where to drop gain?

[daz]

I had a thought about the changes i recently made to this amp today. I had a 220k/220k voltage divider where you see 470/470 circled in red in the schematic below with a 500pf on the series R as a treble peaker. I found the 470/470 worked better i guess due to less loading and it doesn't need that cap. Tho i would probably like a bit more brightness and i also am finding the less of a load the better the tone. So my thought was to just go straight to the grid from the coupling cap and a 1M to ground so it would be like a 1M gain pot wide open. But when i do that theres too much gain and i get blocking, or oscillation at least. So the question is, where would you drop gain to eliminate that? I tried a split load on that .022uf coupler that leads to the divider but while i dropped a lot of gain the issue persisted. So again, where would you drop gain?

[g1]

I'm a bit confused with your use of the term "gain" where I think you are meaning signal level. The 470K's form a voltage divider that affects the signal level going into the grid. The actual gain is the amplification factor of the stage which is a function of the Rp/Rk ratio.
When you adjust the divider level, the gain is still remaining constant isn't it?
Not sure if you are wanting to drop the signal level going into the stage or change the amplification level of the stage?

[daz]

I'm a bit confused with your use of the term "gain" where I think you are meaning signal level. The 470K's form a voltage divider that affects the signal level going into the grid. The actual gain is the amplification factor of the stage which is a function of the Rp/Rk ratio.
When you adjust the divider level, the gain is still remaining constant isn't it?
Not sure if you are wanting to drop the signal level going into the stage or change the amplification level of the stage?

I didn't mean to insinuate that 220/220 is less signal level than 470/470 if thats what you mean. I'm talking about changing the ratio of the divider like turning a pot up. A 470/470 divider is like a gain pot turned 1/2 way up. The coupling cap connected right to the grid along with a 1M to ground is like the gain pot full up. What i am asking is this.....in that latter scenario i get what sounds like oscillation on the verge of blocking. So if i were to set it like a 1M pot full up, what could i do elsewhere to relieve that? I just want to see what it sounds like with less of a load there and see how bright it is.

[Chuck H]

If you want to make that stage a little brighter and increase gain a little (from current schematic that is) you can try adding a cathode bypass cap across the 820R cathode resistor. Any value that floats your boat. As small as .022uf and as large as 22uf. You'll get a perceived treble boost up to the standard Marshall .68uf value. After that you'll be boosting mids as well as highs and so on into bass frequencies. To avoid blocking you can add a grid resistor (right on the tube socket). Something like a 68k or 100k that you'll have on hand should do but you can go much larger if you need to. You won't hear much difference in tone (other than a lack of blocking distortion) with the 68k and 100k values. You can also change your coupling cap feeding that stage to a smaller value. That will decrease the time constant considerably and almost certainly reduce any blocking. And I dare you to hear a difference between the .022uf you have now and a .01uf You might think you do, but you'll have to wonder if you're lying to yourself. So any detriment is far outweighed by the faster discharge rate for the coupling cap.

This is where you tell me that you've tried some of this stuff and it didn't work

[loudthud]

Have you tried a grid stopper in the 1Meg all the way up senario? Maybe 100K or ???

[daz]

This is where you tell me that you've tried some of this stuff and it didn't work

LOL!!!! I actually AM laughing out loud. Yep, ALL of those, and many times. Thing is tho, i don't want any more gain and the cap across the 820R, which for a long time was always there in this amp, adds too much gain. And i wanted to see how it sounds with a 1M load there with the signal unhampered as a open pot. The combo of min load and no treble cut due to a divider there is intriguing to me if only i could do that w/o adding gain.

[daz]

Have you tried a grid stopper in the 1Meg all the way up senario? Maybe 100K or ???

No, but i have had one there many times up to 470k and low as 100k and i have tried them at times when i had the same issue. they never worked when it was this bad, and if it did it would take a big enough R to kill high end. But i'll give it a try since the amp has differences from the last time i used one to tame issues like this.

