limit of "4.1V". Using that zener in the Spice analysis shows the Li-ion battery at 4.08V, just before zener action at 4.1V. Inspection shows the same action as in the "resistor" version except there is more current flow b/c of the low zener impedance of 100 ohms.

During my tests, the Li-ion battery started at 4.17V at the time that the zener was installed. The battery steadily discharged at apx 35ma as the terminal voltage steadily decreased to 4.08V. This zener has a 100 ohm zener impedance which is consistent with the observed current flow.

So this concept works nicely but it's a bit "tougher" on the battery. Also not EVERY zener has the desirable characteristics of the NXP zener. However, this shows the versatility that can be employed in arriving at a solution.
Analog Forced Reception Mod
Today was a good example of the need for Forced Analog [FA] reception. As the weather got bad and started to rain, reception from KUSC @ 91.5Mhz degraded so that the HD) icon was stable for 2 minutes, then would start to flash. When that happens the Sony XDR tuner switches from digital to analog and there are definitely artifacts while switching back and forth!

Pic by Bob Meister, WA1MIK

To say it's not pleasant listening is an understatement. So time for the FA Mod. Many have written about it and perform the required PWB cut, switch addition and wiring interconnection. If you are not proficient at making cuts to 5mil traces that are 2mils apart, then maybe somebody who is should do it for you.

I've indicated where I made the cut to the PWB b/c it's closest to pin 23 and the easiest spot to make a cut w/o having an "accident" and cutting some adjacent trace.

Solder two wires as shown and bring them up to the top side of the main board by snaking them up one of the many holes on the board. I drilled a 1/4" hole on the back panel as shown in the picture to mount the switch. It is orientated up for "HD" and down for "FA".

If you have a DPDT switch, then some have added a blue LED so that the user can see if it's in the FA (or HD mode, your choice) mode. That's all there is to it!

Treble Audio Rolloff Discussion
Testing has reported that the Sony XDR FM Tuner experiences treble rolloff from the PreAmp Ckt and is anywhere from -1.5 to -2.5dB @ 15Khz. The chart shows a typical test of a Sony XDR tuner showing -2.3dB loss at 15Khz.

Every Sony tuner is somewhat different but measurements show that they all seem to fall into this defined attenuation range. Before actually digging into this problem, I simply removed the four (4) caps no matter what treble loss they contributed. In hindsight, I'm not sure I heard that amount of treble rolloff on my Sony tuner. Tuners with circuit failures certainly would exhibit more treble rolloff than a 'good' one.

I employed a K6STI chart that showed the treble rolloff of his Sony XDR tuner of -2.3dB @ 15Khz. After receiving a email from him, I decided to use my own charts. Up to this time, I hadn't really "checked" my Sony tuner for its loss . . . just assumed that mine would be the same or less. I did NOT verify the results myself. There's a lesson here, don't you think??

5Spice preamp ckt simulation showed -0.56db @ 15Khz and -0.89db @ 20Khz. Since Spice is never wrong *, the remaining rolloff is due to the DSP Tuner module. In other words, the Preamp contributes -0.6dB @ 15Khz and the DSP Tuner + preamp totals to -2.2db rolloff.

XDRguy's site shows a "Pete Millett" unmodified Sony XDR rolloff measurement. The sweep results of the PreAmp ckt show -1.5dB @ 15Khz attenuation. So there are some differences evident and we still are not certain "why" there are differences. He also has reported 0.5dB rolloff variations @ 15Khz with different Sony XDR tuners that he has tested.

XDRguy measurement of the Sony XDR Treble Rolloff [from "RF In" to "L+R output"] shows -2.3dB loss @ 15Khz and claims that it's from the PreAmp ckt, not the DSP Tuner module portion of the Sony XDR. The chart at the beginning of this section is taken from that data. However, I believe that the loss is not all from the Preamp ckt itself.
* Certainly the results are only as good as the model. If it's incomplete, so are the results! When the correct transistor model [see the THD Analysis section below] was located and incorporated into the simulations, the results dramtically changed. Now DC bias levels, THD and input drive levels to produces clipping could be accurately determined. Also, I suspected that the characterization of DSP Tuner's output impedance wasn't correct. After some simulations, the correct value was found and has been used in ALL simulations. The simulation results now shadow [see the Frequency Response [FR] & Total Harmonic Distortion [THD] section below] real world performance.

What's causing the difference
So what is causing the difference?? Doing many Spice simulations, I found that 3300pf causes greater than 2dB loss @ 15Khz. The Sony XDR ckt board is two sided with traces on the bottom side and "wire-jumpers" used to connect crossover points. They can add 3-4pf / jumper if the trace connection is over "ground". But it would take a lot of jumpers to make 3300pf! Not a feasible answer to the problem.

Another idea was a "missing" component (a component that is on the board but doesn't appear in the PDF schematic) but none was found.

Another explanation was the values of C12 and C22 are larger than stated. One clue to this idea, is that removing C12 and C22, the attenuation drops to fractions of a "db". If that is true, that would throw out two arguments; it could be layout sensitive and there's a "hidden" component. However, the theory was not supported when measurements of the removed caps, C21, C22, C11 and C21 show values close to nominal.

