Help please! 1973 Marshall Super Lead - EL34 to 6550s bias question

[Helmholtz]

Honestly I can't read the numbers in the attached schematic, but I found one showing 82k "grid leaks". BTW, I think that the max specified grid resistance refers to the total DC resistance between grid and cathode, so the DC resistance of the bias circuit also adds.

[Leo_Gnardo]

Honestly I can't read the numbers in the attached schematic, but I found one showing 82k "grid leaks". BTW, I think that the max specified grid resistance refers to the total DC resistance between grid and cathode, so the DC resistance of the bias circuit also adds.

I agree Helmholtz. Also, what I observe is that the inverter/driver tube has a hard time driving those 82K's. I often see one side of a sine wave collapsing before the amp reaches clipping. Subbing in a 12AT7 instead of Marshall's 12AX7/ECC83 allows maximum clean power. OTOH many folks are looking for Marshall's "crunch" distortion. A simple swap of drive tube may help dial in the sound, depending on what the player is looking for.

[Helmholtz]

....of course lower value power tube grid leaks are not very attractive (to amp designers) because they load down the PI and thus reduce grid drive.

(Edit: Didn't see Leo's post above. Server response currently is very slow)

[vintagekiki]

... I can just get them down to around 50-55ma. About -51vdc at the grid. Plate voltage is holding steady at 455. They're glowing blue-ish with no red plating....so far. Grid stoppers are still 220k though ...

For your plate voltage 455V
ClassAB Ia = 44 mA
High AB Ia = 58 mA
If you measuring current at the cathode, besides anode current you should have added current Ig2 (5 - 10% Ia)

http://web.archive.org/web/20110621211356/http://diyguitarist.com/Images/BiasChart-KT88.jpg


https://www.grangeramp.com/bias_cal.php

Tip
Bias feed resistors (100 - 220 kOhm) and grid leaks resistors (1.5 - 5.6 kOhm) does not participate in value of bias voltage.

Question
We lost in the comments about anything and everything.
What is really about in the thread?

[Greg_L]

I think this has happened to all of us. When you're sure you've checked and rechecked your work, and "it" (whatever "it" happens to be) still isn't working. Then you change things back and "it" is working. Odds are if you rewired for 6550's again that it WOULD work. You'll never know where an error might have been though. Just take peace in the fact that it hardly matters now. Once you bump the bias voltage a little and install the 150k bias feed resistors (you might want to measure those if you reuse the ones you pulled) it's all the same and good to go. Glad it's worked out.

Thanks man. I too am glad I sort backed my way into it working...so far. I still haven't fine tuned it or played it. But what I've got looks promising.

Actually 6550 manufacturers recommend a max. grid resistance of 50k per tube! But 150k for 2 tubes in parallel (corresponding to 300k per tube or exceeding the limit by 500%) seems to have worked with Marshalls.

I've done the opposite to what I'm doing now. Meaning, I've converted 6550 amps to EL34. In doing that on Marshalls I've always removed the 150k for 220k. In my admittedly limited experience, every 6550 Marshall I've ever seen has had 150k.

What does not connect in series 150 KOhm and 68 KOhm so with mini switch have a choice 150 KOhm or 220 kOhm

[ATTACH=CONFIG]55614[/ATTACH]

That's a cool mod, but I'm not getting into all that!

The earlier Unicord schematic does show the 6550's have 47K grid leaks for the 6550's. The JCM800's used 150K's.
https://drtube.com/schematics/marshall/1959mk2u.gif

Interesting! Can't say I ever saw 47K's in that position. Occasionally 82K though, factory installed.

Honestly I can't read the numbers in the attached schematic, but I found one showing 82k "grid leaks". BTW, I think that the max specified grid resistance refers to the total DC resistance between grid and cathode, so the DC resistance of the bias circuit also adds.

Looks like 82k on that schematic. Mine's a "MkII" though and I know they made changes yearly in the early 70s. That 1970 Unicord schematic doesn't totally jive with some of the later Marshalls. I'm just going by the generic Marshall 100w schematic that calls for 150ks for 6550s.

I agree Helmholtz. Also, what I observe is that the inverter/driver tube has a hard time driving those 82K's. I often see one side of a sine wave collapsing before the amp reaches clipping. Subbing in a 12AT7 instead of Marshall's 12AX7/ECC83 allows maximum clean power. OTOH many folks are looking for Marshall's "crunch" distortion. A simple swap of drive tube may help dial in the sound, depending on what the player is looking for.

For your plate voltage 455V
ClassAB Ia = 44 mA
High AB Ia = 58 mA
If you measuring current at the cathode, besides anode current you should have added current Ig2 (5 - 10% Ia)

http://web.archive.org/web/20110621211356/http://diyguitarist.com/Images/BiasChart-KT88.jpg

[ATTACH=CONFIG]55616[/ATTACH]
https://www.grangeramp.com/bias_cal.php

Tip
Bias feed resistors (100 - 220 kOhm) and grid leaks resistors (1.5 - 5.6 kOhm) does not participate in value of bias voltage.

Question
We lost in the comments about anything and everything.
What is really about in the thread?