[big_teee]

Looks a lot like the JCM800 Cascade 2204 without the low side.
Your amp has a lot more gain built in to the first and second stage.
The 2204 has on 1a Cathode a 2.7k and .68uf, and on 1b Cathode a 10k resistor no bypass Cap.
Also on the input a 68k resistor, you have a 10k.
On my 2204 I tried bypassing the cathode on 1b and reducing the resistor, and it fed back to easily.
You may want to replace the 1.5k and .68uf on 1b with just a resistor, until you get it controlled, then go from there.
T

[Chuck H]

Great suggestions. Thing is, there are a butt load (technical jargon) of differences to be heard from variable preamp tube bias scenarios. Try bumping the cathodes in tiny amounts!!! Like 1.8k or 2.2k instead of 820r or 1.5k! There will be additional plate voltage that brightens the amp on that stage but gain will be decreased slightly due to linearity from off center bias! If this is done on any clipping stages it starts to lean into asymmetrical territory which accents second order harmonics while still retaining a slightly stronger and brighter attack!!! Oh daz... You've only just begun to experiment

[soundguruman]

Generally, it works better when V1B cathode is 5K or 10K instead of 1.5K bypassed.
The way you got it there might trigger oscillations. You got too much gain in one tube, dude.
You also might try tuning V1 around, so that pin 7 is the input instead.
You might try 50uF in the power supply instead of 20.

So, what I would go with is
V1A cathode- 1.5K bypassed by 2.2uF instead of .68 OR use 4.7uF OR 6.8uF (the + of the cap faces the cathode)
V1B cathode- 10K
and then it should work better.

[daz]

I know i can do those things, but the thing to remember is i came to the values in this amp thru a lot of experimentation and i hate changing certain things because when i do i always end up going back because they just work. The .68uf/1.5k on A and B just can't be changed w/o changing the tone in a way i don't like. However, i will try that again. I tried loudthud's grid stopper idea and just like i said it still caused blocking when i turn the gain up almost full. And that was 330k stopper. I've just done everything imaginable on this thing s you know. I have nearly 10 years F**king with this amp ! I will try the 2.7k on V1. Problem is that makes the .68 too much so i have to lower that value or it gets real thin, and they the amp starts changing a lot. Leave either w/o a bypass cap and amp loses clarity /def horribly. Theres not many scenarios i haven't tried numerous times, but i have not tried no attenuation at V1B>V2A like i'm doing now which is why i am trying to figure out how to control it. When i used the stopper just now the amp worked fine with gain down to where i usually keep it with a 12k, and i liked the brightness i got. But i got too much blocking too early to even consider it. Then i tried 330k and it still did it tho at a higher gain setting. But it also lost that brightness i gained. With 120k and the gain up just enough to not cause obvious blocking, the tone and sustain was better than ever. Thats why i feel i need to have no attenuation at that point. Anyways, if i bias V1 different, say a 2.7k and still use a cap won't i still be getting the same amount of gain? And like i said, i can go w/o a cap or the amp will turn to mud. That i can pretty much guarantee. Been around the block too many times with this amp.

[daz]

Generally, it works better when V1B cathode is 5K or 10K instead of 1.5K bypassed.
The way you got it there might trigger oscillations. You got too much gain in one tube, dude.
You also might try tuning V1 around, so that pin 7 is the input instead.
You might try 50uF in the power supply instead of 20.

I have tried 40 and 60uf. I have a multi can with one unused 40uf section. I have a wire soldered to the unused one with an alligator clip on the end so i can try various values at various nodes. Just tried 60uf ....no change. On the pin 2 vs 7 thing, what will that do? They are just identical tube stages and trying a different tube be as likely to change which side has most gain, yes? (assuming thats what u r saying)

[big_teee]

Generally, it works better when V1B cathode is 5K or 10K instead of 1.5K bypassed.
The way you got it there might trigger oscillations. You got too much gain in one tube, dude.
You also might try tuning V1 around, so that pin 7 is the input instead.
You might try 50uF in the power supply instead of 20.

So, what I would go with is
V1A cathode- 1.5K bypassed by 2.2uF instead of .68 OR use 4.7uF OR 6.8uF (the + of the cap faces the cathode)
V1B cathode- 10K
and then it should work better.

I agree with guruman on this.
Try 1b un-bypassed, and try the bypass on 2a. I think I have a 10 or 15uf on 2a.
I've been all over my 2204 also, and never could control it when driving both 1a, and 1b as hard as you are.
But, whatever works for you is fine.
T

[daz]

Try 1b un-bypassed, and try the bypass on 2a.
T

Thats an idea. I'll give that a try. Never thought to do that when I left 1B unbypassed, which always caused mud. But with 2A bypassed i suppose that could work. Being 820R tho i may have to raise that to 1.5. I'll try that now, thanks.

[soundguruman]

I don't think V1B should be bypassed at all. Too much noise, and too prone to oscillations.
BUT, with a really careful layout, it might be possible. I would not.

You see, anybody can build a high gain input stage, but getting the noise and oscillation under control is an art.
This is where a GUITAR is the best test, to see how far it can be pushed.
The O scope does not tell you all that, it tells you some of it, granted for sure.