One final argument concerns the output impedance of the DSP Tuner Unit. Finding the impedance that delivers half-power into the PreAmp ckt (as the DSP Tuner's "load") will give the required value for the source resistance (Rs) used in Spice simulations. Since the DSP Tuner is DC biased (1.7V) at output pin 27 for LEFT channel, a DC blocking cap (C19 = 1uf) and series resistance (R9 = 470) is required to prevent short-circuit failures in the DSP Tuner IC.

Using Spice, the output impedance [Z] shows to be apx 1200 ohms. And this value gives a good fit to measured results, so all schematics have been modified to use this value. All Spice simulations were re-run. This new value has changed some of the component values required for "Treble Rolloff - 2nd Method" and "K6JRF PreAmp Ckt Mod - 3rd Method". In general, all steady state AC frequency measurements have changed along with the recommended cap values. This relatively high impedance plus the four caps [C11, C21, C12, C22] contribute to the rolloff at higher frequencies.

The latest 5Spice analysis shows different results, more in line with bench measurements. In order to be complete, I've added the MUTE ckt into the schematic so that the "total" ckt can be analyzed.

XDRguy asked about R302 (47) and C302 (470uf) ckt components. The R-C ckt forms a low pass filter which decouples the 8.5V power-line from other ckts using this voltage. The attenuation at 10Khz is apx -100dB!

This RC ckt feeds power to the PreAmp LEFT and RIGHT channel so the power consumed in the opposite channel (RIGHT) must be included. The Rrchn resistor represents the DC current drawn by the RIGHT channel preamp, apx 3.25ma. In previous analysis I hadn't included both the LPF and Rrchn. Since both effect the operating DC BIAS for the LEFT channel, these ckt components s/b included. Tnx Dave! The new DC bias values are shown in blue next to the ckt node. Note that there's no change to the "AC" steady state analysis.

Results show -2.14dB @ 15Khz with nominal values for all circuit components.

So, in summary, the observed PreAmp ckt treble loss has variation from different authors ranging from -1.5dB to -2.5db. Until I actually make some measurements myself, I still believe that the DSP Tuner Module plus the PreAmp ckt cause the observed attenuation at 15Khz but, to date, measurements by others don't confirm the idea. But, bottom line, we now have a realistic circuit simulation model for an unmodified Sony XDR FM Tuner's PreAmp Ckt.

Total Harmonic Distortion [THD] Analysis
I stated earlier that you shouldn't expect to measure excessive THD from three (3) caps, two (2) transistors and a few resistors! So I decided to see what 5Spice says about the LEFT channel of the SonyXDR Preamp ckt.

5/8/16: Spice model for 2SC3052 transistor employed to update results.
The original analysis used a "2N3904" NPN AF transistor for Q102 and Q103. After finding a correct model
[2SC3052] for those critical transistors, ALL ANALYSIS were repeated and, as expected, some results changed!

The 5Spice analysis shows that I was partially correct! At low signal levels (0.25Vpk), the preamp's THD performance is quite good but due to the limited operating headroom, THD rises with higher signal levels. The results are shown in the chart.

The "stock" preamp has a Low Pass Filter [R/C network R101 + C12 and R6 + C22] that is intended to attenuate high-frequency 'buzz' beyond 20Khz. It also begins to rolloff starting at 5Khz so this action lowers the THD but also causes excessive treble rolloff as you saw previously.

To 'tame' the LPF network, remove the two (2) capacitors (LEFT channel) and that reverses the treble rolloff. The THD results is shown under the "Mod %" column. The THD is basically unchanged with or without the 4 caps although there is a general rise at higher frequencies.

At higher signal input level, 0.75Vpk, the THD rises as shown under the "Mod %" column. This is due to the limited overhead operating range. Voltage values are noted in the schematic. Higher input signals forces the ckt to approach the voltage "rail" so THD will rise. The operating DC bias numbers in blue are shown in the schematic.

I thought that lowered THD is still possible by making a few ckt changes. However, after doing a visual inspection, it is not that easy to change ckt values on a SMD manufactured board. Certainly it can be done but after some Spice analysis, the lowered THD that would result isn't that significant.

However, adding Negative FeedBack (NFB) to this ckt dramatically lowers THD by forcing the ckt to operate in its linear range at the expense of lowered preamp gain. This is a small price for such a large performance improvement. Contrary to Yahoo forums "technical" experts, NFB for this ckt, does NOT add phase/time delay, it just lowers THD*

And that only takes one (1) added resistor per channel!

The schematic shows the NFB resistor as Rnfb = 39K as this is the "middle" of the acceptable range of values. The NFB resistor reduces the internal preamp's THD by forcing it to operate in its limited linear range. Looking at the THD chart above in the "NFB" column clearly shows lowered THD, as low as 0.011% @ 0.25Vpk signal level.

Note that low frequency Bass Rolloff [BR] occurs when NFB is added. It's due to the extremely low value of C19, 1uf and must be increased to correct the BR. The added cap is "Clfb" and can be soldered to TU1 DSP Tuner module, pin (27) and then to the wire jumper JW14 on the underside of the Main Board (LEFT channel). Details in the "Installation" section below.

The BR is -0.7dB with NFB resistor, Rnfb = 20K.

