Traynor Biasing Questions and Pulling 2 tubes

[bcmatt]

Ok, sorry about the delay in progress... things got really busy with work these last couple weeks. I finally got to spend a little bit of time with this today.

I was curious about the old biasing method.
Before pulling 2 tubes and installing the bias measuring resisters, I wanted to see how the old biasing method compared.

If I set the bias according to the schematic (even the the resister value is not right) with 8V across R54, it actually drops to about 5V across R54 when 2 power tubes are pulled out.

Then if I check the bias at that same setting (after installing a couple 1R resisters on the Cathodes, it reads about 70mV across the 1R Cathode resisters!

With a plate Voltage of 468VDC on those 2 remaining tubes, I set the bias to about 33mV across the 1R resisters (about 60%). I thought I would check the R54 voltage drop and it reads at about 2Volts across R54.
I don't care about the schematic's suggested biasing method anymore anyways, it's just confusing... especially when the schematic values are not accurate for those key resisters.

So, I am trying to figure how I might fit in the VVR stuff for the screens so that I can switch to some KT88s. I like to mount a fair size heat sink (that I collected from old computers) underneath the chassis, but things are very tight under there with all the power tube sockets, transformers, and the fair size fan. I'm trying to decide if I may just remove the two empty tube sockets and use that space... It would allow the fan to hit the heatsink well if I did. It would just be a pretty permanent commitment to this being a two-power-tube amp.

[martin manning]

What about mounting a heat sink or even just an aluminum plate inside using the choke attachment bolts and some stand-offs? I'd be reluctant to make any big changes at least until you decide that you really like the KT88 with reduced screen voltage.

The cathode resistor bias method is more straightforward IMO. I'd just go with that and forget about the screen resistor voltage drop.

[bcmatt]

What about mounting a heat sink or even just an aluminum plate inside using the choke attachment bolts and some stand-offs? I'd be reluctant to make any big changes at least until you decide that you really like the KT88 with reduced screen voltage.

It actually doesn't have a choke. Those screws are for a vintage 120V fan that blows on the power tubes...

Fan.jpg

It perhaps could be removed and the heatsink could go in its place...
If I was to just try some aluminum plate inside the chassis... what would you say would be reasonable dimensions? I found a stack of small aluminum signs in my shed that I could maybe try cutting some rectangles out of to stack inside the chassis. I'm just not sure how much mass the mosfet needs to dissipate it's heat properly.

aluminum.jpg

[martin manning]

I think a 3" x 3" piece would be ok, at least for a test.

[bcmatt]

Ooop! I just found a smaller heatsink that is a few millimetres shorter than the chassis when I stand it up on it's side:

Heatsink.jpg

I'm not sure how I would secure it... maybe silicon it down to the chassis, then mount the mosfet to it with a small bolt-hole drilled through the heatsink.
I just realized that my new mosfets didn't come with the mica shields...

[martin manning]

How about a piece of aluminum angle (a hardware store item), screwed to the heat sink and then to the chassis using the fan attachment bolts?

[bcmatt]

I finally had the silicone insulators show up, but they were for the tinier type of MOSFETs. I ordered some others yesterday, but at least they should show up this weekend and I can finally try this.
I do like the idea of an L-bracket mounting the heatsink to an existing chassis bolt. I'll see what I have.

[bcmatt]

My appropriate sized silcon for the mosfet should arrive tomorrow and I am all ready for it. I didn't have any resisters between 220K and 470K so I put in a 1M trimpot that I had on an old VVR board instead of the 2 resisters.

20220603_185244.jpg

20220603_185254.jpg

[bcmatt]

So, I got this Voltage reducer finished being installed. I squeezed a couple of the KT-88s installed as well. One is practically touching the power transformer. I also had to bend the fan bracket away from the power tubes because it was going to touch that same Kt-88. Does this look problematic?

20220604_161420.jpg

With 464V on the plates, I have 299V on the Screens... That's 64% but with the trimpot I could dial in whatever might be safest... not sure what would be better.
I also wasn't sure what sort if bias I should be setting. I dialled in 38mV which would be maybe 17 Watts per tube. Again, I'm not sure what range I ought to be going for. It seems to sound pretty good when I play... I think slightly brighter than the 6Ca7s, but plenty of power... much more than I would ever need. I'm open to experimenting within what would be considered "safe" ranges to see what gives ideal tones.

I have tonnes of tweakability drive and EQ-wise because this Amp has the HRM overdrive circuit added and the Reverb removed. So, in addition to the regular tonestack, it has and extra Drive, Treble, Middle and Bass added to the front panel for the drive circuit (which is always in).

Anyways, this is a fun project so far.

[martin manning]

This is breaking some new ground!

i agree it's nice to have the trimmer there to experiment a bit and find the best tone.As long as the screen voltage is somewhere close to 66%, say or 60% to 70%, it should be ok. The bias should be set as you normally would, at 60-65% of max plate dissipation, and you will have to readjust the bias if you change the screen voltage, of course. Right now you are pretty cool at 17W.

