Traynor YVM-1 Plate Voltage Sag Issue

[soulsonic]

I haven't posted here in forever - but I also haven't seriously tinkered with tube amps in forever... and now here I am fooling around with amps again.

I bought a Traynor YVM-1 "Voice Master" a couple years back for dirt cheap from a local music shop in upstate NY (I don't live there anymore, in northern WV now, same difference, honestly).
The amp "worked" when I bought it. Plugged it in at the store and strummed a few chords - was functional and loud, so it was well worth what they were asking

Things I notice right off the bat:
This amp suffered a catastrophic failure at some point in its past. The fibreboard directly under the large shared screen grid resistor is burned all the way through with a gaping hole. The resistor in its place obviously doesn't match to the style Traynor used, so I can assume it's a replacement for one that burned up. It is a 1K, and at least 10-15w in size. One of the leads coming from the output transformer has been cut short with a new wire spliced on to make the connection to the tube socket. I'm assuming this wire must have been close to the resistor that got burned and it's insulation must have been damaged - whatever the case, that wire got damaged badly enough they chopped it off and replaced. Since the board was ruined in the spot where the screen resistor had been, someone had added an insulated turret to mount the resistor. All the wires attached to the output tube sockets have little paper numbers taped on them for ID, so I'm assuming either the sockets have been replaced, or the person just desoldered all the wires to make the repairs after the resistor burned. The amplifier seemed to work fine in this condition.

Here's the problem. I had it set up and biased happily with a pair of KT66 that I thought sounded great with it. I was running the amp hard into a reactive load DI box for headphone jamming when I noticed the sound started sagging and dipping down really bad. I hadn't noticed this happening before - it had been a couple years since I started playing with mods and it had been working well up to this point. The switch from the large shared screen resistor to the two individual screen resistors is the most recent change, not long before this happened/ I looked around back and saw the tubes flashing brightly while playing. I stopped playing, but it flashed a few more times and then the fuse blew. Assessing the situation afterwards, I found one of the tubes had blown - real bummer, these tubes weren't old and I had gotten a sweet deal on them from Angela when they were selling them for less than half of what they are now (and that was only 2 years ago!). I couldn't find any other damage in the amp or any obvious reason why it would start flashing so vigorously inside the tubes like that.

I stuck a used set of EL34 in there and rebiased. I had no problem getting the tubes set to 70% dissipation and everything is happy at idle. I have 454v on the plates and 450 on the screens. Running a clean sine wave through it directly into the PI, I was able to get 30w output before clipping, with no obvious crossover distortion. As the power increased, the B+ voltage would sag. As it would begin to clip, it would get flyback voltage "shoulder spikes" and the plate voltage began to sag considerably. At 50w of output, the wave was a very squared trapezoid with prominent shoulder spikes. At this point, the plate voltage had sagged to around 398v-400v, and the screen voltage was around 422v. I know the screen having a higher voltage is a huge warning sign, and the screens were drawing 35mA each at this point - a serious malfunctioning condition that can kill the tube. I think this is the heart of the problem and what killed my dear KT66.


My question here is: why is the plate voltage sagging lower than the screen grids?

[soulsonic]

Inspecting the amp, I see the first couple banks of filter capacitors are only rated for 450v. These are old Mallory multicap cans - two of them with two 40uF sections, rated at 450 volts. The first bank immediately after the rectifier is one can with the two sections together in parallel for a theoretical 80uF, at 450v. Remember I said this amp at 454v on the plates at idle. I think this may be the issue. I think that 50+ years of this amp running that cap at its absolute limit has caused it to become faulty. i measured the caps with a Peak ESR60 meter and it gave results that left me still questioning their condition. I measured both multicap cans (they are identical). The first one coming from the rectifier gave on-spec measurements of approximately 40-50uF for capacitance, but the ESR seemed unusually low for such an old-tech cap - around 0.42 Ohm per section. This would seem like a good low number for a modern cap, but it seems highly suspicious to see ESR that low on a 50 year old cap. On the second multicap, I got again very low 0.40-ish for one section, but the second section gave an ESR of around 2 Ohm?! Incidentally, the section that had the highest ESR also is the one that was subjected to the lowest voltage over the years, so I can't help but think this is closest to the original spec. In comparison, a  newer 10-year-old 100uF 450v cap gave me an ESR reading of around 0.4 Ohm.
 