 * It does that by reducing ckt gain ensuring that the ckt operates in its most LINEAR range.
These forums are packed with self-named "technical" experts that don't know the difference between a "hairpin" (no, it's not a woman's hair product!) and a "decibel". Their experience is to BUY tuners, listen and get together to "judge" the better/best tuner. Note that they haven't actually MEASURED THD so everything is strictly qualitative! This process relies on the ears and the quality of the hifi gear. With different venues, equipment and "ears", the results are tainted!

However, this process can be fine if they realize their lack of quantitative analysis. Sadly, they don't and make statements that some FM tuners with medium to high THD levels sound better than a low THD tuner. This is simply ridiculous. An FM Tuner is ALWAYS designed for low THD as is possible within the ckt's constraints. It is LOW THD that makes a FM tuner become an audiophile treasure. For the non-technical, this is the ability to reproduce a "note" exactly as it was played w/o coloration! NFB was employed in the SonyXDR Tuner to lower THD and made it produce extremely clean (low THD and ZERO time delay) audio as detailed in the following sections.

Want Lowest THD . . .
For me, I want the absolute lowest THD possible with the SonyXDR preamp. To that end, making the NFB resistor, Rnfb = 20K gives the lowest THD even for input signals (from the TU1 DSP unit) greater than 1.0Vpk. In fact the preamp won't clip with the DSP Tuner's output level = 1.5Vpk! The chart shows the response at various signal levels.

Note the THD @ 0.25Vpk is under 0.01%, 0.04% at 0.75Vpk and 0.05% at 1.0Vpk. These are excellent numbers for basically a one transistor amplifier (2SC3052) followed by an emitter follower (2SC3052) output with both operating on limited, single voltage rail!

THD in "dB" . . .
Express THD in different units, "dB".

THD is normally expressed in percent but technical people get a greater "feel" when THD is expressed in "db".

THD is defined as the ratio of the rms voltage of the harmonics to that of the fundamental component. This is measured by using a spectrum analyzer to obtain the level of each harmonic and performing an rms summation. The level is divided by the fundamental level and stated as the total harmonic distortion expressed in percent.

The distortion of an audio device indicates the extent to which a sinusoidal test tone signal is modified by non-linear distortions of unwanted overtones and harmonics. It is therefore a measure of the resultant harmonic distortion.

At the "low" level of 0.05% THD, the db equivalent is -66db. That means the distortion level is more than 2000 times down from the fundamental! So if the room's sound pressure level is 10W, the distortion level is 5 milliwatts!

At a "higher" level of 0.5% THD, it's at -46dB. The distortion product is more than 200 times down from the fundamental. For the 10W example, the distortion is 50 milliwatts!

Do you honestly think that you can hear these distortion products??
Note that the overall preamp gain is reduced when NFB is employed. The stock preamp has 1.84x gain input to output. Adding NFB (68K) lowers the overall gain to 1.46x. For the SonyXDR preamp w/ Rnfb = 20K, the gain is 1x. For my audio system, the gain loss makes the other equipment's output balance out perfectly and I have a lot of "headroom" in the Sony power amplifier!

Employing NFB, Spice says that there is -0.7dB @ 20hz Bass Rolloff [BR]. It is inversely proportional to the value of the NFB resistor; 68K = -0.3dB; 20K = -0.7dB. When NFB is used, it requires the addition of cap, Clfb = 10uf in parallel with C19 = 1uf. This removes the bass rolloff and flattens the response to -0.05db @ 20hz.

Note that with 1.5Vpk drive signal without NFB, it is clipped on the negative rail, THD = 12.1%. With NFB, the output is clean w/ THD = 0.038%. Not too shabby! With the 20K NFB resistor installed, your SonyXDR tuner's preamp WILL NEVER CLIP on the strongest INPUT signal and will recover audio that has the range of 0.0x% THD [lowered by -40 [min] to over -60dB] to boot! Note also no delay time [zero crossing point is identical] or phase shift effects added with NFB.
The chart shows the expected levels of THD vs the selected NFB resistor with range between 68K and 20K. In general, the lower the resistor value, the lower the resultant THD. The reason is simple; the lower resistor value reduces the preamp's gain from apx 1.9x with-out NFB to 1.0x with 20K. The use of a 20K NFB resistor guarantees that the preamp will not clip and, at the same time, will produce the lowest THD possible for the preamp. It is the natural choice.

If you compare the preamp gain vs NFB resistor values shown in the chart, it is a no-brainer that NFB is the way to go. The onset of clipping for the SonyXDR internal preamp is just above 1.0Vpk input. The THD is over one order of magnitude higher w/o NFB. This is probably why the SonyXDR FM Tuner's performance without NFB makes it sound "dull".

The tradeoff for this THD improvement is a loss of gain through the preamp, a small price to pay for the lowered distortion level. It's an understatement to say that NFB has improved the quality of my preamp's audio output!

NFB Resistor + BR Caps Installation . . .
I recommend that you use a resistor similar to this one at Mouser  20K, 1/8W film resistor and solder between the junction of R110 and C101 on the "input" side. And between C102 and R109 on the "output" side for the LEFT channel. For the RIGHT channel, between R210 and C201 on the "input" side; between C202 and R209 on the "output" side as the picture shows. Suggest that you use shrink sleeving so it can't short on the many other parts in that area.

When complete, you can check with an ohmmeter; since the two 20k resistors are isolated, a resistance measurement should read 20k ohms. If it reads other, then you have soldered the NFB resistor on the wrong side of the R and C combination.

