My Ab763 Has Gas

[ajeffcote]

What's the best way to reduce the lows that make the speaker fart without reducing the "depth" of the sound? I don't want to lose audible lows, but I want them very tight and no speaker farts. Already changed the last two .1's that feed the output tubes with .047's. It's better but I still want it a little tighter.

[Chuck H]

You can do some math to figure out pass frequencies of all the coupling and bypass caps ad nausium. Trim it all to where frequencies useless to the speaker are squashed but 90% of what the speaker can use comes through. The speaker will still fart. The reason is that it's only in the range that the speaker can use anyway that the speaker farts. If the speaker encounters a frequency below it's range it simply doesn't try to reproduce it. That's the nature of it's range. Speaker farting is often best solved with a different speaker.

The amp itself farting in the PI and output stage is a common complaint for AB763 amps. Trimming the PI coupling caps is a good start. You can trim them lower than that and not lose any bass. Maybe .022uf. If your using something larger than .0022uf for the coupling cap feeding the PI you should trim it down to either .0022uf or the stock value which will be much lower.

I think the AB763 tone stack starts to act funny if you change the bypass cap for the stage feeding it. I usually prefer to reduce the value of the bypass cap the stack is feeding. Try a 1uf or a 2.2uf cap. The stack will still work like it did but there will be a 4 to 6db reduction in bass to the stage that feeds the PI. You probably won't hear a reduction in bottom end. Just a tightening of the bottom end. Any coupling caps over .022uf should be reduced to .022uf. Considering the impedances of the stages following the tone stack a .022uf cap is passing anything the speaker can reproduce. Higher is fartier (new word) only.

I often see the bias set too hot in these amps. A hot bias will break up sooner but also sacrifices definition. A cooler bias will have better dynamics. As long as you don't get disagreable amounts of crossover distortion you can cool the bias and get better bottom end much of the time.

Old power tubes have crappy bottom end.

NFB mods that reduce the amount of feedback loosten the bottom end.

Last but not least... The AB763 circuit overdrives the power tubes first. Followed closely by the PI and then the preamp stage feeding the PI. So... If your amp (not your speaker) is farting, you are simply dialing too much bass in the EQ for the amp to reproduce. Since the amp is accurate and clean up to the point where the power tube start to clip, if the amp is farting there is too much bass for the power tubes to reproduce. The "correct" method for getting maximum bass and no farting from an AB763 circuit is to dial up the amount of overdrive you want, set the mids and treble where they sound good and then set the bass control to zero. Now dial up the bass until it starts to fart and then back it down a hair to where it doesn't. The amp is now producing all the bass it can. Any more bass going through the circuit will only add flab, not more bass. It's not unusual with an AB763 amp that is heavily overdriven to see the bass pot set to two. Or even zero. Really. With a .047uf mid cap there are actually a lot of lows getting in even if you don't use the bass pot. Trust me. Most AB763 amps are farty because the preamp is capable of supplying more bass than the amp can reproduce when clipping. You MUST back the bass pot down or your just wasting watts on flab.

JM2C

[Steve Conner]

I agree, if you want rid of the farting you will have to settle for a certain loss of low end.

You can't filter out the frequencies that cause farting while keeping the ones that cause audible low end, because they're the same frequencies, or very nearly so.

[ajeffcote]

Thank you for the reply. I tried reducing the actual tone stack, but I didn't like that and returned it to stock values. The cap feeding the PI is still stock, at .I usually prefer to reduce the value of the bypass cap the stack is feeding. Try a 1uf or a 2.2uf cap. 001., I don't think I want to go lower there. Might try it anyway, huh?

"I usually prefer to reduce the value of the bypass cap the stack is feeding. Try a 1uf or a 2.2uf cap." Did you mean a .01 or .022?
There is an .022 in line with a 500P before the reverb driver. Is that the spot, maybe reduce the .022 to .0.01?
After the reverb recovery tube there is a .1/47K where the signal (I think) is mostly bypassing the vibrato circuit and fed to the 220K mixer before the PI input cap. Change that .1 to maybe a .047 or like you suggest, an .022? I'm thinking that would be the best spot. I know I can try all these options to see what I like the most, and I will. I'm just thinking out loud and bouncing my un-educated thoughts off of you who know. And I very much appreciate your help.
About the tone control settings, I do turn the bass down as the volume goes up. The way it is right now after I hit 6 on the volume the bass goes to 0. at 3 on the volume I have the bass down to maybe 4. The bias is set fairly cold, I think it's about 16 watts dissipation at idle where I have it now. I bias a little cold for the trem circuit, which is the bias mod type, no opto-isolator. NFB resistor is standard value. Although I like the sound with less NFB, that gave me an oscillation problem.
I guess what I want/need to do is reduce the bass produced by the pre-amp to a point that is more manageable with the tone controls. Thanks for the help.

