# Thread: Screaming Bright Switch Cap??? - 1974 Fender SF Twin Reverb Master Volume Push Pull Switch

1. Originally Posted by TelRay
V3 ANALYSIS

This is how the signal looks like at V3 pin 6... and, damn! that thing was hot... I had to put the scale at 5 V/DIV (0.5 ms/DIV)

same calculation as before: 3,141 px between peaks and 44 between oscillations is about 28.5 KHz

If I disconnect the REVERB INPUT the wave increases in size and the oscillation disappears

Getting back to the oscillation issue:

It shows strongest at the reverb driver V3 plates, so I assume it is generated by V3 (I have been suspecting V3 for a while). Probably a higher harmonic of the plate signal triggers a parasitic resonance of the reverb transformer, cable capacitance, reverb transducer arrangement.

A possible cure would be an RC snubber (resistor in series with cap) across the reverb transformer primary. To find ideal values will require some experimenting. I would start with 5K in series with 220pF. The idea is to use as much resistance and as little capacitance as necessary to damp the resonance. Too little resistance and too high capacitance will degrade the reverb signal.

Remains the question, how the HF couples to the direct signal, as the reverb springs won't transduce such HF. I assume a coupling via the power supply. Could you scope the supply at point A (red wire of reverb transformer primary)?

1

2. Originally Posted by TelRay
MV is set at @ 4, same settings as in post #225
(anyone has the curve for this 1M-JT potentiometer so it can be empirically demonstrated that at that setting the NFB would not be noticeable?)

by "see" I understand you mean connecting the scope to the MV pot wiper, right?
because electrically we have already measured:
CLOSED LOOP MV Pot Wiper = 1.64 V AC
OPEN LOOP MV Pot Wiper = 0.98 V AC
But I wasn't to know the pot or input level hadn't been changed between posts #225 and #238

If nothing has changed -
Closed Loop: 1.64 * 4.21 = 6.9V output
Open Loop: 0.98 * 7.57 = 7.4V output

The output difference is < 1dB at the speaker.

2

3. en Loop: 0.98 * 7.84 = 7.7V output
FWIW, he measured Go= 7.57 (post #238).

1

4. Originally Posted by Helmholtz
FWIW, he measured Go= 7.57 (post #238).
Thanks, I mistakenly used the calculated Go.

1

5. I will recommend that, whenever possible, use the master up full, reduce channel volume as required.
The amp is based around the non-master version, so full master gives a better indication of 'original' design.
If you need to reduce the master to get the oscillation, that is fine, but for things like NFB test, it should be full up.
I'd really like to see a video like post #208 with the master up full. It would certainly help clear up the NFB questions.

2

6. Yes, otherwise the gain increase when open loop may be getting counteracted by the fall in input impedance (thanks to a midway setting of a 1M master vol).

1

7. Originally Posted by g1
I'd really like to see a video like post #208 with the master up full. It would certainly help clear up the NFB questions.
With the master full up I think you'd see the output voltage nearly double when the 820R was disconnected (the same as it would with the master CCW terminal connected to ground) because the signal voltage on the wiper would be constant.

Why did they connect the master CCW terminal to the NFB input?

1

8. Originally Posted by g1
I will recommend that, whenever possible, use the master up full, reduce channel volume as required. The amp is based around the non-master version, so full master gives a better indication of 'original' design.
If you need to reduce the master to get the oscillation, that is fine, but for things like NFB test, it should be full up. I'd really like to see a video like post #208 with the master up full. It would certainly help clear up the NFB questions.
Again, completely logic. It makes total sense that the NFB works at its best with the MV at 10

NFB

OSCILLOSCOPE

SCOPE at SPKR OUT
VIBR CH VOL @ 3
MASTER VOL @ 10
820R being disconnected / connected during the video

for the last part of the video I've increased the VIB CH VOL to 3.5

VOLTAGES
MV @ 7, VIBR CH VOL @ 3 (nice clean SINE WAVE)

NFB CLOSED
MV Pot Wiper = 1.90 V AC
FBK V = 0.84 V AC
DUMMY = 7.78 V AC

Gcl = 7.78/0.84 => Gcl=9.26

NFB OPEN
MV Pot Wiper = 1.68 V AC
FBK V = 0.00 V AC
DUMMY = 12.15 V AC

Go = 12.15/1.68 => Go=7.23

is the 12.15 to 7.78 VAC increase translated as about +4dB? now there should be a noticeable difference.