Be careful on the amount of "heat" (650-700 deg F iron) that is applied to the junction being soldered. I use a small flat blade screwdriver to hold the wire to the part after the joint has been 'fluxed'. Too much heat and the SMD cap or resistor will be moved or, worse, slide off.

The addition of 10uf cap is somewhat problematic. The BEST solution is to use a small, low voltage axial tantalum cap. These are small enough so they can be soldered to the underside of the Main Board [MB] and still be able to mount the MB on the posts.

The tantalum caps 10uf, 6V solid tantalum capacitor arrived from Mouser. They are 2.8mm (0.1") diameter and 7.8mm (0.3") long.

The picture shows the convenient way to mount the small tantalum caps; the LEFT cap is soldered to pin 27; the RIGHT to pin 26. The negative leads solder conveniently to JW14 and JW15 respectively.
If you prefer using a 'standard' aluminum cap, the picture shows how to do this. The 10uf, 16V radial caps and mounted on the top side and the holes in the MB allow "ONE" lead (+) to extend to the MB bottom side. It's soldered on the appropriate pin of TU1 module, DSP Tuner Unit.

For the LEFT channel - Pin 27; for the RIGHT channel - Pin 26. The LEFT cap's negative lead solders to JW14; RIGHT cap's negative solders to JW15.

Frequency Response [FR] & Total Harmonic Distortion [THD] Measurement
This was the first set of measurements made by Dave, XDRguy * using my mods on a SonyXDR-F1HD. As he is interested in improving his external preamp, so am I for the SonyXDR's internal preamp.

He modified a stock SonyXDR that he was preparing for sale, so he could make FR and THD measurements with his labratory grade equipment.

For the THD measurements, he modified a stock SonyXDR by removing four (4) caps [C11-12, C21-22] and adding a 68K NFB resistor. Results show apx 0.05% THD across the frequency band from 20hz to 30Khz under low signal level. THD is flat until a slight rise at 29Khz = 0.07%. Impressive indeed!

However, with lower NFB resistor values, the THD can be further reduced!

To that end, Dave modified another stock Sony XDR to include the NFB resistor = 33K ohms and Clfb=10uf. Then ran a FR and THD analysis producing the following two (2) charts.

The FR chart shows the results of the Sony XDR with K6JRF mods. The FR is flat from 20hz to 20Khz +/- 0.12dB. The chart shows the THD for two (2) values of NFB resistor and a 5Spice data from a Distortion Analysis.

For the THD measurements, the AQUA colored curve shows the actual measured THD for the modified Sony XDR with NFB = 33K ohms. The BLUE curve shows the projected THD+N with the recommended NFB resistor = 20K ohms. This curve shows THD less than 0.07% from 40hz to 28.1Kz.

The data shown in the RED curve shows the Spice THD analysis with drive signal, Vin = 0.75Vpk. The THD is less than 0.04% from 20hz to 20Khz.

The THD curve in RED follows the shape of the other charts except for frequencies below 60hz. THD analysis of this kind require a circuit settle-time due to ckt R-C time constants. Measurements performed as in a 'transient' analysis will show much higher THD numbers than are actually there when the THD measurement is made if the ckt has NOT been allowed to settle. Differences of an order of magnitude [0.25 vs 0.02] are commmon. I suspect that the THD measurements didn't settle so that accounts for measured differences. All three (3) curves show good correlation except as noted.
*  XDRguy made the above measurements and I'm pleased with the results. For those who have been following my Sony XDR mods, they are made from 40+ years of electronics design experience and verified with Spice simulations. However, I don't have the test equipment to make FR and THD measurements. Dave's work, on the other hand, is anchored in testing and measurement. Again tnx for quantifying my XDR modifications.

Spectrum Analysis
When the FCC approved the transmission of stereophonic sound, it had to be compatible with FM mono rcvrs. To do this, the 0 to 15Khz base band part of the multiplex (MPX) signal had to contain the left (L) and right (R) channel information (L+R) for monophonic reception.

Stereophonic sound is achieved by amplitude modulating the (L-R) information onto a suppressed 38Khz subcarrier in the 23 to 53Khz region of the base band spectrum. A 19Khz pilot tone is added to the multiplex signal to enable FM stereo receivers to detect and decode the stereo left and right channels. The composite base band signal format meets the backwards compatibility needed for FM mono receivers while simultaneously providing enough information for FM stereo receivers to decode the left and right stereo channel outputs. Today’s MPX signal includes a 57Khz subcarrier that carries RDS and RBDS signals.

Using my modified Sony XDR, a spectrum analysis showing KUSC and KPBS made on 12/11/16. Object was to see how much actual BW is employed by the more popular FM radio stations. Was surprised to see the KPBS_HD3 employs only 8.9Khz of operating bw.

In contrast, the KUSC_HD1 spectral data was taken from a live broadcast of the LA Phil with the LA Master Choral group. Note that 16.1Khz of operating bw was measured. Also note the large amount of "rumble" shown in the 20hz to 40hz area. For contrast, a plot of white noise is shown to be very flat across the 20hz to 20Khz frequency spectrum showing that the audio recorder/player produced an accurate spectral analysis chart.

Data was captured by use of Spectra Plus-SC program. Details here

External Preamp
Using a modern day opamp to replace the discrete transistors used in the SonyXDR Preamp should provide 'clean' audio to your receiver's amplifier to make room filling high fidelity.

