Traynor Biasing Questions and Pulling 2 tubes

[bcmatt]

My YGL-3a has a discrepancy between the schematic glued under its lid and the actual resister that you are supposed to measure an 8VDC drop across in order to bias the amp. Does it matter that the actual resister is 240R when the schematic says it should be 470R? Wouldn't that majorly affect how cold or hot you end up biasing the amp? I happened to notice a YBA schematic (that also uses a quad of 6CA7s) asks for an 8V drop as well, but it's resister is 1K. So maybe it doesn't matter? Or perhaps the smaller this resister is, the colder it causes the amp to be biased... making for longer lasting tube life as newer and newer designs were used?

YGLPOWER.jpg.jpg

YGL Tube Sockets.jpg

So, really my question would be, if I wanted to pull two of these Mullards to save for later or use in a different amp, which two would I pull? Would it be the ones with or without the 22R Resisters on pin 4? How would this affect the biasing method as well?
Should I be switching to installing some 1R Resisters and employing a different biasing method?

[thetragichero]

why not pull all four and throw new production 6ca7 in there? if they're all good and matched a quad of mullards is about 4 times as valuable as a pair
or play em like they are and enjoy it

your picture shows two 240r resistors so i'm guessing that the 470r was split?

[martin manning]

So, really my question would be, if I wanted to pull two of these Mullards to save for later or use in a different amp, which two would I pull? Would it be the ones with or without the 22R Resisters on pin 4? How would this affect the biasing method as well?
Should I be switching to installing some 1R Resisters and employing a different biasing method?

I'd pull the ones without the 22R since those screen stoppers would help prevent oscillation. I would also switch to 1R's on each cathode to set idle bias since you can monitor each tube. Setting idle bias using total screen current is unusual, and by setting a voltage across a single shared screen supply resistor you don't have any idea what the individual tube currents are, or even how well the current is balanced side-to-side.

[bcmatt]

why not pull all four and throw new production 6ca7 in there? if they're all good and matched a quad of mullards is about 4 times as valuable as a pair
or play em like they are and enjoy it

your picture shows two 240r resistors so i'm guessing that the 470r was split?

Actually, the other 240r resister is R54 which is also supposed to be 470R according to the schematic.
It turns out out don't have any decently matched pairs of 6Ca7s or even EL34s available. I thought I had an ok collection, but discovered none are very decently matched except the pair currently in my Express. It would seem now is not the time to try to buy new ones... everything is unavailable. I guess I have a quad of Kt-88s that might be matched... I may try switching to those...

[bcmatt]

So, really my question would be, if I wanted to pull two of these Mullards to save for later or use in a different amp, which two would I pull? Would it be the ones with or without the 22R Resisters on pin 4? How would this affect the biasing method as well?
Should I be switching to installing some 1R Resisters and employing a different biasing method?

I'd pull the ones without the 22R since those screen stoppers would help prevent oscillation. I would also switch to 1R's on each cathode to set idle bias since you can monitor each tube. Setting idle bias using total screen current is unusual, and by setting a voltage across a single shared screen supply resistor you don't have any idea what the individual tube currents are, or even how well the current is balanced side-to-side.

Thanks Martin! I will try this. I will feel better being able to monitor each tube. I do remember trying out a couple if these Mullards in my Express a few years ago and was schocked how perfectly matched the random couple of them I grabbed were... They've probably only been sitting in this Traynor for about 46 years.

I'm actually contemplating installing VVR in this amp as well just really look after these tubes. While my Express seems to handle the VVR just fine with a heatsink on the mosfet, my JCM800 2204 fried the mosfet with too much power. I'm thinking that even with only 2 tubes, this Traynor might also be a bit much, and am trying to figure out if there is a two-parallel mosfets alteration I could do.

[martin manning]

...I have a quad of Kt-88s that might be matched... I may try switching to those...

A pair of KT88 could really sub for the quad of 6CA7, but you would want either a higher primary impedance or a lot of voltage sag to keep the plate dissipation under control. With the existing primary impedance, a quad of KT88 would like to have screen voltage well below the plate (~150V) to place the load line correctly, so a straight swap is probably not a good idea. You could use a MOSFET to bring the screen voltage down, though. This would be a VVR for just the screen voltage, still feeding the screens through a shared screen resistor. I agree with you on the limits of the Hall VVR's power handling capability. It's claimed that it will work for amps up to 50W, but I've heard of many failures too.

[bcmatt]

...I have a quad of Kt-88s that might be matched... I may try switching to those...

A pair of KT88 could really sub for the quad of 6CA7, but you would want either a higher primary impedance or a lot of voltage sag to keep the plate dissipation under control. With the existing primary impedance, a quad of KT88 would like to have screen voltage well below the plate (~150V) to place the load line correctly, so a straight swap is probably not a good idea. You could use a MOSFET to bring the screen voltage down, though. This would be a VVR for just the screen voltage, still feeding the screens through a shared screen resistor. I agree with you on the limits of the Hall VVR's power handling capability. It's claimed that it will work for amps up to 50W, but I've heard of many failures too.