Has anyone here experienced plate voltage sagging because the filter caps were being pushed past their rated voltage? Or is this typically caused by another problem not related to the caps? It seems like flyback voltages are an issue here. Would installing flyback catch "tube saver" diodes help here?
Should I replace the output tube sockets? Others have suggested to me the flashing from flyback may have caused an arch on the socket which may be causing the plate voltage sag from something shorting along the arch path. I don't see any obvious arch damage on the sockets, and I have cleaned the thoroughly, but I know sometimes this is impossible to see.
Would replacing those multicap cans likely solve the problem?  What other things may be the cause of the plate voltage dipping below the screen voltage?

[AlNewman]

http://www.valvewizard.co.uk/pentode.html

That's an interesting read.  Especially the screen grid part.

[tubeswell]

If you’ve had significant fly back spikes, and fly back spikes are now routinely in the signal under heavy signal conditions, the winding and lamination insulation in the OT may already be compromised. If you disconnect the OT and run 5VAC across the secondary, what VAC are you getting on the primary? (You probably want galvanic isolation on your VAC source for this test - I.e. don’t just run a variac from a mains point).

[PRR]

> I know the screen having a higher voltage is a huge warning sign

That is Tetrodes, true Tetrodes. Not pentodes. Not "beam power tetrodes" (which is just an evasion of a patent on power pentodes).

If you see how an amp works, when playing LOUD, the plates are below the screens much of the time, but up/down too fast to see on a meter.

454V on a 450V cap is not instant death. But 50 years might do it.

You can, like I did on my 2002 Honda, keep beating it until parts fall off. Patch-up, repeat, repeat.

[soulsonic]

My goal is for the amp to be reliable and not toast my tubes anymore. I'm willing to do whatever is necessary to achieve that goal, even if it means I need to replace the OT. I bought this amp so cheaply that I'm still coming out ahead even if I have to drop $150 on a good OT. Would be more than worth it if it saves me $$$ in tubes no longer being damaged.
I will see if I can scare up a 5v transformer, or at least something close, to send a voltage thru the OT.
Plans for now:
Test the OT for damage - replace if necessary
Replace charred circuit board with Hoffman-style turret board (still have parts on-hand that I ordered probably 12 years ago)
Replace the main filter caps with ones rated for higher voltage
Possibly replace output tube sockets


Beside testing the OT, can anyone recommend anything else I should be looking at?

[pdf64]

I don’t see any reason to think the OT is damaged.
Have you got a schematic, both original and your amendments? This one doesn’t seem to align with your description of the original design https://el34world.com/charts/Schematics/files/Traynor/Traynor_voicemaster_yvm1.pdf

People using bias calculator web pages tend to aim to idle KT66 at 21W, which seems too high for a valve type that’s roughly equivalent to a 5881 / 6L6WGB.

It’s completely normal for the voltage of unregulated power supplies to sag as current draw increases, as with an AB output stage.

Please note the enormous voltage spikes and don’t attempt to probe the output anodes of an overdriven output stage.

For our purposes, 454V = 450V.
The difference is only apparent now because we’ve got high resolution digital meters.
Keep in mind that valve amps and their components were designed in the days of analog meters and slide rules.

What V AC do you measure on your mains and heater circuit?

[soulsonic]

Yes, it isn't like the schematic in Hoffman's library. It's much closer to the version attached here. But it wasn't exactly like that one either. It seemed like a mashup of the two. It has the inductor and 4/8/16 ohm output taps, like shown in the schematic hosted here, but it has the Mixer Out/Power Amp In jacks as shown in the attached schematic, and it had EL34s in it when I bought it. It also has the power supply with full wave bridge and standby lifts the negative end. It has a 100K feedback resistor that appears stock - neither schematic shows that. Most other values match to the schematic attached here.
I measured 6.5vac on the filaments.

[pdf64]

It would be really beneficial to have a schematic we can base suggestions on. Of the power amp at least.
Would it be feasible to amend the schematic from the previous post?
Even just by hand, take a photo of it, and upload to imgbb etc?

