• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

6550 screen voltage

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have two Stromberg-Carlson Model AP-437 mono power amplifiers and the schematics under the hood for the two are different but they are wired the same. One schematic, the one that is how they are wired shows a 2k serial resistor to the screens and an 8k shunt resistor for a screen voltage of 375. The plates are ~520v. The other schematic shows a 470R serial to the screens and zero shunt resistor, the screens are only 5v less than the plates, which to me is too high.

Has anyone seen these amps before and why are there two variations? My thinking is maybe the 6550s were failing with such high screen voltage and so they decided to lower them via the resistive divider.

The 8k shunt resistor is above the chassis, the type that stands straight up and they both are failed. Someone else added large wirewounds that I do not know the rating but look like 20 watts and they are screaming hot and blackening stuff underneath. I currently have mounted 50w chassis mount resistors under the transformer cover using the chassis as a heatsink and it seems to be a little better. But I am curious about why there are two different versions. And maybe going with just the 470 serial screen resistor can be fine but I have little experience with the 6550 and that seems too high to me.

What do the 6550 experts think out there? What should I do? Leave the large chassis mount shunt resistors and lower screen voltage or go ahead and run the screens at 515v?
 
Administrator
Joined 2004
Paid Member
I suspect it might be the other way around as a cost saving measure, although there will be parametric changes due to the differences in screen voltage.

I designed and built a 120W UL amp for a client/friend 2 decades ago with a quartet of 6550 running at 550V and 25mA quiescent current each, those tubes were finally replaced at my insistence after 18 years of continuous service. (Svetlana 6550C - apparently I got some good ones) Note that the screens would have been a couple of volts higher than the plates except for the stopper resistors on each screen. (I no longer remember the value, I think it was 100 ohms, but I also used 330 ohms in some amplifiers)

I am not sure there are any good 6550 these days..
 
I suspect it might be the other way around as a cost saving measure, although there will be parametric changes due to the differences in screen voltage.

Oh okay so you are thinking that maybe the extra 20 watt resistor top side was failing and so they said lets just run the screens at anode voltage. That makes sense to me since both the resistors have failed. The bias voltage must be higher with the higher screen voltage which if I change it I would modify for this.

I also don't have any plate characteristic graphs with the screen at 515v, and I do not know what the load impedance the output iron is but I could easily calculate it, the former is the more difficult thing to obtain. So I don't know which operating condition the output transformer is optimal for. If your assumption is correct in that the first iteration is the one with the 8k shunt resistor then I would think the output transformer load is optimal for the lower screen voltage.
I am not sure there are any good 6550 these days..
This is what worries me. Although the amp does have tired looking Tung-Sol 6550's with metal bases in them currently they won't last forever.

What's a guy to do? :confused:
 
Administrator
Joined 2004
Paid Member
I think you'd find with a commensurate shift in bias and screen voltage that the operating conditions will be virtually the same. (This is the problem and benefit of pentodes.. LOL)

Makes sense if you are using them as a pair to make them both the same.

Reliability and cost both could have been considerations for the change.

Your best bet is probably a set of Russian made Genalex Gold Lion KT88, these are what I have been using to retube old client amps. So far so good over a couple of years use. (Get them matched and burned in, this usually weeds out the questionable ones)
 
Thanks for the tip about the Genalex Gold Lion KT88, I have heard good things about them. II was using Penta Labs KT88's a few years ago with a 6550 tube amp I had.

I already outfitted the both with the chassis mount 50 watt resistor for the lowered screens and I will see how that goes for a while. The chassis where I mounted the resistor doesn't seem to get too hot so I am happy. If there are any issues or anything I will just run the screens at 515v.
 
Shortened service life is an ongoing expense that never ends. :mad: IMO, 1 way or another, the cost "bullet" of holding g2 B+ down has to be bitten. AFAIK, the only 6550/KT88 variant with truly tough screen grids is the scarce/costly GE 6550A.

While the regulators already mentioned are not low cost, they are, in my mind's eye, clearly superior to any sort of dropping resistor arrangement.
 

PRR

Member
Joined 2003
Paid Member
You've got over 25 WATTs in that divider. Is it cold in your house?

That amp can't need over 300V on screens. Use a MOSFET to make a low-loss dropped supply. Total screen circuit dissipation will be 1W-2W.

I don't think 500V/300V in hi-fi, with reasonable idle current, is any kind of 6550 abuse. I've run them much harder no-problem. I did not have any 6550 from this decade-- but a ~2005 Sovtek appeared as robust as a Kentucky G.E. or a vintage Dyna 6550.
 
It's not cold in my house.......................now LOL.


I didn't engineer the dang things so I can't tell anyone why they chose to do it this way.

There is not much room under the hood for any boards and the chassis is shallow, I don't think it's even 2" deep.

I do agree, I shook my head a few times at the amount of energy put forth into heat trying to drop the screens. :hot:


Here is the original schematic attached inside the amp. I circled where the 10 watt 2k serial resistor and 20 watt 8k resistor are. The replacement 20 watt resistor inside the chassis was VERY hot so I replaced them with 50 watt ones inside the transformer cover. This is what i used and it's much better than the wirewound 20 watters.

EDIT: I also replaced the 2k resistor with a 20 watt ceramic so it doesn't get as hot either. I went with 3k3 instead of 2k and I changed the 8k to 8k2, I have 360v on the screen and 520 on the plate.

F1074147-01.jpg
 

Attachments

  • screens.jpg
    screens.jpg
    428.8 KB · Views: 216
Last edited:
The absolute max screen volts for 6550's is 550 Volts, pentode mode only. this means UL is definitely out for 550V supplies. Some manufacturers choose to ignore this, CJ for one. They also seem to think we have 230V mains supply, when it can get up to 253Volts. this means the unfortunate valves are seriously overdriven and flashover and consequent damage follows.
You could try KT120's which seem more tolerant in this respect.
 
Administrator
Joined 2004
Paid Member
I'd run them at 350V - performance will be just fine.

I suspect the resistor picture is just a stock photo and not what you put in your amps. (It's 8.2 ohms)

That voltage divider is consuming something like 50mA, if you have space you could install a pair of gas tubes running at 20 - 30 mA or so, greatly improve screen regulation ( = less distortion) and reduce power consumption slightly.
 
Yes there is the divider + the screen current across the 3k3 and then just the divider current through the 8k2. So about 55mA through the 3k3, and 45mA through the 8k2. I don't like it but that's how it was designed.

I like the idea of gas tubes, I just actually put some 0D3's in my 6HJ5 amp (mainly for looks) but I want to keep the looks of these original, and the circuit as original as possible for resale purposes. People that buy this stuff want it as stock as possible.

So it looks like I will keep it as is with the screens down around 350, the tubes should be happier and the performance should not suffer.
 
Fender had a FS400 amp that ran 6550 ge's at 350 volts on the screens and 700! on the plates

I run 6550s and KT88s this way all the time with no problems. You will need good quality sockets though as 700+ volts on the octals is pushing it, but the tubes hold up fine if the screens are limited to 300-350V. Where else can you get around 100 watts clean with a pair of common tubes?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.