6V6 marshal/basman style amp farting out at high volume

[elipsey]

Hi,

My homebrew bassman/marshall style 6V6 amp farts out on the low E string at high volumes. It also seems like it has way more power driving the output tubes than is necessary. I would characterize the amp as aggressive. I don't need the amp to be all around cleaner, per se, but I really don't want the low E to fart out in a sudden and ugly way. It sounds pretty good at lower volumes, so I even forgot about this problem for a while untill I cranked it up again. I have played other 6V6 amps (DR, heavily modified PR, supersonic 22) that seemed to have more clean bass, and that just generaly behave more gracefully at similarly high volumes. I'm looking for guidance on where to go from here.

I have done a lot of different things with this amp along the way. I built it two years ago to give to a friend, but I kept messing with it, and never told him. So far, I've tried using the preamp circuits from a 5f6a, a jmp50, a jcm800, jcm800 w/ channel switch, and reverend hellhound, and the PI has also been through several iterations. I read Randall Aiken's LTPI design guide, and I have used the methods he descibes to implement my own PI stages, in various ways.

I sympathize with daz when he talks about how he can't stop tinkering in this thread. Also, I'm thinking that as he discussed in that thread, preamp voltages might be a missing peice in my amp. Anyway, through all of this, there has been a sudden onset distortion on the E that I don't care for. I don't know whether it's blocking distortion or not, but I have tried several things to address that.

To clean up the bottom end, I tried:

• carefully checked and cleaned all contact points with deoxit
• Tried the amp with another cabinet, and tried it's cabinet with another amp, so I know it's the amp that's farting out, not the speaker or cabinet rattle or something.
• carefully checked for cold solder joints and simple mistakes of that kind
• Various smaller caps at the PI input, down to .001
• Various smaller caps at the PI output, down to .01
• Switched from 12AZX7 to 12AT7 in PI
• 47K PI plate resistors to operate 12AT7 in higher current region of curve
• Lowered bias resistors to 100k
• Increased power tube grid stop resistors to 8K2
• Various tail resistor values from 22K
• Various 6V6 tubes and bias points.
• Different rectifier types for different B+
• Tried 6L6GC and 5881 instead of 6V6. The 6L6 tubes were louder but still farty at high volumes.

Many of the things on that list seemed to help indirectly by reducing bass, but it always would still fart out at least a little bit, and seemed like I was just taking away bass instead of actually solving the problem. I get about 18 watts of clean output on a dummy load with 1khz input signal, and I get a nice clean signal on the scope from the PI input. B+ is about 410 with 5U4GB. PT is DRRI, OT is Weber DR transformer with 6K6 primary and 8Ohm 12" speaker.

Here's the current schematic:


EDIT: I modeled my PI circuit after the one in a Concert II. I attached the Concert II schematic because it was requested, but removed it because it's apparently confusing. Here's a link to it, if you insist. There is one schematic attached to this post now; what you see is what you get.

So here are a couple of things I might try next:

• Experiment more with the PI bias
• Lower the preamp voltages to reduce drive to output stage
• Different output transformer

The preamp voltages are really high, 350 for PI, and over 310 for the rest. I'm thinking maybe this is causing higher driving voltages to the output tubes than I need, and contributing the generally aggressive character of the amp. I wonder if the bassman/marshall preamp is just producing way more output than 6V6s can handle gracefully. I also wonder whether 6L6s are current starved with 6K6 output primary, and maybe that's why they don't work any better.

Any advice?

[g1]

The schematic attachment does not work. Post an accurate complete schematic of the amp as it is now. Also the Concert II schematic.

[elipsey]

Ok, sorry about the broken link. Fixed in the edit, and added the rest of the concert II schem, thanks.

[Chuck H]

I'm not 100% on what YOUR circuit is. The first schematic is accurate to your amp except for the PI? Why not change the values on the first schematic so it's complete? What is the third schematic supposed to be for? It looks like a reprint of the ConcertII schematic PI section. Are all the values visible in the third schematic the values that are in your amp. What about the NFB loop? We need a complete and accurate schematic rather than a several schematics that are partially correct. Even with some explanation it gets fuzzy. Reading a schematic is fine. Reading two with explanations of which is accurate and where is dicey.