The clearance between the upper right tube and PT does look a bit tight. Cooling is the main concern, but you have the fan so it may be ok. Is there any other pair of sockets that would have a little more clearance, like upper left and lower right? As long as you have one tube on each side of the OT it should be fine, but I seem to recall only two sockets had a screen stopper.

[martin manning]

PS, you are miss-matching the OT/speaker to get a 4k primary, correct?

[bcmatt]

Yes, I now consider this 4 ohm amp an 8 ohm amp instead.

Yes, I was sticking to the sockets with the screen stoppers. I guess I could reconfigure the sockets to have the screen stoppers on the other sockets instead. That was going to be my plan when it looked like the KT-88s wouldn't fit... but once I took the chassis out of the cabinet, I realized they actually "just" fit. The fan seems to do a pretty good job though.
Can I just have the B+2 hit a terminal strip that then hits a screen stopper for every individual tube? Can you think of a reason why only two stoppers were used? 4 would be safer, no?

So it looks like I have now preserved my quad of Mullards.

20220604_174039.jpg

I may save them for my Express in like 10 years, when they're worth as much as small island.... or a tank of gas.

[bcmatt]

I never did post the final pic of the mosfet on the heatsink:

20220604_175948.jpg

So, how does one decide what a reasonable range of bias for a tube that can handle considerably higher plate voltages? With Plates at 464V, you'd think I could be taking advantage of the "lower" plate voltages to get a little lower volumes out of this amp.

I don't know if this makes sense, but if a KT88 can be capable of 42W and Plate Voltages up to 800V... then I treat my more modest 464V Pates like like I'm scaling the power to be like 24W tubes. Then, setting them at a 65% plate dissipation bias setting would mean I'm trying to hit about 34mA. Am I crazy? Should I expect something to totally ruin the tone?

If I just use this:
https://www.amplifiedparts.com/tech-art ... calculator
It says I ought to try and get between 45mA and 53mA... but if my plate voltage was 100V it would just want me to bias those KT88s between 210mA and 245mA in order to get that 35W per tube. Seems pretty crazy... And I don't feel a need to achieve 35W per tube unless that is the only way to get good tone.

I'm willing to experiment with different biases to hear what sounds good, but I am curious if going too low can be dangerous for something in the amp.

[martin manning]

...I was sticking to the sockets with the screen stoppers. I guess I could reconfigure the sockets to have the screen stoppers on the other sockets instead. That was going to be my plan when it looked like the KT-88s wouldn't fit... but once I took the chassis out of the cabinet, I realized they actually "just" fit. The fan seems to do a pretty good job though.
Can I just have the B+2 hit a terminal strip that then hits a screen stopper for every individual tube? Can you think of a reason why only two stoppers were used? 4 would be safer, no?

You can try the other sockets as they are, and see if there's a problem. If there is, you could add two more stoppers at the socket pins. Traynor decided they were necessary, and that one per side was enough, but that was for the 4x 6CA7 arrangement. It may be fine as is for the KT88's.

So, how does one decide what a reasonable range of bias for a tube that can handle considerably higher plate voltages? With Plates at 464V, you'd think I could be taking advantage of the "lower" plate voltages to get a little lower volumes out of this amp... If I just use this:https://www.amplifiedparts.com/tech-art ... calculatorIt says I ought to try and get between 45mA and 53mA... but if my plate voltage was 100V it would just want me to bias those KT88s between 210mA and 245mA in order to get that 35W per tube. Seems pretty crazy... And I don't feel a need to achieve 35W per tube unless that is the only way to get good tone.

The percent-of-Pa max method is a simple technique to (in most cases) avoid over dissipation and reduced output tube life. You only need to bias high enough to avoid crossover distortion (theoretically just over 50% conduction angle). If you are below that it will just sound really bad, and a bit more usually produces better tone. As you go higher you need to avoid over dissipation, and where that limit is depends upon the power supply's voltage and voltage sag characteristics. Here's a good read on biasing: https://www.aikenamps.com/the-last-word-on-biasing

I'm interested to hear more about how this amp sounds with this FET-reduced screen supply as you get it dialed in, and I'm curious about how the Zener current limiter plays into the sound. It should protect the power tube screens and the FET, and I wonder what would happen if you were to reduce it further.

[bcmatt]

I'm interested to hear more about how this amp sounds with this FET-reduced screen supply as you get it dialed in, and I'm curious about how the Zener current limiter plays into the sound. It should protect the power tube screens and the FET, and I wonder what would happen if you were to reduce it further.

Do you mean what would happen if you reduced the Zener's voltage type, or reduce the screen voltage more with the trimpot?

So, ya, I'll experiment with lowering the bias enough so that I can hear that crossover distortion, and figure out at what point it gets dialled out.... and then if I can hear any improvement as I go higher.