Along this line, I decided to employ a LT1632 dual opamp in a 5Spice simulation to see what improvements could be had over the SonyXDR's internal preamp. The LT1632 specs are impressive and should provide significant improvement processing the recovered audio.

My design was captured into 5Spice which serves as the instrument that 5Spice "reads" and then analyzes. The ckt design is straightforward inverting amplifier using the LT1632 opamp to simulate the LEFT channel of the SonyXDR tuner. SonyXDR's 8.5V is used as the power for the opamp. Two resistors (Rdcb1 + Rdcb2) establish 4.25V bias to the non-inverting opamp input (pin 3). By using "ac-coupling" (Cin, C5), there is more than +/-3.5V signal swing range that will handle the largest recovered signal from the DSP Tuner Unit, TU1, pin 27 for LEFT channel.
The previous chart has been updated to show the THD values for the SonyXDR internal preamp and the K6JRF external preamp using the LT1632 opamp. The THD in general is three/four-orders-of-magnitude lower than the internal SonyXDR preamp. It clearly shows the difference between a multi-transistor IC amplifier vs one gain stage amplifier (Q102) + emitter follower output (Q103). Impressive indeed!

To make the point clearer, consider 1Khz THD @ 0.75Vpk drive to both preamps:
Sony internal preamp modified plus NFB = 0.015%
K6JRF external preamp = 0.0000047%
That's four orders of magnitude lower THD!

In defense of the internal preamp, the single transistor amplifier Q102, 2SC3052 does maintain excellent linearity under the limited voltage range and is dramatically improved with NFB. Even though the K6JRF preamp shows extremely low THD levels, those can't be measured with current instrumentation. A typical measurement will show 0.01% - 0.02%. With NFB employed in the internal preamp, THD levels are below 0.05% even at 1.5Vpk signal drive levels.

To use the external preamp to attenuate the high frequency "hiss", the frequency compensation capacitor (Ccmp), when set to the appropriate value, can effectively attenuate this noise. The right chart shows Ccmp set to 100pf and 10pf. The ckt does dual-work but there's a slight THD increase at higher frequencies. However, this ensures over -17dB attenuation @ 1Mhz as the chart shows.

In XDRguy's standard upgrade package, he offers a dual channel preamp that replaces the internal preamp. The bench measurements show THD at the limit of the S/N + N ratio of his test equipment which has been verified with my preamp 5Spice analysis. Note that K6JRF's external preamp is NOT XDRguy's version.

If you perform mod #1, all you need do is remove C12 + C22 for LEFT channel, C11 + C21 for the RIGHT channel and add 10uf each to C19 and C20 for LFB rolloff. Next I recommend that you add a NFB resistor (20K) to each preamp channel to ensure that you will get as low as possible THD (0.004% @ 0.25Vpk) recovered audio.

Internal vs External Preamp Conclusions
I've decided to stay with the internal preamp with my NFB mod = 20K ohms. Finding the correct Spice model for the 2SC3052 transistor resulted in ACCURATE DC Bias and THD distortion results which in turn, has given excellent correlation between Spice simulations and real world measurements.

Some have reported that listening to a stock SonyXDR is "hard" on the ears. Spice analysis shows Treble Rolloff and high THD numbers making that conclusion very probable. However, removing four (4) caps, employing NFB and performing the BR mod, THD is now in the range of 0.004% at low (0.25Vpk) and 0.03% at high (1.5Vpk) signal levels! There's nothing "hard" about those numbers. Additionally, it will NOT clip under strong signals as it did w/o NFB. IMO, NFB makes the SonyXDR Tuner an audiophile contender!

On the other hand, the K6JRF LT1632 designed external preamp produces the lowest THD, orders-of-magnitude below the internal modified preamp. However, the modifications to the Sony chassis to incorporate this external amplifier are more complex than adding one cap (Clfb) and one resistor (Rnfb) per channel (after removing C11,C12,C21,C22).

As you have seen with verified results, the internal preamp produces unacceptable THD w/o the K6JRF mods and the use of NFB. With them, the THD levels are low enough that it shouldn't be a problem even for highly discerning ears!

If you decide to make a change, there are are three (3) methods to fix the treble rolloff problem;
Mod #2 and #3 do not use NFB and BR. The preamp runs w/o feedback and could be overloaded producing high THD audio.
Note: Mod #1 is mandatory in preparation of NFB + BR use.
1-   Remove four (4) chip capacitors [C11, C12, C21, C22] - this results in a flat response to 20Khz but does not remove the ultrasonic noise. No big deal b/c you can't hear it anyway!
2-   Add one cap to neutralize the treble rolloff - this will correct the treble rolloff w/o removing any caps. Signal THD is is OK at low-to-medium signal levels. Not as good as method #1.
3-   Replace the four (4) chip caps with new values - this will give a balanced low and high freq rolloff and attenuate ultrasonic noise. Preamp audio THD is low (0.01% @ 10Khz for LOW signal levels) but can/will be overloaded on high signal levels.

Chip Cap Removal - 1st Method
I have a IC and chip-part de-soldering tool, so I chose to remove the four caps.

The PCB layout picture shows that the four caps are located inside the BLUE rectangle. With the proper tool it is relatively easy to unsolder each cap. The key is 'take-your-time' . . . and all will be well.

