Using KT88 tubes in lieu of 6550 tubes

[TomCarlos]

MEF Team...

While I await parts to finish off the Marshall 2000 amp, I thought it might be best to open a separate thread to ask this question.

For the Marshall, all three Revisions of the amp call for 6550 output tubes. In the amp before me, the prior service tech installed JJ KT88 tubes. A Cathode resistor was also installed for each row of tubes. Since I am not overly familiar with KT88, I thought I would run this by you.

For the circuits I have seen using KT88s, they are connected in a Cathode Bias configuration. This amp is different because it was designed for Fixed Bias. So what we have here is a blend of both.

I am attaching the portion of the schematic for the output tubes - and this is based on the Series 2000 Revision 1 and Revision 2 schematic.

As you can see, each row of KT88s has a HUGE 270 ohm, 10 Watt Resistor. Here are the key voltage measurements.

Plate: 660 volts
Cathode: 24 volts
Grid: -57 volts

Current through the Cathode resistors is 24 volts / 270 ohms = 82.9ma and divide that by three tubes to get 27.63 ma each.

So it looks like the tech decided to run these tubes at 50%.

I guess my question is this.... If the KT88s are nearly identical, why use a Cathode resistor? Why not just use the Negative Bias control?

Thanks, Tom

[g1]

You will have to tweak your math a little as those are 250R, not 270.

[TomCarlos]

You will have to tweak your math a little as those are 250R, not 270.

I guess that is what happens when you have the amp chassis backwards!!

[Jon Snell]

KT88 and 6550 are very similar, the 6550 can't handle quite as much power as the KT88 but not that anyone would notice.
I see no issues.

[glebert]

How far negative does the grid bias control go? Seems like you are at -81V equivalent to be at 50% dissipation.

[Leo_Gnardo]

Looks a belt & suspenders approach. If you don't trust one method, use both. Some safety margin for those valuable big output tubes.

FWIW I like KT88s, used a lot of JJ in past years. I know tubes aren't "supposed to" have tone, but I find there's often a fat plummy tone that shows up with KT88 in use, a bit of that classic FM radio announcer's voice. Fryette/VHT amps seem to have this built-in and I'm sure KT88's they use help.

[TomCarlos]

Thanks guys..... at this point, I see no reason to change the Configuration.

Glebert, the Negative bias is at -57vdc. I guess there is room to fiddle a bit. But when you are in Channel B, the Distortion channel, I don't believe changing the bias will make that big of an impact. I was more concerned with "does anyone see any issues with the hybrid approach" and it looks like you gents answered that.

By the way, as I am working on the Marshall Series 2000, I sent an email to Marshall Amps in the UK. I did receive a response. But as you can imagine, there is very little in the way of records or Service Bulletins on this amp. If you look at Revision 3, they definitely made some changed. But for this amp, I will leave it as-is.

Again, my appreciation.

[Helmholtz]

Glebert, the Negative bias is at -57vdc.

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

[TomCarlos]

Got it... thank you Helmholtz and Glebert.

[Helmholtz]

As you asked:
(Partial) cathode bias works by local NFB and has some stabilizing effect on the circuit.

[mhuss]

The cathode resistors will also help even out the current sharing between the tubes if they're not well matched.

Genalex had specific cooling requirements for KT88s, I don't know if they still apply to those made today. They recommended spacing of at least 4 inches center to center, and that the sockets have the 2&6 pins lined up along the inter-tube ("valve") axis.

[TomCarlos]

The cathode resistors will also help even out the current sharing between the tubes if they're not well matched.

Genalex had specific cooling requirements for KT88s, I don't know if they still apply to those made today. They recommended spacing of at least 4 inches center to center, and that the sockets have the 2&6 pins lined up along the inter-tube ("valve") axis.

Thank you. Interesting point you have about the tube installation recommendations and it cause be cause to see if I could find the Genalex doc that makes that recommendation. I read other threads that mention horizontal mounting and space between tubes due to heat.

UPDATE - This amp's circuit (Revision 2) was designed for 6550 tubes. With the KT88 tubes installed, the Center-to-Center spacing is approx 2.5" and space between tubes is approx 1/2" or less- not good if these tubes are running hot. I guess someone decided to make this change (based on the argument that KT88s might sound better than 6550 tubes) but did not consider the heat factor. So let's leave the "sounds better argument to posts that can be found all over the internet."

UPDATE- This amp was originally designed for KT77 Tubes !!

And as a weekend warrior hobbyist technician, I frown on swapping out tube types and making circuit changes. If you take away the cathode resistors shown in the posted image, this amp is a 6550 design. So this is something I will be discussing with my friend as the amp I am working on belongs to him. If he has a set of 6550s, maybe we revert back to that, remove the cathode resistors, and I reset the bias.

[glebert]

Before I took out the cathode resistors I would probably try jumping or tacking power resistors in parallel. By lowering the cathode resistance in a few steps you can see if the idle current can be set to a reasonable point without the cathode bias contribution.

I know datasheet values are often conservative, but 6550 max screen voltage rating is considerably less than KT88s, by at least 100V. In this amp it seems like KT88s would probably be "in spec" and the 6550s "out of spec" in terms of that parameter.

[TomCarlos]

Before I took out the cathode resistors I would probably try jumping or tacking power resistors in parallel. By lowering the cathode resistance in a few steps you can see if the idle current can be set to a reasonable point without the cathode bias contribution.

I know datasheet values are often conservative, but 6550 max screen voltage rating is considerably less than KT88s, by at least 100V. In this amp it seems like KT88s would probably be "in spec" and the 6550s "out of spec" in terms of that parameter.

And perhaps that is why the Output Section was revised as shown in Revision 3, June 1986. I believe that by this time, the amp was out of production.

Now then.... Revision 1 WAS designed for KT77s - I just found an image of that schematic on the net and am posting here.

Revision 2 switched to 6550 tubes but the resistors attached to the power tubes remain the same value.

Revision 3 changes... Based on 6550s

-Newly added 470 ohm, 20 watt resistor after the choke and before Screen resistors
-Newly added plate resistor, 10 ohm, 5 watt
-Changed value of Grid resistors, from 1.5K to 47k.

One thing to note here.... I and others have found errors in the Marshall 2000 Schematics. So what you see in Red on Revision 3 is just one such instance.



And another thing... You will NOT find "JCM 800" or "Lead Amp" stamped anywhere on this amp. It is the "Marshall Series 2000" !!!

Thanks!!!

[Tom Phillips]

... Genalex had specific cooling requirements for KT88s, I don't know if they still apply to those made today. They recommended spacing of at least 4 inches center to center, and that the sockets have the 2&6 pins lined up along the inter-tube ("valve") axis.

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.