My bias feed resistor, if I'm talking about the right one, is 27k. If I drop it to 15k...like the schematic, I can get my bias close for 6550s. The bias pot is nearly maxed to one side, so I need to juggle some resistors still.



My "hottest" tube is about 55ma. The rest are a little lower. I'd like to bias them on the cool side and then tweak to taste.

[Greg_L]

Also, I'd like to offer my sincerest thanks to all of you that have offered your expertise and help so far. I'm learning a lot from "problems" like this. I wish there was some way that I could return the favor, but please take some satisfaction in knowing that you aren't wasting your time on me. I'm listening, I'm learning, I'm trying, I'm making progress, and I couldn't do any of that without help from you guys that don't know me and certainly don't owe me anything. So, thank you.

[Chuck H]

Also, I'd like to offer my sincerest thanks to all of you that have offered your expertise and help so far. I'm learning a lot from "problems" like this. I wish there was some way that I could return the favor, but please take some satisfaction in knowing that you aren't wasting your time on me. I'm listening, I'm learning, I'm trying, I'm making progress, and I couldn't do any of that without help from you guys that don't know me and certainly don't owe me anything. So, thank you.

Just don't go back to the archives and read all my crap from over twenty years ago when "I" was doing all my learning here. That'll be thanks enough

[Greg_L]

Just don't go back to the archives and read all my crap from over twenty years ago when "I" was doing all my learning here. That'll be thanks enough

Lol. Deal.

[Leo_Gnardo]

My bias feed resistor, if I'm talking about the right one, is 27k. If I drop it to 15k...like the schematic, I can get my bias close for 6550s. The bias pot is nearly maxed to one side, so I need to juggle some resistors still.

[ATTACH=CONFIG]55619[/ATTACH]

My "hottest" tube is about 55ma. The rest are a little lower. I'd like to bias them on the cool side and then tweak to taste.

OK now I know which R we're talking about. It wasn't clear, now it is. BTW first time I'm seeing that particular schematic. Notable, the way their draftsman listed "USA variant" parts in a separate column. Other schemos I've seen have the values right on the parts, with alt's in parenthesis. In any case it's evident 27K is too much one way and 15K too much the other way. Perhaps an 18K or 22K would be just right.

[vintagekiki]

In any case it's evident 27K is too much one way and 15K too much the other way. Perhaps an 18K or 22K would be just right.

Since the bias voltage is insufficient because the quiet current is still large, it may be possible to increase the trim pot from 25 kOhm to 50 kOhm.

[Greg_L]

OK now I know which R we're talking about. It wasn't clear, now it is. BTW first time I'm seeing that particular schematic. Notable, the way their draftsman listed "USA variant" parts in a separate column. Other schemos I've seen have the values right on the parts, with alt's in parenthesis. In any case it's evident 27K is too much one way and 15K too much the other way. Perhaps an 18K or 22K would be just right.

I think what I'm going to try is...with no power tubes installed...juggle resistance values until I can get around -50vdc at the grids with the bias pot near it's middle setting so I have wiggle room either way. The resistor at the bias diode (bias dropping resistor?) seems to have a nice effect on this from the testing I've done so far.

Since the bias voltage is insufficient because the quiet current is still large, it may be possible to increase the trim pot from 25 kOhm to 50 kOhm.

Yes if I need to go that route, I will.

[Greg_L]

OK now I know which R we're talking about. It wasn't clear, now it is. BTW first time I'm seeing that particular schematic. Notable, the way their draftsman listed "USA variant" parts in a separate column. Other schemos I've seen have the values right on the parts, with alt's in parenthesis. In any case it's evident 27K is too much one way and 15K too much the other way. Perhaps an 18K or 22K would be just right.

Bingo. I strapped a 56k across the 27k to get the bias drop to about 18k. That gets my 6550s nailed down at 53-55ma with the bias trim pot in the middle of it's range. Good call.

[Leo_Gnardo]

I think what I'm going to try is...with no power tubes installed...juggle resistance values until I can get around -50vdc at the grids with the bias pot near it's middle setting so I have wiggle room either way. The resistor at the bias diode (bias dropping resistor?) seems to have a nice effect on this from the testing I've done so far.

Remember that without power tubes installed, all voltages will read higher than they will with power tubes installed. The current draw of their filaments plus current draw on the high voltage winding will also reduce the bias voltage by a volt or two, maybe even three, when the power tubes are installed. So, shoot a little over the target & you'll land right where you want to be.

Whups, it appears I was a minute late... no matter. Good to see it worked out!

[Greg_L]

Remember that without power tubes installed, all voltages will read higher than they will with power tubes installed. The current draw of their filaments plus current draw on the high voltage winding will also reduce the bias voltage by a volt or two, maybe even three, when the power tubes are installed. So, shoot a little over the target & you'll land right where you want to be.

Whups, it appears I was a minute late... no matter. Good to see it worked out!

Thank you. Yeah my thought process was to get the negative readings at the grids close first, without the tubes installed just to keep everything safe. Then from there I put the tubes in and fine tuned. Seems like it's worked. I'm getting nice readings on each tube now and the bias pot is right in the middle with room to spare either way.

And the grid leaks are back to 150k.

Now.....to play it and see what happens.

[Helmholtz]

How would you describe the sound difference between EL34s and 6550s?