unortodox design....will this work

[daz]

I have a EL34 build that i had spent many months tweaking....i'm sure some of you remember. (bad memories maybe Anyways, i long since nailed the tone i was looking for and finally used it at a gig where i was relieved to find it exceeded my expectations. I have no reservations saying that of all the marshalls i used in the past, this sounds better. I was extremely happy. It has a very high gain preamp.

That said, there is one issue i have never been able to figure out. It's not a issue that really affects the tone in any way unless you play much louder than i ever do. But still it bothers me because i have several people wanting me to build them this amp and i don't want ANY flaws in it. Heres the deal....at extreme volumes a extremely ugly distortion sets in. We're talking almost full volume. I am fairly sure it is the PI being slammed too hard because it has a PRE PI master right after the tone stack ala JCM800. So we know it isn't happening before that because with full preamp gain it sounds godly till it's almost cranked. The way i first realized this was happening was that i installed a POST PI master at one point and left the pre master in too. When i would turn the pre master all the way up and used the POST PI master down, the sound would happen till i turned the pre master down a bit.

The sound is a complete turnaround....it's so ratty and buzzy it's horrid. Like pegging the needles on a mixing console sort of. And the bass completely disappears. So picture the worse OD pedal you ever heard and imagine turning it's OD all the way up and turning the bass off and the mids and treble all the way up.

So we know it isn't before the pre PI master because if it were it would sound like that at any volume if the gain knob is maxed. And we know it isn't the power tubes because with the post PI master way down it does it as long as the pre pi master is way up. See what im saying here?

This is where the thread title "unorthodox design" comes in. I have no idea why this amp does this, but i'm wonder if it would be worth trying to add a voltage divider right at the cathode of the cathode follower. Theres already a 100k to ground, so adding another R and messing with values, or removing the 100k and adding a pot to experiment. Or another idea, put a second pot after the pre pi master so as to have 2 masters in a row one after another, then with the first one all the way up, thru the second one down till the nasty OD disappears then replace it with a voltage divider of the values i measure at the pot when turned down to where it cleared up.

Thoughts?

Schematic...

http://s24.photobucket.com/albums/c3...urrent=amp.jpg

[defaced]

Wouldn't changing the 100k mess with the operation of the cathode follower? I'd put the voltage divider you're talking about after the tone stack. I've considered doing the same thing to a build I'm working on because I want more control at lower volume ranges. I would think that with two voltage dividers in a row you may need to put a bypass cap across one of them to maintain the high end in your amp.

[cbarrow7625]

Daz,

The first thing that occurs to me is that you should increase the value of the 470 Ohm and/or 10K Ohm resistors in the PI.

Increasing the 470 Ohm to 820 or 1.2K will reduce the gain of the PI somewhat but it will also give the PI a little more headroom before it goes into hard clipping. Increasing the 10K to 15-22K will give the PI better balance and also reduce the gain a little bit.

If that tone is not to your liking then I would just start by putting a 100K resistor between the tone stack & MV. Keep increasing it until the problem goes away when dimed.

Just curious, does the problem happen with the NFB in & out or only in one switch position?

[daz]

Happens with or w/o the feedback loop. I will try some of what you said in combinations too. I have tried a 1.2k at the PI cathode already and the tail i tried increasing too, tho i don't remember the size. whether i did both together i can't recall, and whether i tried diming the amp to see if that helped this issue or did it for other reasons i can't recall. But i'll give it a go again and try the resistance after the tone stack too. Thanks.

[ETR]

It may just be breaking into oscillation at high settings. If so layout could be the culprit.

[daz]

Don't think so. It's a bit too much to be oscillation i think. Nite and day.

[cbarrow7625]

BTW, you can try higher resistance values in the PI. Higher values will just change the tone a bit more.

One more thing to try is to reduce the 1M grid resistors in the PI to a lower value (as low as 100K). A lower value for the grid resistor will give the PI better "control" over the voltage swing and may help to tame the condition. Remember that in a PI the grid resistor is bootstrapped, so the actual resistance as seen by the grid & PI coupling cap is around 10x the value of just the resistor. A 1M resistor actually acts like a 10M resistor in this bootstrap arrangement. Take that into consideration if you think you will need to increase the PI coupling cap (i'm guessing not). It could just be blocking distortion in the PI. Lowering the PI grid resistor would fix it if this is the case (and probably have less effect on tone than any of my other suggestions).

[Bruce / Mission Amps]

Don't think so. It's a bit too much to be oscillation i think. Nite and day.

I'd mess with the PI gain also but I'm thinking it could be an oscillation too and what you just said above even reinforces my thinking there....

[MWJB]

The schematic only tells half the story (in some cases not even that much) any photos of the layout?

[daz]

I'll try that. In fact i have already thought about trying 220k's in place of the 1megs. Won't be able to till next week tho. Thanks. As for gut shots, no...don't have any.

[Ptron]

What about blocking distortion? That's the first thing I thought of. If that's what it is then I think cbarrow7625's suggestion:

If that tone is not to your liking then I would just start by putting a 100K resistor between the tone stack & MV. Keep increasing it until the problem goes away when dimed.

is probably the way to go.

[daz]

Right, and i kept a mental note of that and all things mentioned. I fully intend to try that, and probably first since it's a 2 second change. Might use a 500k pot and see where if anywhere within that range is stops.

[Merlinb]

Right, and i kept a mental note of that and all things mentioned. I fully intend to try that, and probably first since it's a 2 second change. Might use a 500k pot and see where if anywhere within that range is stops.

If you put the resistance after the MV (in series with the LTP grid) then you can eliminate blocking without sacrificing your max gain level. 100k sounds about right.

[ETR]

Am I missing something? What's unorthodox here?

[daz]

Putting a voltage divider before or after the master volume in a JCM800 type circuit i would say is unorthodox since i have never seen it done. Trying to drop gain at the output of the CF is also something i never saw done. Anyways, those a moot points because everyone who's answered so far has offered up suggestions i can use, and thats what the title was meant to evoke. So it worked didn't it?!