1

9. Originally Posted by Helmholtz
The voltage difference must be due to a loading effect by the input resistance of your measuring instrument. What did you use? If you measure across the cap with a DMM, you should get zero voltage.
I was measuring the AC Voltages at both sides of the .1 uF / 200V cap with a Fluke 177 DMM with the ground attached to the amp's chassis
You are right, if I put the ground probe on the other side of the cap I get 0 V AC (unless I push both CH VOL and MV to 10 and then I get about 60 mV)

0

10. Originally Posted by Helmholtz
Could you scope the supply at point A (red wire of reverb transformer primary)?
sure thing! i couldn’t get a clean sone wave here, this is the closest i’ve got

0

11. I am surprized not to see any 120Hz ripple. What is the peak-to-peak signal voltage?

Are you sure you measured at point A (in little square), the "448V" supply that connects to the 20µ/500V filter cap and the TR2 choke?

1

Gcl=7.78/1.9=4.1 ( you used the FB voltage by mistake)
Go=7.23
Go/Gcl=1.76

Means that Go is 76% or 4.9dB higher than Gcl.

The scope signals only show a Go/Gcl ratio of roughly 4/3. The reason for this is that the NFB signal adds to the input signal and increases the apparent Gcl. This effect should reduce at full MV setting.

In fact with this MV wiring the effective feedback ratio and Gcl change with the MV setting and the theoretical formulae can't work properly as they assume the same PI input signal with and without NFB.

1

13. Originally Posted by Helmholtz
Means that Go is 76% or 4.9dB higher than Gcl.
Thx man for the re-calculation and correction

Does this close the NFB chapter of this "book" (post)? With the following conclusion:

- NFB is working
- the effect of disconnecting the 820R will be only noticeable at close to 10 MV settings

0

14. Originally Posted by Helmholtz
I am surprized not to see any 120Hz ripple. What is the peak-to-peak signal voltage?
Are you sure you measured at point A (in little square), the "448V" supply that connects to the 20Âµ/500V filter cap and the TR2 choke?
If you guys have learnt anything about me in the past 256 posts is that: I am not sure of anything
But certainly more sure than on post #1 and thanks to your support.

With the exception that... many posts ago @ #85 we decided to eliminate the 560 pF cap connecting V3 pin 1-6 to "A" in favor of the "blackface" 2.2K / 25uF 25V arrange replacing the 470R connecting V3 3-8 to ground (compliant with Fender DWG 010182 Rev C, which is the one before the addition of the PULL BOOST to the MV)

(old photo with still original components, for illustration only)

the scope settings were:

the SCOPE was set at 20 mV/DIV so I think the answer to your question is 4.75 DIV * 20 mV/DIV = 95 mV (after turning the VIB CH VOL up to 10, biggest waveform)

is that right?

0

15. Originally Posted by TelRay
Thx man for the re-calculation and correction

Does this close the NFB chapter of this "book" (post)? With the following conclusion:

- NFB is working
- the effect of disconnecting the 820R will be only noticeable at close to 10 MV settings
In essence, yes.

What is the max output power into the 4 Ohm dummy load just before clipping?

0

16. Originally Posted by TelRay
If you guys have learnt anything about me in the past 256 posts is that: I am not sure of anything
But certainly more sure than on post #1 and thanks to your support.

With the exception that... many posts ago @ #85 we decided to eliminate the 560 pF cap connecting V3 pin 1-6 to "A" in favor of the "blackface" 2.2K / 25uF 25V arrange replacing the 470R connecting V3 3-8 to ground (compliant with Fender DWG 010182 Rev C, which is the one before the addition of the PULL BOOST to the MV)

(old photo with still original components, for illustration only)

the scope settings were:

the SCOPE was set at 20 mV/DIV so I think the answer to your question is 4.75 DIV * 20 mV/DIV = 95 mV (after turning the VIB CH VOL up to 10, biggest waveform)

is that right?
Point A is correct. Can't read the scope settings from your picture (always make sure that the red var(iable) knobs are in their locked (=calibrated) positions).

Two proposals:

1)Reinstall the 560p cap and see what happens to the HF. The cap acts as the simplest version of a snubber. It might be necessary to wire a resistor of some k in series.

2) Solder an additional (good) E-cap of 20µ/500V between point A (positive) and the ground side of the V3 cathode resistor. Ideally there should be no signal voltage at point A.