After removing the filter caps (C11, C12, C21, C22) the Spice analysis shows a flat output response to 20Khz. However, the ultrasonic noise that was previously attenuated will now be present. You can't hear it but it's still there. The cap removal (C11, C21, C12, C22) method has this drawback.

In summary, for the PreAmp Ckt itself, the -2.5dB treble rolloff can be reduced to a low level by removing the four capacitors and the overall THD lowered by adding a NFB resistor.

Tuner Treble Rolloff Spice Analysis - 2nd Method
I decided to check the results using 5Spice using the Service Manual [SM] schematic. When captured into 5Spice the Spice compatible appears as shown on the left. This is the (LEFT) channel only but both are identical. The load impedance is 47K || 330pf. All reference designators are as labeled in the SM.
The Audio Mute Ckt was modeled and found to NOT effect the frequency response of the Preamp Ckt so it was omitted.

A standard "AC" steady state frequency analysis was run with the recommended cap, 0.01uf, (labeled "Cext" in schematic) and the results are shown in the right chart.

There is apx 0.3dB rise @ 10Khz (BLK line) with the 0.01uf cap that has turned a -2.14 treble droop into a small, 0.3dB treble boost! For older ears that might be welcome but it starts to rise at apx 5Khz so it will sound "shrill". If you want a flatter response, as shown on the RED line, you should use 8200pf [0.0082uf] cap.

Mouser is a great source for quality parts are decent prices. They have a Vishay Disc Ceramic cap, 8200pf, 50V rating for $0.26 each. Cap size is 3.8mm wide by 3.8mm thick and will fit under the ckt board. Direct link to 82000pf Ceramic Cap here.

As to where to attach the caps, consult the PWB picture above showing where to solder the added caps.

K6JRF PreAmp Ckt Mod for "Flat" Response - 3rd Method
If you want a balanced freq response, the following changes result in balanced LOW and HIGH frequency response from the Sony XDR Audio PreAmp ckt.

The schematic shows the LEFT PreAmp channel response @ TPv4, in the RED for "before" and "after" ckt value changes. This chart spans frequencies from 20hz to 20Khz showing the family of curves generated with different values for C12 and C22. Stock value of 22000pf gives the most attenuation whereas 470pf gives the least loss.

Replacing the two caps with 470pf gives a balanced low and high frequency rolloff = -0.2dB from 20hz to 20Khz. This is the recommended value.

The advantage of this method is the ultrasonic noise spectrum coming from the DSP Tuner module is attenuated. Note you can't hear it (unless you have dog ears!) but that's what the filter network (R101 + C12 and R6 + C22) does. Completely removing C12 and C22 defeats that purpose.

Don't forget the RIGHT channel PreAmp. Make the same changes for C11 and C21. That's all that is needed!

"T" Pad Attenuator
For my setup, the stock Sony XDR tuner's output is a bit too hot for both of my satellite and OTA rcvr's recovered audio, so I added a 3.25db attenuator to make it more compatible with the overall system sound level. Used two (2) RCA unbalanced attenuator "carcass" to mount the three (3) resistors as shown.

With the use of NFB modifications to the internal SonyXDR preamp, I no longer need to use the 3.25dB attenuator. The preamp output level is compatible with my satellite and OTA equipment output levels.
Yagi Design Details Section

This section has four (4) separate design sub-sections;
- 1st is 4el yagi for XLNC @ 104.9Mhz
- 2nd is 5el short boom yagi for KUSC @ 91.5Mhz
- 3rd is 6el long boom yagi for KUSC @ 91.5Mhz
- 4th is 7el long boom yagi for KUSC @ 91.5Mhz
- 5th is Radio Shack UHF yagi for OTA HDTV

XLNC 4el Yagi
Using Antenna Optimizer (AO), the as-delivered stock yagi was optimized for XLNC @ 104.9Mhz. The difference between stock SL yagi @ 104.9Mhz vs optimized yagi is shown below;
Desc--- Stock----XLNC-Opt
F/G --- 7.5db --- 8.1db
F/B --- 8.3db --- 11.3db
SWR -- 2.43 ----- 1.13

The yagi was put up about 3 months ago and it has worked spectacularly for XLNC. The distance is apx 40 miles from my house. When KUSC FM @ 91.5Mhz was 'discovered', it was replaced by the 5el, then the 6el yagi (described in the subsequent sections).

The summary report files can be downloaded and run in AO should you have the program.
The "stock" AO design file modeled @ 98Mhz is here.
The "optimized" [XLNC @ 104.9Mhz] AO design file is here.

Download both files and rename the 'stock' file to "FMSL4EL.ANT" and the other to "FMSL4ELO.ANT" to run with your version of AO.

KUSC 5el Yagi
For KUSC @ 91.5Mhz, the yagi should have at least 9.5db F/G (109 miles) so that dictates a 5 or 6el beam. Preliminary design show that 9.3db is attainable on the same boom (66") with 5els. This AO design shows good "specs" for the short boom.

Desc--- Stock----KUSC-Opt
F/G --- 7.4db ---- 9.3db
F/B ---13.8db --- 10.2db
SWR --- 1.6 ----- 1.04

As before, download the design file here and rename it to "FLSL5ELS.ANT" to run in AO.