[pdf64]

AB763 is generic, covers too many amp models; reverb, output tubes, OT and power supply are all variables within it.

'I bias a little cold for the trem circuit, which is the bias mod type, no opto-isolator'

That's not AB763 then. The circuit used for a bias vary trem may affect the tendency for bias shift.

'The cap feeding the PI is still stock, at ...'
What is that value? I find that 1nF is plenty.

Although I like the sound with less NFB, that gave me an oscillation problem'
Best to get the amp stable with no NFB; otherwise the implementation can be prone to parasitic oscillation, especially when overdriven. It's bad practice to use NFB as a band aid, in this regard.

I think that the coupling to the power tube grids is one area where SF 70's designs are often an improvement; 0.022uF coupling caps feeding 100k grid leak resistors are more resistant to bias shift. However, the much larger coupling cap to the phase splitter (maybe to compensate for any bass loss from this?) tended to make the amps too bassy
Chuck, I seem to remember that pre-amp clipping (asymmetric), caused by the reverb driver, is the next section to clip, after the power amp.
How are you identifying that the phase splitter is clipping?
I'm looking into this at the moment, and am finding that the voltage swing from regular LTPs is way higher than is required by the power tubes; it's just that power tube grid conduction makes it look like the phase splitter is clipping (chopping off the tops -only- of the phase splitter outputs). That's why bias shift / blocking distortion is an issue with power amps, and why the Paul Ruby mod evens things up (by performing a mirror image function on the bottom of the signals).
Take out the power tubes and scope what the LTP can actually do.
Even with my 5E3, the power tubes appear to hit grid conduction before the cathodyne clips, but the margin is less. Also a 5E3 analysis is complicated by the pre amp power supply sag, resulting from the power tube screens being fed via the big dropper resistor.
Pete.

[Chuck H]

I think the best thing to try for now is changing the cathode bypass cap on the second gain stage (the tone stack recovery triode) to 1uf (not .1uf or .01uf). This won't change the tone stack "feel".

Is this amp already modified or is it a build? As noted, there is no AB763 circuit with a bias vary trem. Also, were it an actual AB763 you would have two channels. If this amp is modified or a build there could be some differences that may seem unimportant to the laybuilder/modder but could be causing your troubles.

Also, have you tried the amp with a diferent speaker cabinet to see if it could be the speaker that's farting out?

[drfunkphd]

Put the circuit back to stock values except for the following. Try these one at a time in this order:

1. Change out the electrolytics in the power supply. The power tubes are supplied by 2x100 uF caps if I remember correctly. Change these to 2x220 uF F+T electrolytic caps. If you want to power tube distortion and are turning it up to 7 or above, the stock power supply doesn't filter enough and you get motorboating and farting. Beefing those caps up goes a long way. Bumping the next cap up from 20 uF to 40 or 50 uF helps too.

2. If that doesn't do it, change the electrolytic caps of the cathodes of V1 and V2 to non-electrolytic caps. 1.5 uF is plenty, but try 1-2.2 uF if you want. Fender used 25uF or 50uF caps on the cathodes and that is just too much, plus electrolytics suck there from a performance standpoint.

3. Change the decoupling caps on V1 and V2 from 0.022 to a lower value, 0.01 or lower. Marshall used 0.0022 uF in that spot, Vox used 1000 pF on their bright channel.

If that doesn't help, then consider a bad speaker or output tranny, but starting with a handful of relatively cheap caps is probably a bit more cost effective.

[mort]

If this is one of those modern Jensen speakers(like what comes stock in a DRRI) it will always fart when you push it. I love the sound of those speakers at lower volumes, VERY sweet. But when they get pushed it's fart city. I don't have experience with every speaker out there, but with the ones I have had experience with, that Jensen is the only one I've had farting issues with.

[Ben N]

'The cap feeding the PI is still stock, at ...'
What is that value? I find that 1nF is plenty.

+1. That did the trick for my DR. If that's not enough, I'd say to lower one or both of the preamp bypass caps to 2.2uf or thereabouts, which will (likewise) still leave you plenty of bottom but tighten things up.