1

17. Originally Posted by Helmholtz
1)Reinstall the 560p cap and see what happens to the HF. The cap acts as the simplest version of a snubber. It might be necessary to wire a resistor of some k in series.
Mission #1 completed!

this is the "beneficial" effect of the 560 pF cap, installed as per the drawing corresponding to this amp's schematic seen with the SCOPE connected to the SPKR OUT. I would say it "almost" kills it... only a visible bump where the ripple was before

2

18. following with Hemholz's proposal #1 do you guys recommend I put a 5K resistor in series with the newly re-introduced 560 uF cap?
In that case, should the Resistor be placed between the TUBE and the CAP [POSITION 1] or between the CAP and "POINT A" [POSITION 2]? (might be a super rookie question, yes. I understand one position is High Pass and the other Low Pass)

for proposal #2, does this "Solder an additional (good) E-cap of 20µ/500V between point A (positive) and the ground side of the V3 cathode resistor" mean this?:

and, is this still necessary if the oscillation is killed completely y proposal #1 (after adding the resistor)?

0

19. The additional 20uf cap proposed would be for the power supply circuit to the HV into the reverb transformer. The idea being that any signal on the power supply rail that is interacting may be causing the problem and additional filtration may help eliminate it.

The series resistor is a suggestion that would limit the HF roll off from the 560pf capacitor. It's value would be relative to whether or not the reintroduction of this capacitor stops the oscillation, and then the resistor value could be determined that still stops the oscillation, but also limits the HF roll off of that capacitor in the circuit. It's a matter of drawing a line between compromises. But we don't even know if reintroducing the 560pf capacitor will mitigate the oscillation yet.

1

20. Originally Posted by Chuck H
But we don't even know if reintroducing the 560pf capacitor will mitigate the oscillation yet.
but Mr Chuck, sir.... we do know that re-introducing the 560 uF cap almost completely kills the oscillation (it flattens it out to a remaining “bump”) - see video on post #262
my question was more: should i try to completely “iron” the bump by adding the resistor. if yes, where and what value
and... if this finally kills that problem, is proposal #2 (20/500 cap) still necessary?
thx!!!

0

21. Ah. Ok. Well as I said above, I don't think the resistor is there to make the 560p circuit across the reverb transformer MORE efficient. It's there to make it less efficient so there's less HF loss in the reverb sound. In other words, if the bare 560p cap is doing the job maybe less of that would still do an adequate, but lesser job with less HF loss. It's a matter of compromise. The resistor facilitates the compromise.

I can't help but think there may be other solutions like lead dress changes, ground scheme changes, additional power filtering (as Helmholtz mentioned) or even just shielding. But if you're happy with the tone with the 560p cap in place and it solves the problem there really isn't much to think about.?. It's not like it's an "improper" circuit of itself. Fender probably used it to solve for the very problem you're experiencing after all.

1

22. The series resistor is a suggestion that would limit the HF roll off from the 560pf capacitor. It's value would be relative to whether or not the reintroduction of this capacitor stops the oscillation, and then the resistor value could be determined that still stops the oscillation, but also limits the HF roll off of that capacitor in the circuit.
That^^

But the series resistor - if adequately chosen - would have a second beneficial effect. It extracts/absorbs energy from the HF resonance by dissipation and thus damps or even kills resonance, while a capacitor only might just shift the resonance to a lower frequency.

The complex reverb drive circuit consisting of reverb transformer, transducer, cable, wires and parasitic elements is mainly reactive and could even have more than one resonance. Reactive components (inductors, capacitors) just exchange energy while a resistor "eats" it. As the circuit's behaviour is also influenced by the layout and lead dress, there is hardly a one-for-all optimum solution.

If the 560pF cap suffices, fine. But a little experimenting with different resistor and cap values could be fun.

I recommended to try the additional 20µ/500V filter cap, because there shouldn't be signal voltage on the power supply. A strong effect of the additional E-cap could indicate a somewhat bad (high ESR) filter cap at node A. Easy to try and if not considered necessary, the cap can be used a spare part.

2

23. hi all,

I think it is time for an update on how this thing actually sounds and make sure nobody thinks we are just going for the mathematically perfect amp but there is something still audible going on

- I played the guitar with a nice overdrive pedal with the REVERB INPUT disconnected - nice sound
- CONNECTED the REVERB INPUT - there is something going on on the High Frequency side which sounds like oscillation
- Placed an EQ and cut some dB at 9KHz leaving the REVERB INPUT connected, the sounds cleans up
- Exaggerated the effect boosting at the same frequency to make what I hear more evident

Here a video with all the previous steps in a sequence

Do you guys think that what we have seen on the scope and are trying to kill is responsible for what is heard on the video? (rough calculations on the frequency of the oscillation placed it above 20KHz)

0

24. A guitar signal consists of fundamental(s) and harmonics. Distortion increases the harmonics' content. Typically, the more distortion the more high frequency harmonics. A stronger input signal or a harder attack increases distortion and thus high frequency content.