I have some ideas but I'll need to know if they apply before I trouble typing my anus off.

[elipsey]

I have some ideas but I'll need to know if they apply before I trouble typing my anus off.

I appreciate that. There is one attachment now, it's the amp I've got.

[Chuck H]

Ok... Rather than write everything in text I just made notes to the schem. Unfortunately this doesn't denote what the changes are intended to do, so please ask any questions you have on the matter. Some of the changes are designed to work together. So not making all of the changes could cause confusing results. It can always be put back the way it was. I recommend making ALL the changes for the purposes of evaluation.

[elipsey]

Thanks Chuck, I think I see what you're driving at.

Would it be OK if I replaced the first filter before the choke with a 40uf cap (instead of 47uf as you advised), and used a 20uf filter at the "D" node? That way I could just rewire my cap can, and I wouldn't have to wait for parts. Also, and I'm sorry for picking nits, isn't 40uf the max input filter for 5U4GB?

If that change would be close enough, I can try it right away.

[g1]

I think he suggested 47uf because that is a commonly available value. The 40uf you have is close enough. And the 20uf at mode D should be ok for the time being. You can add another 20uf in parallel (to make 40uf) when you get parts.

[Chuck H]

What g-one said.

[elipsey]

Ok, I tried it what you said, and it's an improvement. I thought I should try it before I asked questions, but I have a few now. First of all, I apologize for the poor and confusing schematics. I have seen some pretty nice schematics on this site; is there a free software I can use to make my own? How about simulation software?

Chuck, I think I understand why you made the suggestions you made. I will type them out (saving you the trouble of typing your bhole off), and maybe you (or someone else) can correct me if I'm wrong:

1. The larger bias resistor and bypass cap on the first stage increase gain and headroom for that stage.
2. The treble peaking cap added between the 1st and 2nd stage balances the increase in bypass cap at V1a, such that overall voicing of the preamp remains somewhat similar to the previous configuration.
3. Increasing filter for V1 from 8uf to 22uf improves bass in preamp. (I didn't have the cap, so I couldn't do this exactly, but I plan try the 8uf in parallel with a 20uf filter in my cap can, while compensating the dropping resistors so that V1 voltage remains the same).
4. 220k resistor added to voltage divider before the cathode follower stage attenuates output of first two stages for less distortion from cathode follower.
5. Larger tail resistor reduces PI output while improving balance.
6. PI is biased colder by 820R bias resistor for reduced output, maybe adds a bit of compression.
7. Larger 220K bias resistors improve PI clean headroom.
8. Change dropping resistors to reduce PI plate voltage and lower PI output.
9. Increase first filter for better bass headroom from power stage.

It seems like the overarching strategy is to reduce PI output while increasing preamp gain and headroom. Is that what you were going for? Did I interpret the purpose of each part change correctly?

[Chuck H]

1. The larger bias resistor and bypass cap on the first stage increase gain and headroom for that stage.

Yes and no. The bias resistor is intended to improve linearity. Asymmetric clipping tends to make bass notes fuzzy and undefined. The bypass cap increase is just good practice. The first gain stage is the most prone to noise and an unbypassed cathode can be a great source. Fully bypassing the first stage cathode eliminates any noise in that circuit from going on to be re amplified by all the following stages. There isn't enough gain between stage 1 and 2 to do much clipping so this is just good practice.

2. The treble peaking cap added between the 1st and 2nd stage balances the increase in bypass cap at V1a, such that overall voicing of the preamp remains somewhat similar to the previous configuration.

Yup. In fact, if you increase that 470k resistor to 1M it'll be an almost dead on copy of the frequency balance you had with the smaller bypass cap. But the higher resistance adds noise and I wanted to keep the number of alterations reasonable.

3. Increasing filter for V1 from 8uf to 22uf improves bass in preamp. (I didn't have the cap, so I couldn't do this exactly, but I plan try the 8uf in parallel with a 20uf filter in my cap can, while compensating the dropping resistors so that V1 voltage remains the same).