[soundguruman]

1. More headroom in the preamp, higher plate voltage, less sag in power supply.
2. Better speaker, such as EVM.
3. Reduce value of cathode bypass capacitors.
4. Higher quality coupling capacitors.

[Chuck H]

Agree. With caveats.

Many AB763 amps (there were several) already run generous voltages in the preamp and power amp. Increasing voltages in such an amp can get closer to the dreaded Super Twin sound. But if there is room to increase voltages then this can absolutely help. With a second caveat. Changing the power transformer is expensive and may require retrofitting. A simpler solution is emminently probable. But a good idea along your line of thinking is that a solid state rectifier (rather than the typical tube rectifier, if this amp is using one) may reduce sag. Unless the OP doesn't want to change the attack envelope of the amp.

"I" don't like EVM speakers. And I would offer that speaker choice is too subjective to suggest a model without a reference from the OP on tonal preferences other than they don't want it to fart. But the suggestion of trying the amp through another speaker cab has already been offered. That should indicate if the farting is being caused by the amp or the speaker/s. And if the current speakers are indeed not producing adequate bottom end, or at least partly responsible for the farting, then a speaker upgrade is a good suggestion in general.

"Reduce value of cathode bypass capacitors." is too vague. There are many in an AB763 circuit. I wouldn't, for example, reduce the reverb drive or recovery cathode bypass capacitors. Or even the mixer stage capacitors. Further, many players don't like the way the tone stack behaves when changes are made to the first gain stage. So that pretty much leaves the tone stack recovery stage. And a value for that has already been suggested.

Most quality film caps are created virtually equal WRT audio frequencies. I wouldn't expect any performance change unless the caps that are in there are grossley bad.

[jrfrond]

Cut the lows at the FIRST stage by reducing the cathode bypass cap to 1uF or under. Because this affects the LF rolloff of the stage at -3dB/octave, you will still have good lows, but are rolling off the ultra-low "mud" frequencies that can cause that "rolling", "farting", intermodulation distortion at laster stages when the amp is pushed into overdrive.

Keep those .047 caps as your final couplers, because their reduced time-constant in that area will help to keep "blocking" distortion at a minimum when the amp is dimed.

I also like to dispense with the old .1/.047 tonestack caps in favor of two .033 caps. The more "modern" value of .022 is sometimes a bit too dry-sounding to my ears. I also like to up the value of the treble cap to 500pF.

[pdf64]

'reducing the cathode bypass cap to 1uF or under. Because this affects the LF rolloff of the stage at -3dB/octave, you will still have good lows, but are rolling off the ultra-low "mud" frequencies'
Rather than rolling off low frequencies, using this method they're attenuated by up to ~5dB?
Pete.

[soundguruman]

'reducing the cathode bypass cap to 1uF or under. Because this affects the LF rolloff of the stage at -3dB/octave, you will still have good lows, but are rolling off the ultra-low "mud" frequencies'
Rather than rolling off low frequencies, using this method they're attenuated by up to ~5dB?
Pete.

some will be attenuated 5db others more
the lower the frequency the more attenuation
just to give you an idea in a mesa boogie its 22 uF, in a marshall it might be .68uF
you really have to play with it to get it how you like it.

[soundguruman]

What's the best way to reduce the lows that make the speaker fart without reducing the "depth" of the sound? I don't want to lose audible lows, but I want them very tight and no speaker farts. Already changed the last two .1's that feed the output tubes with .047's. It's better but I still want it a little tighter.

Cathode Bypass Capacitor Calculator

TSC

Here are some handy calculators, to help you predict your low frequencies, tone stack,
and all that good stuff...

As you can calculate, a .68 bypass gives you a -3db roll off of about 200 hz and a 22 uF gives you almost subsonic at 15 hz.
Of course, changing the other values in the circuit will affect that also.

[pdf64]

'some will be attenuated 5db others more
the lower the frequency the more attenuation'
I don't see it like that; it's a shelving filter, so with the regular 100k plate / 1k5 cathode resistors, 5dB is the most attenuation there is.
Just to point out that the potential for freqency shaping from varying cathode bypass values is limited.
Reducing the value of coupling caps is the way to make significant changes to low frequencies.
Pete.

[soundguruman]

What's the best way to reduce the lows that make the speaker fart without reducing the "depth" of the sound? I don't want to lose audible lows, but I want them very tight and no speaker farts. Already changed the last two .1's that feed the output tubes with .047's. It's better but I still want it a little tighter.