I can hear an increase of higher frequency harmonics in your video when connecting the reverb input. This might be increased distortion caused by oscillation, but it is hardly the HF oscillation itself you are hearing. Furthermore HF oscillations don't produce a specific, identifiable sounds on their own. So we can't really be sure about the cause of the effect.

I recommend to try my proposals above and report back. I would also expect the effect to reduce by attenuating the grid drive of V3. IIRC, this has already been proposed.

2

25. I'm not hearing anything I'd really call 'glitchy' in that one. The one that kind of stood out to me was post #121 on page 4.
Are you still able to reproduce that glitch?

1

26. I'm not hearing anything I'd really call 'glitchy' in that one.
Agree, but it seems strange that connecting the reverb transducer should change the frequency response/content of the direct signal at all. Would be interesting to see the influence on a square wave signal. Maybe a capacitive coupling between adjacent wires or a spill-over via the power supply.

Question to the OP: Not very probable, but is there a connection between the green reverb transformer secondary wire and the junction between the 1k and 470k resistors (as with the original Master Volume Boost Switch MVS)? If yes it should be removed!

2

27. Originally Posted by Helmholtz
I recommend to try my proposals above and report back. I would also expect the effect to reduce by attenuating the grid drive of V3. IIRC, this has already been proposed.
Jawoll!!! will do!

can you please confirm the placement of the resistor and cap for proposal #1 and #2 as shown on post #263? want to make sure i got it right.

I understand that for proposal #1 the resistor needs to be placed between V3 pin 6 and the 560 pF cap to act as a Low Pass filter (it would be a High Pass filter if I put it after the cap, right?)
You have originally proposed a 220 pF cap with a 5K Resistor, according to one of those online RC Filter calculators the cut-off frequency of that arrange is 145 KHz.
If I now have a 560 pF resistor to get the same cut-off frequency I would need to start with something like a 2K Resistor (142 KHz).

For proposal #2 I just need to make sure that the placement of the ecap and polarity is right (as described in the schematic con post #263)

Originally Posted by Helmholtz
I would also expect the effect to reduce by attenuating the grid drive of V3. IIRC, this has already been proposed.
Are those the 47K resistors added as grid stoppers on V3 on post #75?

vielen dank, sir!!!

0

28. Originally Posted by g1
I'm not hearing anything I'd really call 'glitchy' in that one. The one that kind of stood out to me was post #121 on page 4.
Are you still able to reproduce that glitch?
mmm... i might get lost in the terminology but more than "glitchy" it sounds "farty" (is that even a word?)... sounds like a farting speaker cone
I can try to reproduce the same sound as on that post #121. Will put it here possibly tomorrow using the same clean guitar sound I used back then.

0

29. can you please confirm the placement of the resistor and cap for proposal #1 and #2 as shown on post #263? want to make sure i got it right.
The placements of the caps as shown in post #263 are correct. You may try to connect the lower end (marked "2") of the 560pF cap to the V3 cathodes (pins 3&8) instead of point "A" , as done in other versions of the Twin. Maybe this works even better.

I understand that for proposal #1 the resistor needs to be placed between V3 pin 6 and the 560 pF cap to act as a Low Pass filter (it would be a High Pass filter if I put it after the cap, right?)
The order of components in series wiring doesn't matter, won't change anything.

You have originally proposed a 220 pF cap with a 5K Resistor, according to one of those online RC Filter calculators the cut-off frequency of that arrange is 145 KHz.
Not much sense calculating a cut-off frequency using simple RC calculation. The RC network is wired in parallel to the reverb transformer primary, which behaves mainly inductive. So we have an LCR damped resonant filter (with partly unknown component values) that doesn't behave like a simple RC filter.
I just proposed ballpark values that are unlikely to influence the audio range. The idea being to progessively increase capacitance and/or decrease resistance until the oscillation stops (no real calculation involved).

For proposal #2 I just need to make sure that the placement of the ecap and polarity is right (as described in the schematic con post #263)

Are those the 47K resistors added as grid stoppers on V3 on post #75?
The 47k grid stopper doesn't do much to reduce grid drive. It just makes grid current distortion smoother and reduces HF response. Reducing grid drive can be achieved by replacing the 1M grid leak resistor by 2 resistors of e.g. 500k in series and connecting the grid(s) to the midpoint between the resistors, thus producing a voltage divider.

