Using KT88 tubes in lieu of 6550 tubes

[TomCarlos]

The cathode voltage of 24V adds to the bias.
So actual grid-to-cathode bias is -81V.

I am still waiting for a new stereo jack (to replace the Mix Input) - one with longer leads as newer jacks are not the same physical as the old ones. But while I wait for that, I think I would like to revisit this ... maybe I do not understand the significance of the -81 volts.

I was merely taking a voltage reading from (1) the Cathode to Ground and then (2) the Control Grid to Ground.

(1) So are you telling me that if I was to put my DVM leads on the Cathode and Control grid pins I would measure -81 volts? And

(2) Is there a significance to -81v or some other measurement because the tube can only tolerate a certain voltage difference?

In the end, to properly bias the amp, it would be similar to a Cathode Bias formula?

(Plate voltage - Cathode voltage) * (Cathode voltage / Cathode Resistor) = Actual Dissipated power on each tube (and yes I will account for the three connected tubes). Anything else?

[g1]

Negative bias is at -57vdc. I guess there is room to fiddle a bit.

The cathode voltage of 24V adds to the bias.
So actual grid-to-cathode bias is -81V.

Here's a theory. The bias winding is not up to snuff, or the PT is not original and has a lower voltage bias winding.
This would mean there is not adequate bias range and the cathode resistors were added to get around the problem.
Side effect might the relay chatter mentioned elsewhere (when both relays engaged) as relay voltage is also derived from bias winding and would also be low or weak.

If this were true, R303 and R308 values might need adjusting.

[glebert]

I am still waiting for a new stereo jack (to replace the Mix Input) - one with longer leads as newer jacks are not the same physical as the old ones. But while I wait for that, I think I would like to revisit this ... maybe I do not understand the significance of the -81 volts.

I was merely taking a voltage reading from (1) the Cathode to Ground and then (2) the Control Grid to Ground.

(1) So are you telling me that if I was to put my DVM leads on the Cathode and Control grid pins I would measure -81 volts? And

(2) Is there a significance to -81v or some other measurement because the tube can only tolerate a certain voltage difference?

In the end, to properly bias the amp, it would be similar to a Cathode Bias formula?

(Plate voltage - Cathode voltage) * (Cathode voltage / Cathode Resistor) = Actual Dissipated power on each tube (and yes I will account for the three connected tubes). Anything else?

I'm sure others more expert than me will chime in, but to me
1.) Yes
2.) The negative grid to cathode voltage is the "throttle" of the tube. The more negative the more the current is throttled back. With other amps usually the screen voltage is something like 400V and the grid voltage might be -50V. With the 2000 you are working on with the screens at almost 600V you have to have it throttled way back, hence the -81V bias voltage to keep it at 50%. Here's a chart from a Genelex KT88 datasheet. If you use your imagination and extend the x axis (plate voltage) another 100V up to 600, and then imagine where that -80V curve is going, you will probably not be that far off from where you are.

Caveat, I can never tell if Triode curves apply to the circuit in question, but I know it isn't ultralinear. .

[TomCarlos]

I'm sure others more expert than me will chime in, but to me
1.) Yes
2.) The negative grid to cathode voltage is the "throttle" of the tube. The more negative the more the current is throttled back. With other amps usually the screen voltage is something like 400V and the grid voltage might be -50V. With the 2000 you are working on with the screens at almost 600V you have to have it throttled way back, hence the -81V bias voltage to keep it at 50%. Here's a chart from a Genelex KT88 datasheet. If you use your imagination and extend the x axis (plate voltage) another 100V up to 600, and then imagine where that -80V curve is going, you will probably not be that far off from where you are.

Caveat, I can never tell if Triode curves apply to the circuit in question, but I know it isn't ultralinear. .

Ok.... I think it's sinking in....

In the case of Fixed bias where you have a Cathode connected directly to ground, it's simple - you consider the grid voltage and find the corresponding curve.

In this case, we have a hybrid. So with the added 250 ohm resistor connect to a string of Cathodes, and thus putting 24 volts on the Cathodes, we add that to the Grid voltage to come up with the -81 volts.

I tried to extend the graph and keep the same (projected) curvature of the -81 curve. So it might hit at the 60ma point on the Y axis (just a guess). So yes, based on measurements, we are running at close to 30ma per tube - the 50% point.

Thanks for the explanation !!!

[TomCarlos]

Here's a theory. The bias winding is not up to snuff, or the PT is not original and has a lower voltage bias winding.
This would mean there is not adequate bias range and the cathode resistors were added to get around the problem.
Side effect might the relay chatter mentioned elsewhere (when both relays engaged) as relay voltage is also derived from bias winding and would also be low or weak.