Yes. It's there to give the LF the reserve power it needs to remain dynamic.

4. 220k resistor added to voltage divider before the cathode follower stage attenuates output of first two stages for less distortion from cathode follower.

Nope. It decreases gain in that stage only. We already increased gain in stage one a little. The second stage is more apt to get overdriven and there's a limit (subjectively) to what sounds good. Since gain was increased slightly in stage one (and elsewhere) this compensates in stage two.

5. Larger tail resistor reduces PI output while improving balance.

It doesn't actually decrease PI output by any significant amount. But the added balance may keep the bass a little tighter and improve the function of the NFB loop on the LF.

6. PI is biased colder by 820R bias resistor for reduced output, maybe adds a bit of compression.

Nope. The cooler bias on the PI is actually to add a little headroom and reduce PI clipping a tad. You were biased hot. You may be closer to center with the cooler bias. This may actually sound LESS compressed.

7. Larger 220K bias resistors improve PI clean headroom.

Yes. This does increase drive to the power tubes, which may be detrimental to the farting issue, but was done as a concession to fidelity. Which may prove more important in the long game.

8. Change dropping resistors to reduce PI plate voltage and lower PI output.

Yes.

9. Increase first filter for better bass headroom from power stage.

Yes. The higher capacitance allows for more current reserve on dynamics. When the amp can't reproduce a dynamic the failure is audible in one way or another. Unless it's your speaker that's farting this should definitely be an improvement.

One other change I would suggest is changing the coupling cap between stages 1 and 2 to a lower value. Other than some JCM800's (which are all boom and fizz IMHO) you don't usually see a cascade type preamp with .022 coupling both preamp stages. The first coupling cap is usually smaller. You can get pretty small before you notice a loss of LF. But what you may notice is that by rolling off the LF that you can't get through the speaker anyway you'll free up the amps efforts to better produce audible LF more clearly. I'd suggest a .0022uf for starters.

EDIT: It wasn't easy to notice on the schem but I now see that you're using a 220p treble cap. That cap combined with the 56k slope resistor is actually cutting your upper mids quite a lot. Upper mids are important to give the bass notes some definition. It won't prevent farting, but it may balance thing and make any farting less objectionable if you increase the treble cap to something like 330p or 390p and increase the slope resistor to 68k.

My goal before was not to cut bass, but improve how gracefully the amp handles it. There are still three other things that could be done in this vein. One is the rectifier tube. The 5u4 is pretty poor tube for sag and bottom end clarity. If your voltages aren't too high you might consider trying a 5ar4. I think this would increase your Vp by about 20V, so it's a significant bump. Check your plate volts to be sure the tubes can handle it. Another would be the bias. I didn't ask before because I didn't want to overwhelm you. If your bias is on the hot side that can reduce LF clarity. Too cold doesn't sound good either. Make sure you're biased to about 70% dissipation for the tubes. The last thing would be to use a bassier speaker. Then you don't need to turn up the bass on the amp so much trying to get the amount of bass you want.

[mhuss]

One minor point - the rated max grid resistance for 6V6 and 6L6 tubes in fixed bias is 100k (or 0.1M as they like to say in the datasheets). Obviously, nothing blows up when you exceed this, but they had that limit spec'd for a reason.

[olddawg]

Just out of curiosity, what are you using for an OT? IMHE, the choice of an OT can make a big difference in bass response.

[elipsey]

One minor point - the rated max grid resistance for 6V6 and 6L6 tubes in fixed bias is 100k (or 0.1M as they like to say in the datasheets). Obviously, nothing blows up when you exceed this, but they had that limit spec'd for a reason.

Thanks, that's interesting. I didn't know there was a spec for that.

[elipsey]

Just out of curiosity, what are you using for an OT? IMHE, the choice of an OT can make a big difference in bass response.

FWIW, it's a Weber DR style transformer with a 6600 Ohm primary. I guess that impedance is a little bit lower than textbook for 6V6 p-p, so maybe that would allow a little more distortion, all other things being equal.