How much higher impedance are we talking? When pulling two 6Ca7s, I was planning to use with an 8ohm load (it only has a 4 Ohm tap). From what I can find online I think the OT has 2K primaries. Would two KT-88s justify using a 16 ohm cab? (I recently rewired my 4 ohm cabs to be 16 ohm cabs.) Either way, should the screen voltage be lowered? Or is that just a situation when the primary impedance is on the lower end?
If I did decide to install a mosfet to somehow reduce screen voltage, how would that affect the amp using regular EL34s?

[martin manning]

Doubling the load to 4k by miss-matching would work for 2x EL34 or 2x KT88, but with the KT88's you need to lower the screen voltage. If you had a MOSFET VVR on the screen, you would just turn it all the way up or bypass it to use EL34's.

[bcmatt]

Doubling the load to 4k by miss-matching would work for 2x EL34 or 2x KT88, but with the KT88's you need to lower the screen voltage. If you had a MOSFET VVR on the screen, you would just turn it all the way up to use EL34's.

Ok, gotcha! I'll think this through. I guess the ideal would be to use a trimpot internally for this so it doesn't get messed with accidentally. What kind of power handling are we talking about for this Mosfet (IRFPE50PB)on the Screen voltage? Is this going to produce significantly less heat than VVR on plate would cause?
I guess this sort of VVR would require minimal parts. No need for the extra resisters, just a pot, mosfet, zener and a 10R 5W resister?

[martin manning]

I drew this up a year ago for a KT88/6550 screen supply. I'd use fixed resistors instead of a trimmer, and a big MOSFET since screens can draw significant current. I'm showing another part number, but it's basically the same as the one you have for the plate supply in your recently restored VVR. This circuit could be built on an aluminum heat sink or on the chassis, and switched in for KT88 and out for EL34 between the screen node and the shared screen resistor, which I would keep to get some additional screen sag (reset the bias when changing tube type, of course). The ~2/3 voltage ratio should be good for running either 2 or 4 KT88 with the appropriate primary impedance. With the 430V B+ and reduced Vg2, the KT88's should last a long time. This would be an interesting experiment.

PS looks like in your amp you have 2x 240R @ 10W in series feeding the screens.

[bcmatt]

I drew this up a year ago for a KT88/6550 screen supply. I'd use fixed resistors instead of a trimmer, and a big MOSFET since screens can draw significant current. I'm showing another part number, but it's basically the same as the one you have for the plate supply in your recently restored VVR. This circuit could be built on an aluminum heat sink or on the chassis, and switched in for KT88 and out for EL34 between the screen node and the shared screen resistor, which I would keep to get some additional screen sag (reset the bias when changing tube type, of course). The ~2/3 voltage ratio should be good for running either 2 or 4 KT88 with the appropriate primary impedance. This would be an interesting experiment. With the 430V B+ and reduced Vg2, the KT88's should last a long time.

PS looks like in your amp you have 2x 240R @ 10W in series feeding the screens.

Cool, thanks!. That looks pretty much along what I was imagining... but would be afraid to just guess at on my own. Not that I feel like I really understand how it works aside from "VVR is magic!". I specifically don't understand how the zener plays into this circuit either.

I'm not sure what you call that resister, but 1 of those 240R 10W resisters is the resister between B+1 and B+2 (also 470R on the schematic) and the second one is the screen resister we have been talking about (That is also 470R on the schematic).
Would you suggest that I ought to increase either of those? Would going with a bigger value on the one closer to the screens help increase the sag?
Also, I'm assuming this VVR would be inserted between these two resisters (with the second big filter cap being before the VVR as well)?

I'm going to look at where I might fit this in physically, but I'm liking the idea of potentially putting this on a toggle switch. Would that potentially cause any issues (danger to semiconductors)? If a switch could handle it, would it be better to just pull the whole VVR in and out, or toggle between two sets of resister values 330K/1R and 680k/1M? A KT88/EL34 switch might look fancy... but really not necessary since the amp still needs to be opened up to rebias anyways.

[martin manning]

That's what I was thinking, a DPDT that connects the screen supply to the tube bases as they are now, or routes it through the MOSFET reducer. I'd leave the dropping resistors as they are, and just insert the reducer in place of the wire that goes to the screen pins. That Zener conducts when the combined drop across the 10R plus the FET Vgs reaches its Vz. At that point it starts overriding the voltage signal coming through the 100k.

[bcmatt]

Sorry, I realize I keep referring to parts of the schematic that I didn't include in the photo earlier:

YGLPOWER.jpg.jpg

So I don't seem crazy right now, I should note I replaced the image in my first post with this one as well.

[GerryJ]

A bit off topic, but my late 60s-early 70s Traynor amps are 'rated at 115v. Most modern AC current is ~120v. I know some people use devices that are simplified variacs (ie, they go down on voltage, but not up) reputedly for tone, but also to preserve tube life & transformers in...really old tweed era amps, dialing down to 110-115 volts, the original voltage.
Aside from tone stuff, is this necessary for longevity & also tube life for the old Traynors? Their huge transformers & general overbuilt quality makes me think not. Also, I don't dime or attenuate these amps. Thanks!

[martin manning]

I wouldn't worry too much about 120 vs 115, that's less than 5%. Some places in the US see 125 or more. Not a bad idea to bias on the cool side, if you expect to encounter higher voltage in the future.