If this were true, R303 and R308 values might need adjusting.

G1 - The winding in question produces 104 Vac. The Revision 1 schematic has R303 at at 10K. My amp has this 10K. Later Revisions changed R303 to 2.7K. I mentioned R308? I think you meant R304. In any event, I think I will leave the bias setting where its at for now. I don't want to mess with this. I have no idea when the Cathode resistors were added.... maybe someone did that when they dropped in the KT88 but still working off the Revision 2 schematic? Who knows.

I went back to Post #10 by Helmholtz and #11 by mhuss. I think they were offering an explanation of why the 250 ohm Cathode resistors exist. And perhaps it was more for stability and leveling out potential differences in current flow in the tubes? I just don't know.

My only concern with the amp is seeing tubes right up against each other!! For a lay person, the first thing that came to mind is heat transfer from one tube to the other and back again!!

For my friend, I will merely point out what we have, your comments regarding spacing between tubes, etc and let him ponder it. Maybe he says, go with what we have and roll the dice. Or if this becomes his main amp, maybe he will ask about 6550 tubes and see if there is more space between tubes and then deciding on whether we stay with Revision 2 (and decide if we keep the Cathode resistors) or move to Revision 3. Rev 3 scares me because we are having to add yet another very large resistor in the circuit (a 470 ohm, 20 watt and plate resistors. I just don't see where I put those in an already busy area. And with Rev 3, I guess we can eliminate the Cathode resistors and recalculate a bias point.

By the way... I did exchange emails with someone at Marshall Amplification in the U.K. I think we (the MEF group) have more documentation and knowledge about this amp than they do. After 40 years, I wouldn't expect for Marshall to have all documentation on this amp. But besides schematics, it would have been good to keep Service Notes on why the changes in the output section and decision to move from KT77s to 6550s with no other changes in the circuit. And then we have the changes from Rev 2 to Rev 3, where they continue to show 6550 tubes but now we have changes in the accompanying circuitry.

Oy Vey ... what a mess !!! But at least the amp passes a signal and is "shredder loud."

[catalin gramada]

I.ll say get lost of bjas resistors in power cathode in this amp. They will induce local feedback with unwanted effects, specially in a shared configuration reducing output power and raise the output resistance and compromise the damping factor of the amp. The tubes looks cramped indeed, mount a cheap fan on a side. Just my thoughts.

Late: btw. I.ll think you.ll be more happy with kt88...perhaps. Just tested some 6550 JJ sent me by error in place of kt66 I asked for. Man 6550 made by jj have no mids. It is a very particular tube, extreme bassy-if this is "fat sound" by any means.

[mhuss]

Good point.
This type of application information is routinely ignored by many modern "designers" that assume that they can just add additional tubes to increase an amplifier's power output. That works in the short term but is not necessarily acceptable for tube life and overall circuit reliability. I see no reason why it should no longer apply for the recommended spacing. The socket orientation could change if a manufacturer were to orient the guts differently with respect to the base pins. However, I don't think that's likely.

Many amps designed for KT88s also have slots in the chassis between tubes to allow for cooling air to circulate. They do run hot when used at 70% dissipation.

[catalin gramada]

And 6550 installed

[TomCarlos]

Great photos guys.... I will keep a copy of those images for discussion with my friend.

Again, the amp chassis "is what it is." If my friend plans on using this amp at "11", we would have a problem. But if he is playing at a low or moderate volume level, and especially when biased at 50%, he'll probably be ok - that is my hope!!!

[Helmholtz]

(1) So are you telling me that if I was to put my DVM leads on the Cathode and Control grid pins I would measure -81 volts?

Yes, use your meter and see for yourself. Black meter lead to cathode instead of ground.
Effective bias voltage (Vgk) always is grid voltage minus cathode voltage: Vgk = Vg - Vk.
Using your numbers Vgk = - 57V - 24V = - 81V.

(Plate voltage - Cathode voltage) * (Cathode voltage / Cathode Resistor) = Actual Dissipated power on each tube (and yes I will account for the three connected tubes).

That formula will do. As always, cathode current is the sum of plate and screen currents.
Actual plate current is about 3% less than cathode current.

BTW, I assume the original KT77s were not quite up to the task.
KT77s were meant as a replacement for EL34s having similar specs.
6550s (and KT88s) are more powerful and mechanically rugged.

[tedmich]

Many amps designed for KT88s also have slots in the chassis between tubes to allow for cooling air to circulate. They do run hot when used at 70% dissipation.

Some folks (the now departed ?) John Chambers of Champ Electronics in UK did not add much space!


https://web.archive.org/web/20211205...ct/const5.html

Sadly no gut shots exist for his 16x KT88 rack mount monster...