The antenna was constructed today and the plots show good gain (9.3dB) w/ 10db F/B ratio. The second chart shows the non-linear spacing and irregular lengths to get high F/G on a short boom. And that means, it doesn't have a lot of operating bandwidth b/c it was designed for receiving KUSC @ 91.5Mhz.

The 'guts' consist of a stock Stellar Labs 4el yagi with a 66" boom length and an extra element. The construction of this yagi is different; the boom is 0.712" square, the elements are not tapered, are 0.396" in dia. The folded dipole driven element extends above (1.8") and below (1.5") the boom center. Never have seen that before. Attached in the feed element is a waterproof 300 to 75 ohm transformer. There is no need for an external balun. Overall it's very strong but light, apx 1.5 lbs, perfect for So Cal.

KUSC 5el Yagi Results
To say it worked as designed, is an understatement. Today's weather is cold (60F) and it has been lightly raining, so there's absolutely no inversion. After boresighting the yagi to KUSC (0 degs), the Sony XDR showed three bars plus HD) signal . . . solid! But as the weather worsened, the HD) was 'blink' meaning that the signal was too weak to use the Sony internal HD decoder. The HD0 would come and go indicating that more F/G is needed. This yagi would be good for the middle area, apx 50 miles max from KUSC.

The pics below compare the XLNC-4el (1st) with the KUSC-5el (2nd) and a "preview" of the KUSC-6el yagi.

The difference in angles, 5degs mag for 4el vs 0 deg mag for the 5el, probably has to do with the gain and pattern of the 4el vs 5el. The supplied "manual" shows that the boom-to-mast clamp is between the driven el and reflector. When I mounted the 4el yagi, the boom-mast clamp was at the beam's balance point which just in front of the driven element. But for the 6el, the boom-mast clamp is on the director 2 side of the 1st director, making for less interaction of the driven element.

Since the 4el was designed for 104.9Mhz, the F/G and F/B at 91.5Mhz is apx 5db. Not so with the 5el. Still the boresight heading is 0 degs mag so the signal is definitely bouncing off "something" to get to my house at that angle. The 6el yagi result is different and answers the beam heading question. Please read on.

KUSC 6el Long Boom Yagi
During some of the recent weather conditions (heavy rain and cold), the HD) is 'blinking (weak) and the analog signal dropped to "2" bars! Albeit a short time, it did drop. Since I have two (2) Stellar Labs 4el yagis, there is enough material to make a long boom 6el yagi. This design will be optimized for F/G as was the previous 5el yagi but on a long boom for significantly more F/G.

Using AO, I've completed the newest design. Being 109 miles from the KUSC xmitter site, I could use as much F/G as I can get. To that end, here's my latest design; 6 els on a 8'-9" boom.

Desc--- Stock---- KUSC-5el---- KUSC-6el
F/G --- 7.4db ---- 9.3db ------- 11.0db
F/B ---13.8db --- 10.5db ------- 11.3
SWR --- 1.6 ----- 1.03 ---------- 1.02

As before, download the design file here and rename it to "FLSL6ELL.ANT" to run in AO.

The construction of the Stellar Labs yagi is a home-brewers delight; it is light but extremely strong. Every element is a solid, one piece tube unlike other yagis where the element is rolled aluminum sheet so the tube has a "slit". It's very easy to extend the boom using an "insert". The "seam" will be sealed so no moisture can penetrate.

Each yagi's boom is 66" long. Apx 41" will be cut from one yagi to make the required length, 105". I plan on joining the two booms with a some sort of a square insert, possibly aluminum. It needs to be apx 0.640" L x 0.640" W to extend the boom length to 8'-9" (105").

The "picture" gives you an idea of what it will look like when complete.
The left chart shows the AO design parameters in the azimuth view. Great F/G with little F/B ratio. F/B ratio is not important where I am; all other FM stations are beamed from Mount Soledad and that puts their signal into the "side" of the antenna where there's 30db or more side null.
KUSC 6el Yagi Photos
The yagi was built and here's some pics showing the construction. Most HAMs save aluminum as if it was gold. So I had a good selection of aluminum tubing for this project. The main need was for an "insert" to "seam" the two halves. Found a 5/8" square hardwood dowel which measured 0.635" L x 0.635" W x 24" L. I used the complete piece and the boom is razor straight.

This morning I took down the 5el short boom yagi and installed the 6el long boom yagi. When I connected the coax to the Sony XDR, there were 3-bars and FLASHING HD) symbol . . . same as before! With the no-inversion weather we are experiencing, this is normal for a lower gain antenna. But this yagi is supposed to have enough gain and it's the main reason why it was designed and built.

My sadness turned to happiness after boresighting the yagi. The HD) was SOLID! 10 hrs later and it's still solid.

One main reason for the difference in beamwidth [BW] of the stock yagi, 67 degs vs 46 degs for the 6el yagi. The KUSC beam heading is 335 degs per their location map above and the best HD) reception was found at 335degs! To say I'm pleased is an understatement. The yagi was supposed to give the extra gain that was needed but sometimes "Murphy's Law" sets in!

I can state that if you want a LONG RANGE yagi, the 6el design is for you. 109 miles with 3bars and HD) solid!
KUSC 7el Long Boom Yagi
For those who want the biggest yagi with max gain and have the material plus time to build it, here's the yagi for you!

