6550 or kt88 screen voltage

[imaradiostar]

I'm working on a build using a pair of 6550's and a Fender Twin Reverb power TF. It seems most designs using 6550's are either ultralinear or use some sort of system to drop the screen voltage to around 300 volts.

If I run the screens up around the plate voltage (say 400-450 volts) will it be harder on the tubes than the normal lower voltage? I assume it'll take more grid 1 bias voltage to keep them turned off. Will this result in the need for a lower z bias supply and more grid drive or will it result in higher transconductance and gain. It's tough to tell from the datasheet I was looking at last night but I assume it'll take less grid drive but more bias voltage.

What about running the screens from a choke and a seperate pair of diodes? It seems like this would be simpler than a regulated screen supply but I don't know if it would work as well in the real world.

I once read that if you're going to regulate the screens you should regulate the bias supply...any truth to that in low fi guitar amp land?

jamie

[Merlinb]

Originally Posted by imaradiostar

If I run the screens up around the plate voltage (say 400-450 volts) will it be harder on the tubes than the normal lower voltage?

Yes, absolutely. Many people have problems when they try and treat a KT88 like any old EL34/6L6GC. If you want to run them with a high screen voltage, you either need to use a very low OT impedance (around 2k to 2.3ka-a say) which forces you close to Class B, or you need to use a really big screen stopper resistor (like 10k or maybe more) to stop the screen melting, which can give a squishy sound that most people dislike.

Running the screen at ~300V only requires one extra resistor to the 'normal' topology, and will save you a heap of cash in tubes in the long run!

[loudthud]

With a B+ limited to 450V, 6550's will run fine with screen stoppers of 1K to 1.5K. For an OT, 3.5K is about right. Sort of a 70's Marshall MK-II setup. 6550's really need more B+ to show their advantage over 6L6s.

[mooreamps]

Well, what is harder on the tubes is the higher plate currents, hence the higher plate dissipations. The higher screen volts allows for higher plate current, given at the same level of bias.

I have an amp that runs a single KT-88. The plate volts is around 600 V ; but the screen runs only at 300 V. I do this with a split power supply. Now, in your case, since screen current is low, you might be able to get away with a voltage divider type network to drop the screen volts down a little.

btw : I also run an EL-34 in this same amp, and I run the screen on it no higher than 300V as well.

-g

[imaradiostar]

The output tf's I'm planning to use come out around 4k:8,4&2 ohms and are considerably larger than a blackface twin output- hoping they'll be a good match for a pair of 6550 type tubes. I'd imagine that these outputs are less than ideal as plate voltage gets closer to 600 volts- maybe something like 5k to 6k6 would be better.

I'm trying to do something a little different from normal. I might instead use the twin power tf and one of these output tf's with a quad of eh 7591a's. I really like the way they sound in another one of my builds. While 8k is probably on the high side for fixed bias it might work well with cathode bias. I like that they require very little drive and this can make for a cleaner design with less components.

I thought about using an additional small PT to get the right voltages to run the screens, bias and plates at ideal voltages. I would probably use this idea to bump up the plate voltage a bit.

jt

[imaradiostar]

I'm replying to my own thread like I need the ego boost or something.

I used an antek inc power transformer on a previous build and thought it worked really well. The 4t400 has two 400 volt secondaries which should give me around 550 volts for the plate supply. If I use a small swinging choke and a second diode bridge I'll have around 350 volts for the screens and preamp. I'll probably use a cap type bias supply like an ampeg. The antek power tf is pretty small and light and is rated to put out more than enough power to make 100 plus watts with a pair of KT90's. Four tubes probably would make double that without too much trouble using the same PT. It's only $50 so it'll be another great experiment.

Anyone try the EH KT90's yet? I think it'll be neat to build a guitar amp with them. I'm going to try an "overly pentode" sounding amp with a 6au6 out front and a pentode as part of the phase inverter. I wasn't planning to use any negative feedback but the pentode into a concertina splitter should allow for easy NFB injection at the cathode in case I change my mind. I have a hunch that my cousin (a studio engineer) would love an amp like this so I'm going to build it as a gift for him. There are some designs guys have shared here that lead me to believe a design like this should work well and be very pedal friendly, something that seems to come in handy in his work.

jamie

[defaced]

We should compare notes when we finish. I'll be using the 4n400 for a 4x KT-88 build over the next month or two I don't plan to drop the screen supply though because the JJ screens are rated at 600v. I've run them at 500v and didn't have any problems, though I also wasn't looking for them.

[Steve Conner]

I understood the situation to be as follows:

KT88s were originally rated for ultralinear and triode use, so the screens can stand quite a high voltage.

6550s can't stand as much screen voltage, but they have more "inner mu", so they don't need as much screen voltage to generate the same power as KT88s. This makes them more suited to tetrode operation at high power.

Nowadays the line between the two tube types is pretty blurred, it seems that the Russian tube makers use the same guts for both. I've used both types at 475V with a 1k shared screen resistor and a 6.6k OT, and got about 60-70W.