Dumping excess Gain

[Fred2202]

I have been lurking around these forums for about a year and have learned a lot. I have modified five different PA and Hi Fi amp chassis for guitar use. A couple have turned out very good...especially one I built on an old Filmosound 16mm projector chassis. The others sound OK but could be better. The biggest problem I have run into with the others is how to dump excess gain in the preamp and smoothing out rough sounding/grainy distortion. The question I have is what is the best way to dump excess gain in a preamp and/or smooth out rough sounding distortion...low value grid leak resistors, grid stoppers on preamp tubes, low value plate resistors (even at 100K I still have too much), high value cathode resistors..or a combination of all of the above. Any comments would be apppreciated.
Most of my chassis have two 12ax7 tubes for the preamp and a 12ax7 phase inverter. One has a ef86 running into a 12AX7 and then straight into a single 6l6GB. This one shows some promise but the preamp distortion is still too hairy.

[loudthud]

One thing I like to do is install a trim pot as the grid leak resistor wired as a volume control. It allows you to rapidly experiment with how much attenuation sounds better where, install one on every stage. Then you can go back, measure where the pots are set and install fixed resistors.

Another thing I like to do is install a resistor on top of the front panel volume pot so if you have a bright cap, it still does something when the volume is at max.

[Fred2202]

Dumping excess Gain

Thanks for the advice. I made up a set of 6 jumpers with the locking hooks a few weeks ago and they have helped a lot with patching different components in. Saves a lot of time soldering and unsoldering. The trim pot would allow more fine tuning.
I have been frustrated with a build I am working on based on the P1 extreme schematic (off of the AX84 site). It uses a cascode for the first stage and a single 6L6 for the power stage. I decided to try an EF86 in place of the cascode in V1. The amp sounds fantastic clean but when I turn the gain up the distortion get very harsh. The transition into preamp distortion is somewhat nasty also. I tried grid stoppers on the preamp tubes and this helped a little but not enough. The schematic uses a 2.2 meg resistor after the cascode...I tried this but it made the amp sound very flat and dull with no sparkle whatsoever. I am in the process of trying different grid leaks in the preamp now. If all else fails I may just rip out the EF86 and try the cascode in V1...in other words duplicate the schematic.

[chipaudette]

I've heard only one cascode-overdrive amp and I really didn't like it. Not at all. Ick.

I'm a little confused about your goals...do you want a clean amp (hence your question about dumping gain) or do you want an amp with overdrive (hence your discussion of cascode)?

If you want a clean amp, the easiest way to cut gain is simply to insert a voltage divider. After the coupling cap of one of your pre-amp tubes, put a 500K resistor in the signal path followed by a 500K resistor to ground. This'll cut the gain by a factor of 2 and reduce the volume by 6dB. If that's not enough to get rid of any undesirable pre-amp distortion, replace the 500K in-path resistor with a 1M. Easy. The trim-pot is a great idea too, but you probably have the resistors for a voltage divider on-hand already. Done.

If you want a dirty overdrive amp, i would pass on doing the cascode. I would use a more traditional Marshall/Mesa approach of cascading (not cascode) gain stages in series. In my opinion, though, you won't get a decent sounding tube overdrive unless you've got at least 3 pre-amp gain stages cascaded together (four is even better).

Don't confuse my suggestion for 3-4 gain stages as a suggestion for super amounts of gain...no. You can certainly get a lot of distortion from just two stages, but I don't feel that it will be a useful or musical sounding distortion. For a musically useful distortion, there needs to be some sculpting of the signal and some sculpting of the amount of gain between different stages. IMHO, you can't do the necessary shaping in just two steps.

So, if you upload something resembling a schematic of your amp at the moment, we can see what you've got to work with and offer suggestions on a good configuration that might be easy to implement.

Sounds like fun!

Chip

[Fred2202]

Dumping Excess Gain

Here is a hand drawn schematic of the preamp and power tube section. I do not want make this amp a metal monster. I would like to get a decent clean sound and be able to dial up some mild/moderate type preamp distortion. I already have a nice clean sound up to 1/2 way on the gain pot and turning the master up gives me plenty of volume. Any further it starts getting pretty raspy and starts to sound like blocking distortion..then when the gain control is about 8 - 10 it smooths out and gets pretty crunchy and sounds OK. I put some grid stoppers on the preamp tubes and it did help somewhat. If I go too much higher on the grid stoppers...the amp loses some treble. I tried voltage dividers and it either kills the problem altogether and leaves me with 100% clean sound only, or the raspy transition into distortion is still there but at a lower volume level. I tried upping the plate resistors but it did not seem to help..only gave me more gain. Should I increase the cathode resistor values?
I have two more preamp tube sockets on this amp that are unused and the transformer can handle them. I can wire in another 12ax7 if needed. Any comments will be appreciated.
By the way this amp is quiet as a mouse for a single ended amp. I can peg the volume and you cannot tell it's even on until you play a note.

[chipaudette]

Wow, I'm not used to seeing configurations like this...but, really, I'm not too used to seeing anything other than Fender/Marshall/Boogie configurations. I'm not at all versed in employing EF86 tubes, so I can't comment at all about what's happening there. Also, the tone control surprised me but is probably cool.