[soulsonic]

This is what I got

[PRR]

> It has a 100K feedback resistor that appears stock - neither schematic shows that.

The other plans DO have feedback resistors. And both/all resistances matter.

The one with the coffee stain is 47k:4700. The one PDF64 linked is 22k:470. Now you show 100k:4k7. So: 10:1, 47:1, 21:1. These are the kind of gain/cleanliness tweaks (Pete was a tweaker) you do after you use the rig a while (or a major buyer asks for changes). Or the available values for the Presence pot.

The several rectifier ploys just go with available (this was not a high-production product) transformers or the price of rectifiers. 6L6 vs EL84 is fairly moot in the aesthetics of 1970 (no cork-sniffers yet).

[pdf64]

… I had it set up and biased happily with a pair of KT66 that I thought sounded great with it. I was running the amp hard into a reactive load DI box for headphone jamming when I noticed the sound started sagging and dipping down really bad. I hadn't noticed this happening before - it had been a couple years since I started playing with mods and it had been working well up to this point. The switch from the large shared screen resistor to the two individual screen resistors is the most recent change, not long before this happened/ …

A single shared 1k screen grid resistor will drop more voltage / provide more current and hence dissipation limiting than 2 individual 1k screen grid resistors.

If it had previously been coping fine under the same conditions (?), then reverting that change may be worth considering.

The 2k61 HT dropper between screen grid and LTP supply nodes seems rather low. Allows a higher voltage at the LTP and hence a larger maximum signal output from it.
That can exacerbate bias shift at the output valve control grids when overdriven, possibly leading to blocking distortion. Which kinda aligns with the description of how the amp was sounding immediately prior to the valves failing.
The HT supply to the LTP should ideally be just enough to push the output valves cleanly to maximum output, but the LTP then clips if the level is increased much higher.
That can allow a lot of power amp overdrive without the output valves being temporarily pushed too far into class B.

[soulsonic]

… I had it set up and biased happily with a pair of KT66 that I thought sounded great with it. I was running the amp hard into a reactive load DI box for headphone jamming when I noticed the sound started sagging and dipping down really bad. I hadn't noticed this happening before - it had been a couple years since I started playing with mods and it had been working well up to this point. The switch from the large shared screen resistor to the two individual screen resistors is the most recent change, not long before this happened/ …

A single shared 1k screen grid resistor will drop more voltage / provide more current and hence dissipation limiting than 2 individual 1k screen grid resistors.

If it had previously been coping fine under the same conditions (?), then reverting that change may be worth considering.

The 2k61 HT dropper between screen grid and LTP supply nodes seems rather low. Allows a higher voltage at the LTP and hence a larger maximum signal output from it.
That can exacerbate bias shift at the output valve control grids when overdriven, possibly leading to blocking distortion. Which kinda aligns with the description of how the amp was sounding immediately prior to the valves failing.
The HT supply to the LTP should ideally be just enough to push the output valves cleanly to maximum output, but the LTP then clips if the level is increased much higher.
That can allow a lot of power amp overdrive without the output valves being temporarily pushed too far into class B.

Yes, I was concerned that the change to the two individual 1K resistors may have upset something that had  been fine before. Even before I had noticed the issue of tubes flashing, it seemed like the amp didn't sound "right" compared to how it had been. Guess it's good I didn't throw that resistor away! I think this is likely the main cause of the problem.


Running a clean 400Hz sine wave directly into the input of the PI, I got 30w of output before it started clipping. This seems low, all other things considered (big transformers, etc). Is perhaps the voltage being higher than optimal in the PI causing it to clip sooner than it should, ie; a gain structure issue with too much driving the output? Or is it more likely because of other deficiencies? Or do these amps just not have alot of clean headroom? This is the only Traynor I've owned, so I don't have alot of experience with their sound. I repaired a YBA for a friend once, but that was awhile ago, and I didn't try it cranked all the way, so I'm honestly not sure how clean I should expect it to be.


The 2K61 resistor is not original. It is a replacement put in by whoever had worked on the amp previously. The original likely got damaged when the old screen resistor smoked, because it is located very near that area. But based on the stock schematics, I think this is most likely a replacement for what had been 2K7, so I don't think that is a matter of a faulty repair.