1

30. Originally Posted by Helmholtz
Question to the OP: Not very probable, but is there a connection between the green reverb transformer secondary wire and the junction between the 1k and 470k resistors (as with the original Master Volume Boost Switch MVS)? If yes it should be removed!
Just re-posting this very pertinent question.

2

31. Originally Posted by Helmholtz
Question to the OP: Not very probable, but is there a connection between the green reverb transformer secondary wire and the junction between the 1k and 470k resistors (as with the original Master Volume Boost Switch MVS)? If yes it should be removed!
sorry I missed that one. The answer is, no. The green cable goes to the REVERB IN RCA jack tip and from it a .0022 uF cap connects it to GROUND (as described on post #133, as per schematics of AA1069 and AA270 silverface amps)

1

32. Originally Posted by TelRay
sorry I missed that one. The answer is, no. The green cable goes to the REVERB IN RCA jack tip and from it a .0022 uF cap connects it to GROUND (as described on post #133, as per schematics of AA1069 and AA270 silverface amps)
I don't think connecting the .0022µ cap to the Reverb In jack makes sense. It obviously doesn't help with the oscillation. Actually I never saw it in any version of the Twin before. Is it shown on an original schematic? My AA270 schem doesn't show this wiring.

OTOH, the cap makes sense on the reverb output - especially when using some cap (560p) at the V3 plates, which produces a lowpass effect. THe reason is that having the cap across the reverb output transducer regains some reverb treble that might have attenuated by the V3 plates cap. In addition the .0022 cap attenuates unwanted HF content with this wiring.

1

33. Originally Posted by Helmholtz
A guitar signal consists of fundamental(s) and harmonics. Distortion increases the harmonics' content. Typically, the more distortion the more high frequency harmonics. A stronger input signal or a harder attack increases distortion and thus high frequency content.
Originally Posted by g1
I'm not hearing anything I'd really call 'glitchy' in that one. The one that kind of stood out to me was post #121 on page 4.
Are you still able to reproduce that glitch?
I think I will try to satisfy both posts by trying to play something like on post #121.
Trying to be as fair to the amp as possible here I've changed guitars from Humbuckers (a Gretsch White Falcon) to Single Coils (a Fender Jazzmaster), connected the guitar straight into the amp (no distortion = no added harmonics at the input and changing pickups I believe also not such a hot input signal) and use the following settings: CH VOL 7, TMB (4,3,4) and MV 3
You can hear it on the video below, my conclusions are the following.
I do not hear those "blips of oscillation" in loud passages as in post #121 anymore (as reported several posts back, even if playing with the Gretsch)
I still hear the difference between the REVERB INPUT being connected or not (as on the video in which I used the overdrive)
At the end of the video i looped the sections in which I can hear that High Frequency "distortion, blip... etc" and took several dB at 9KHz and then boosted at the same frequency to make the problem more evident

I will follow the modifications proposed in the last posts and report back

0

34. Originally Posted by Helmholtz
I don't think connecting the .0022µ cap to the Reverb In jack makes sense. It obviously doesn't help with the oscillation. Actually I never saw it in any version of the Twin before. Is it shown on an original schematic? My AA270 schem doesn't show this wiring.
OTOH, the cap makes sense on the reverb output - especially when using some cap (560p) at the V3 plates, which produces a lowpass effect. THe reason is that having the cap across the reverb output transducer regains some reverb treble that might have attenuated by the V3 plates cap. In addition the .0022 cap attenuates unwanted HF content with this wiring.
I remember that on post #133 I was following Rob Robinette's fixes introduced by Fender for what he called:
"Reduce Reverb Noise & Oscillation: This .002uF cap from reverb pan input to ground was used in the AA1069 and AA270 silverface amps. I prefer the A1172 fix shown at top because it cleans up the pan input and output."
LINK here (you might need to scroll down until you find the Reduce Reverb Noise & Oscillation section"

0

35. Originally Posted by TelRay
I remember that on post #133 I was following Rob Robinette's fixes introduced by Fender for what he called:
"Reduce Reverb Noise & Oscillation: This .002uF cap from reverb pan input to ground was used in the AA1069 and AA270 silverface amps. I prefer the A1172 fix shown at top because it cleans up the pan input and output."
LINK here (you might need to scroll down until you find the Reduce Reverb Noise & Oscillation section"
Thanks, found it.

Seems the AA1069 and AA270 Bandmaster Reverb had the .0022µ cap across the reverb input, while the corresponding Twin versions AA769 and AA270 (https://www.thetubestore.com/later-f...amp-schematics) had the cap at the reverb output - strange.

1

There are currently 1 users browsing this thread. (0 members and 1 guests)

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•