The last thing that deserves mentioning is the BIAS supply filter capacitor. If the value of the cap is increased, the amp will have more bass response.
If the value of the cap is decreased the amp will have LESS bass response.
AND this capacitor has quite a bit of affect over the amp "farting" in low frequencies, when the output tubes are over-driven.
If the cap is "too large" in value, the output tends to go farting into DC.
Very seldom mentioned, but important!

[olddawg]

Sealed cabinet? Different speakers.

[ajeffcote]

The last thing that deserves mentioning is the BIAS supply filter capacitor. If the value of the cap is increased, the amp will have more bass response.
If the value of the cap is decreased the amp will have LESS bass response.
AND this capacitor has quite a bit of affect over the amp "farting" in low frequencies, when the output tubes are over-driven.
If the cap is "too large" in value, the output tends to go farting into DC.
Very seldom mentioned, but important!

I didn't think of the bias supply filter. I installed a larger cap, 100uf if I remember correctly. Think Fender used a 50uf. I will try swapping it out. Thanks.

[soundguruman]

I didn't think of the bias supply filter. I installed a larger cap, 100uf if I remember correctly. Think Fender used a 50uf. I will try swapping it out. Thanks.

I think the highest is 70.
50 is middle of the road "conservative" fender kind of bass response...
I thought about your original question for a while, and finally realized the bias cap.
Although, we never really think about the bias filter affecting the tone of the amp, it really does a lot.

[Chuck H]

Although, we never really think about the bias filter affecting the tone of the amp, it really does a lot.

I never think of it.

[Chuck H]

...Well?

[pdf64]

My view is that for something to be regarded as significant (eg the bias cap value affects the tone of an amp a lot), then it has to be measureable somehow.
My guess is that the bias cap of an AB763 could be changed form 4u7F to 470uF without affecting any output characteristic of the amp >1dB.
Is there a mechanism proposed by which the bias cap affects bass response?
Pete.

[Rick Erickson]

Pdf - I was thinking the same thing as I read this thread. I have never heard a bias supply cap affect tone. I have heard a bad one introduce hum into the output stage, but it is not a tone shaping capacitor. Even if it was, it would stand to reason that a larger value would pull bass response out of the signal since it is referencing the 220K feed resistors to ground through the bias cap. I typically use 100uF/100V here. Only two ways I can think of to reduce "farting". Either turn down the Bass or the Volume. The relationship of power supply filter/bypass caps to tone should be noted however. A larger supply bypass cap in the power supply will increase low freq response. For that reason, one should be cautious when replacing the preamp filters not to increase their stock value by much, unless you compensate by reducing the preamp bass response, via coupling cap values or cathode bypass caps.

RE

[Tom Phillips]

...The relationship of power supply filter/bypass caps to tone should be noted however. A larger supply bypass cap in the power supply will increase low freq response. For that reason, one should be cautious when replacing the preamp filters not to increase their stock value by much, unless you compensate by reducing the preamp bass response, via coupling cap values or cathode bypass caps. RE

Rick,
Every change affects something but I don't understand how increasing the value of a class A preamp's power supply node bypass cap, assuming that the amp was designed properly to start with, will increase the bass response. Can you elaborate about the origin of this information?

Regards,
Tom

[Rick Erickson]

Tom - I had a great response written to this yesterday. but because I forgot to login first - it disappeared.
Guitar amps are intentionally designed with limited frequency response. Feed an electric guitar into a flat response full spectrum amp and it won't sound very good, at least it won't sound like we expect a guitar to sound. This was explained in the original Alembic Bass Preamp manual published in the mid or late 70's. Before that, Jack Darr published "The Electric Guitar Amplifier Handbook" which explained the interaction of the power supply and the various amplifier sections. As such - you can design a guitar amp around a power supply with large filter & bypass caps, but you will need to limit bass response, usually through low value coupling caps. (.001-.022uF) Fender took the opposite approach. His amps had large value coupling caps (.047-.1uF) and small supply bypass caps. (16-20uF). If you increase the value of the bypass caps in the preamp supply without decreasing the coupling cap(s) you will usually end up with a very "flubby" (sorry for the technical jargon) sounding amp. The bias supply cap(s) in question do not function the same way as the supply bypass caps. They provide a path to ground for the signal on it's way to the power tubes, via the bias feed resistors - typically 100K-220K. In this configuration a very small cap size will let more bass signal pass to the power tubes. Because of the high resistance of these resistors it doesn't take much capacitance to ensure a low impedance path for any audible frequency. Anything over 2uF should be plenty. More capacitance might make for less ripple on the bias voltage, but it will not change the frequency response. This is why bias cap values are typically between 10-50uF. You can go larger - I typically use 100uF/100V caps in a Fender amp, but this is more for the physical size of the cap than anything. Too large of a bias cap can be a bad thing. If it takes too long to charge then there might not be full bias voltage on the tubes when the standby switch is engaged. Just one of many factors to consider when designing, or changing an amp.