You have some other design choices that did catch my eye, though. Your grid stoppers of 270K and 120K seem awfully high. If you need them this high, you've probably got some other problem. If you put them there to minimize the effects of blocking distortion, I think they're overkill. In fact, I think that blocking distortion is only a real problem with tube stages that employ cathode bypass caps. My understanding is that it's the recovery time constant with the bypass cap when coming out of blocking that makes blocking distortion so horrendous sounding. I could be wrong. But, if I'm right, the 270K grid stopper isn't doing anything regarding blocking distortion because that stage has no cathode bypass cap. Have you tried it without the 270K?

The biggest problem I see in your design is with the 0.01 uF cap just before your second 12AX7. Why is it there? The way that you've drawn it, there is no DC path to ground for the grid of the following tube. This is a big no-no. Pre-amp tubes self-bias using a DC path from grid to ground. You have none for this stage (2nd 12AX7). I'd remove that 0.01uF cap or, if you really want to keep the cap for some reason, add a 1M resistor to ground either before or after the 120K resistor. Make the mod, fire it up, and see how it sounds.

Finally, if you still can't seem to dial in some decent sounding gain, I think that you should try moving your volume control later in the chain. For maximum flexibility, though, I wouldn't move the existing but...I'd actually add another "volume" pot. Having 3 knobs will give you a lot of freedom to find the sound you like. Put the new knob after your last 12AX7 (before the 6L6). You can now think of your three pots as "gain 1", "gain 2", and "volume".

With this 3 pot configuration, I bet that you'll find that you like the middle pot ("gain 2") always set to some value between 4-6. You'll find that the first pot ("gain 1") is good for controlling the amount of overdrive (low for clean, high for overdrive) and that the last pot is good for controlling the overall volume (low for quiet and high for loud). "Gain 2" could just be left at some fixed value all the time. If this is true, you could then actually remove that middle pot and replace it with an equivalent pair of resistors acting like a voltage divider (that's what a pot is, a voltage divider that is variable).

If you are going to add that pot between the last 12AX7 and the 6L6, you'll need to put it between a pair of caps to isolate the DC bias on either side. You've already got one cap in there (the 0.022 cap) so you only need to add one more. I'd add the extra cap and the new pot PRIOR to the existing 0.022 cap. Start by adding a new 0.047uF cap after the 12AX7 and then follow that cap with your new 1M volume pot. Coming off the wiper of the new pot connect your existing 0.022 cap.


So, those are my suggestions. Definitely start by pulling the 0.01uF cap (or adding the 1M to ground). If that still doesn't have the sound you want, add the extra volume pot and have fun exploring the new combinations of sounds (start with the "gain 2" set to 5 and play with the other two knobs).

Chip

[Fred2202]

Dumping Excess Gain

Thank you for the suggestions. Yes the grid stoppers are on the large side but they do seem to help take some of the raspy edge off of the distortion. The EF86 has more gain than a 12AX7 and these values seemed to help smooth out the sound. I tried even higher values but the amp started losing some sparkle and started sounding flat. I tried voltage dividers and it either took away any crunch sound or just lowered the volume of the raspy overdrive sound. I have amp that I modified that uses a 5879 small pentode (less gain than a EF86) in V1 and I have received quite a few compliments on it. I bought into the EF86 hype and thought I would try it. More is not always better. I will pull that 0.01 cap, add a ground reference resistor and try the extra pot before the 6L6. Thanks again for the suggestions.

[chipaudette]

Just as a note...if you pull the 0.01 cap, you don't need to add the resistor. If you pull the cap, the grid leak will be able to go to ground via the pot that precedes your 0.01 cap. It's an either-or kind of thing...either pull the cap, or add the resistor.

Let us know how it turns out. If that tube grid really had no path to ground, giving it one should have a tremendous impact.

Chip

[Fred2202]

RE: Dumping excess Gain

I pulled the 0.01 cap right before the second half of the 12AX7 and it did make a noticeable improvement. It did smooth out the distortion/overdrive in the preamp. It took the raspy sound out of all but the open low E and A strings. I changed out the 25 uf cathode cap on the EF86 for a 1 uf. Now there is just a little hint of raspiness on the low open E string. All in all it sounds pretty good now. I will probably do a little more fine tuning like trying lower value grid stopper resistors on the preamp tubes, and maybe trying a smaller value coupling cap somewhere as I still have plenty of bass to give up. Thank you for your suggestions.

[PRNDL]

Maybe try using the 2nd half of the 12AX7 as a DC coupled cathode follower. That way you'll have only 2 stages of gain instead of 3.

Also, is the cathode of the 6L6 connected to ground (instead of the 32 ohm tap)?

You may also want to try cathode instead of fixed bias.

[Fred2202]

RE: Dumping excess Gain

I set this amp on the shelf to give myself a breather before going back to tweaking. The amp is fixed bias and the cathode is connected to the 32 ohm tap on the output transformer. I do plan to change this to cathode bias as I think it will give a smoother transition into distortion. The arrangement as it now stands is probably providing a lot of negative feedback. The cathode follower sounds like a good idea and is one I have been thinking about. Thank you for your comments.