The boom length is 12' 6" so it's long and requires a strong boom to support the length. Using the SL Labs yagi as the starting point would require a "truss" to remove the sag. I prefer a reenforced boom to handle the extra length. As of now, I have no plans to build it . . . but you never know???

Using AO, the design info is shown below compared to the 5el and 6el yagis. It's been designed for MAX F/G and low SWR at the design frequency. The comparison data shows;

Desc--- Stock---- KUSC-5el---- KUSC-6el---- KUSC-7el
F/G --- 7.4db ---- 9.3db ------- 11.0db ------- 12.3db
F/B ---13.8db --- 10.5db ------- 11.3 ---------- 11.8db
SWR --- 1.6 ----- 1.03 ---------- 1.02 ---------- 1.11

As before, download the design file here and rename it to "FLSL7ELL.ANT" to run in AO.
The left chart shows the AO predicted F/G and F/B ratio azimuth view. Design for maximum F/G and feed impedance close to 300 ohms as shown by the SWR.

The F/B ratio has not been optimized as it is not important where I am; all other FM stations are beamed from Mount Soledad and that puts their signal into the "side" of the antenna where there's apx 30db or more side null.
Radio Shack Corner Reflector for OTA
Many years ago, I bought the RS UHF Corner Reflector antenna for receiving but never used it. About three years ago, I decided to drop my cable connection for Over-The-Air (OTA) HDTV reception. So the cable company was waived bye-bye and I put the antenna up on a 15' mast at apx 30' over ground.
I've been bitten by the "HD" bug so I replaced my garage FM rcvr with a Kenwood KDC-HD262U rcvr. Features a superb FM rcvr with both digital and analog decoders. If the digital fades or drops out, it switches to the analog mode. Has a single CD player and a USB input. Output power is 22 Watts RMS @ 4 channels and requires 12V which is supplied by small Astron power supply.

Has digital DSP featuring an "EQ Pro" section that lets you tailor the sound to fit your speakers. A superb feature. Price varies from apx $60 to $99. Amazon price was $67.00 but today it's $99 with free shipping! Price varies alot but it's hard to go wrong!

The picture shows receiving KUSC @ 91.5Mhz with both HD) and ST modes. Not bad for 109 miles to the xmitter site.
In order to ensure adequate signal to both FM rcvrs and OTA signal to four HDTVs in my home, I added two (2) distribution amps; a 2 channel FM and a 4 channel VHF/UHF. FM on the left; OTA on the right. As the picture shows, both units are mounted under the eave so it is safe from rain and sun that could produce damage.

Both amps operate on their respective signals; the FM amplifier sends it out two (2) with 11.5dB/channel gain. Note that both amplifiers are NOT "preamps" in that there's no input gain, just output gain to buffer the respective signals for long coax cable runs. The FM cable to the garage rcvr is apx 70' producing apx 2dB loss.

The HDTV input from the OTA yagi is amplified by 7dB/channel and the four channels are distributed via existing coax to each HDTV in my home. I thought it was a good way to use existing coax (from 45 years worth of satellite and cable system wiring) and not have to rewire the whole house. And it works well! The five (5) HDTVs all get OTA as well as satellite. OTA is free . . . satellite is not!
Where I live, Mount Soledad broadcasts Ch 8 and 10 (still VHF) for their HDTV OTA signal. Mount Miguel broadcasts the remaining US HDTV stations while Mexico has one main channel. What all that means is: CH 8 and 10 are off the back of the antenna while Mexico and other US HDTV channels are off the front. So the corner reflector was removed. It received ok but at times, Ch 8 and 10 would black-out due to low VHF gain and multipath, most likely.

So intuitively, I added a VHF element behind the UHF driven element so it could receive VHF signals. The modified antenna received better but still not good enough. Need to consult Antenna Optimizer (AO).

AO showed that the reflector was too long and mounted too close to the UHF driven element. With the corrections made, the summary;
6.8db @ Ch 10 [193.25Mhz] from 180 degs
7.5db @ 570.25Mhz from 0 degs.

So for Ch 8/10 signal reception, the pattern 'reverses' as shown; more gain off the back than the front. The left chart shows the Ch 10 [blue] and UHF [red] pattern.

For CH 10, the feed SWR is very high but its xmitter power is 21KW. Being apx 2 miles away (line-of-site from my roof!) means that there's a lot of signal that can be wasted due feed impedance mismatch! The right chart shows each element's current distribution for CH 10 operation.

The design file CH10-UHF.TXT is here. Download it and rename it to "CH10-UHF.ANT" to run in AO.

The AO sketch shows that the reflector, properly chosen, aids in F/B ratio for both UHF and VHF reception.

Home for SonyXDR FM Tuner
Since the audio/rf "rack" is a bit on the warm side, I built a shelf on the fireplace and that's my Sony XDR's new home. Also it's closer to my "chair" and easy to control with the remote. What would we do w/o remotes? BTW, the sound is just spellbinding. Even my wife likes it!

Hard to see but look closely ... 3 bars + HD) from KUSC @ 91.5Mhz - 109 miles away! SonyXDR tuner with my custom yagi equals great FM stereo! So now I'm done and will just enjoy the FM from my easy chair!

Many times I watch a "sporting" event and select the FM Tuner for the "sound". The sound from a baseball game or tennis match doesn't match the great sound from the Sony XDR Tuner. There's no competition!

Send me Email Icon with your comments