RE

[Bruce / Mission Amps]

... Too large of a bias cap can be a bad thing. If it takes too long to charge then there might not be full bias voltage on the tubes when the standby switch is engaged. Just one of many factors to consider when designing, or changing an amp.

RE

Unless the standby switch is in the high voltage center tap, I think all the bias voltage is developed before the standby switch is thrown on anyhow....
It might be something to mildly consider but I'm thinking only if a or the "current limiting resistor" feeding the diode is a very large resistive value and thus a long time constant is developed.
But in reality.... I think this is possibly just a red herring.

[pdf64]

'If you increase the value of the bypass caps in the preamp supply without decreasing the coupling cap(s) you will usually end up with a very "flubby" (sorry for the technical jargon) sounding amp'
Rick, thanks for your thoughts on that, I hadn't considered that possible design rationale when analysing the 'classics' previously.
Is it something that can be measured? Probably not by basic scope / sig gen, as it may be a dynamic response type of thing.
Have you derived a design 'rule of thumb' for balancing B+ bypass time constants to coupling circuit -3dB frequencies?

' Fender ..amps had large value coupling caps (.047-.1uF) and small supply bypass caps. (16-20uF). '
I guess if that approach is taken too far, positive feedback via the power supply could start to take effect (especially as electrolytics age and develop increased ESR), resulting in some strange sub sonic rumblings; just as well for open back cabs.

One point that comes to mind is that rather than the actual value of the caps having significance per se, it's the impedance in the circuit around them (and the resultant time constants / -3dB points) that create the potentially significant system elements.
eg a 12AX7 plate feeding a 50k (trem intensity) pot via a 0.1uF cap has a reasonable -3dB point of 18Hz, whereas a 12AX7 plate feeding a 3M3 mixing resistor via a 0.022uF cap is -3dB at 2Hz, which is getting close to the -3dB at 1Hz of it's B+ supply (16uF fed by 10k for a BF DR). The analysis is of course incomplete and has ample room for improvement but I think I've got the key stuff.
I guess it could be better analysed using control theory / poles and zeros, but that was a tough module and a long time ago.
Pete

[wizard333]

All those things reduce low end.

When you turn a vintage Fender up, the bass needs to be reduced. Simple way that has worked for decades: Turn the bass knob down as the volume goes up. Ta-da! When volume is way up there, the bass control needs to be at 0 or nearly 0. Big, tight, Marshall-esque sound.

Want to get trickier? All these things reduce bass. If you are saying "well I want to reduce flub but not reduce bass" sorry, you're SOL. That is an oxymoron.

Some other things: You could wire a dual pot at the volume so as volume goes up, you run your tone stack output through a passive high pass filter at the same time. Use a dual 1M pot, wire side 2 so that as the volume goes up, your signal is passed through a very small cap across that resistance, which increases as your volume goes up. You'll probably have to experiment with cap values. Also, the pot taper may not roll off bass at the rate necessary, but it would help.

Cathode bypass caps can be dropped to 4.7uf from 22uf without much perception of thinning up the sound. You'll lose some frequencies below almost all the fundamentals on guitar, but meh. The sound will tighten up but still be plenty full. Not enough to crank the volume without still turning down the bass, but it will help. I like non polar electrolytics. You can get 4.7uf non polars from mouser or even your local Radio Shack. I do that on some amps that *Big amp company* puts 47uf Cks in, because *big amp company* wants to impress little kids at Guitar Center volume levels with how much bass there is, but doesn't care that the amp gets massively farty at gig or rehearsal levels.

1uf was suggested above, but at that point you are noticeably pushing high frequencies more than bass, and the amp will thin up at any volume level.

As noted elsewhere, you can reduce the value of coupling caps. The input cap to the PI is the place to start. Socket it, try various cap values and see what works.