Hot Cat 30 out of higher voltage EL34 amp

[bcmatt]

google matches that heatsink part number to a Dell W4254 heatsink assembly for the OptiPlex GX280.

the problem with Dell CPU coolers is that they often used model-specific ductwork instead of generic clamp-on fans like you might find on a generic CoolerMaster.

That said, the square holes in the top are where they often mount the custom shrouds (tabs on the shroud fit into the square holes), or where spring-type retaining clips would grab-on to hold a clamp-on fan onto the heatsink body.

I'd bet that you could fabricate some kind of clamp to hold a fan directly against the fin assembly if it turns out that you should need it.

Just a thought.

I see what you mean. Ok good to know. I'll see what happens with this setup, and go from there, knowing that I may need to mount a fan directly too this thing. I do think I have some DC fans in my bin as well.
The current old fan is AC. It really helps you know the amp is on, but not accordingly loud or anything.

[maxkracht]

I would hope the massive heatsync alone would be more than adequate cooling, assuming it is properly coupled to the transistor. Never tried, but i'm guessing you could drill and tap the copper block fairly easily for transistor mounting.

[solderhead]

Geeking out on computers for a minute: Most generic PC use separate fans for CPU cooling and case exhaust. Dell increased cooling efficiency and reduced cost by using a custom plastic duct so that the CPU fan would take in ambient case air for CPU cooling and directly vent the exhaust out of the case. One fan did two jobs (case ventilation and CPU cooling) instead of mixing retained CPU heat/exhaust with the rest of the air in the box and relying upon a separate case fan to purge it. The custom duct work made these Dell cooling parts pretty much incompatible with any other PC. They started doing this custom duct stuff in the P2/P3 era. That heatsink came from a P4 that used a custom heatsink and blower assembly.
https://dl.dell.com/Manuals/all-product ... _en-us.pdf
P4 cooling systems have a minimum Thermal Design Profile of 50W. That's huge compared to his needs for this amp, if we assume that a really good thermal junction exists between the planar surface of the heatsink and the backside of a mosfet, which isn't always flat.


How much voltage is the mosfet dropping? How much wattage is going to be dissipated in the mosfet under max load? The OP hasn't posted those numbers, but the total wattage can't be all that much, and the heatsink he's using looks like it will amount to a maximum overkill solution. He's looking at a heatsink that has a 50W TDP under ideal conditions. But ...

That TDP rating for the heatsink is based on Dell's use of custom ductwork to direct flow across the fins of the heatsink, to prevent stagnation of air between the fins. So it isn't really clear how efficiently the heatsink will work without ducted air, but I'm guessing that he's got so much surface area available that his fan will be adequate. If not, then adding some DIY ductwork or directly mounting a fan onto the fins could be something to consider. One also has to consider how effective of a thermal junction exists between the back side of the mosfet case and the CPU heatsink.

This is such an off the wall solution that it's interesting. It would be interesting to know how much power he needs to dissipate, and to take heat measurements with the amp dimed into a resistive load to see how efficient this approach turns out to be. Since we're geeking out on this, bcmatt do you have an IR thermometer?

[bcmatt]

Sorry, I don't have an IR meter.
The truth is, I don't know exactly how much power this MOSFET may have to deal with.
I know the Hot Cat is about a 30 watt amp but sees about 380V on the. EL34 plates.

The last measurements I took of this amp with a pair of fixed bias KT-88s was about 480V on the plates. I'm hoping and expecting that to go down a bit by cathode biasing. Drop a bit more with a couple thermistors on the PT primaries. Drop a bit more with my sag resister before the VVR. I'll be playing a bit with cathode resister value as well, but I guess I should expect it may be more like a 45 watt amp but I really don't know yet.

I know that the MOSFET has to dissipate the most heat when the VVR is set at 50% so I usually avoid that if possible. However, realistically, on most my amps the sweet spot for volume in most rooms is 9 o'clock to 11 o'clock on the VVR pot with the knob all the way down being 10%. So realistically, I may be riding around 40% most of the time, which is close to working the MOSFET the hardest. So I feel like I need to heatsink it to the maximum potential that it may need. But not sure how to measure that beyond putting my hand on the heatsink to check it's temperature after some riffing.

If an amp is biased close to class A, does that mean the MOSFET is working hard regardless of what is being played on the guitar?

[bcmatt]

I cleared out the chassis and decided to swap the OT with the spare I have that is actually the same size but with 4, 8, and 16 ohm outputs for 4.2K primaries. I figure it should be fine replacing the Hammond of same size and weight that was made to be for 4xEL34s 2k primaries at 4 ohm only.... So with just a pair it was going to be 8 ohms at 4k. A little more versatility in cabs would be nice.
I'm still trying to decide where I'll put the filter caps... Maybe on the fibre board since it is so long. I may move The PI tube socket down closer to the preamp.

20250329_115617.jpg

[bcmatt]

It's really slow going, but making some progress.

20250405_201400.jpg

I got distracted by turning my 5E3 mostly into a 6G3

[bcmatt]

I wired mostly from the inputs and preamps up till the PI. I'm still trying to decide which slot I'll put the the PI socket in. I think I'll leave at least one tube socket for a potential FX loop or Reverb in the future if I feel so inspired.
I'm going to start wiring from the other end of the board with filter caps now to help decide where I want the PI to line up.

20250406_121842.jpg

[bcmatt]

I've updated the plan slightly:

Hot Cat Adapted from YGL Filters.png

My 40Uf Filter Caps are rather large for the board, so I may risk it with some little radials I bought doubled up.

20250406_192134.jpg

[bcmatt]

Got a bit more wired.
Larger dropping resister and screen resisters, VVR, Cathode Bias circuit to test, etc. I'm now trying to decide how to lay out the rest of the filter caps:

20250407_122430.jpg

Also, I don't seem to have 10K 2Watt Resisters. I'm would think those Grid resisters on the Power Tubes should be fine with 1 Watt Resisters...

[bcmatt]

I think I just need the protection diodes on the power tube plates, then I can fire it up and start testing voltages.

20250407_175439.jpg

[bcmatt]

Well, some success and some failure last night...
I chased my tail for quite a while due to the B+ output from the VVR being connected on the wrong end of the Zener. It was only giving me a range of 5-30 volts output. After dealing out both the zener and the MOSFET I realized the issue. Bypassing the VVR was giving too high of plate voltages all along the preamp (over 300V). Also the amp was biased too hot at full voltage with the 470 ohm cathode resisters (somewhere around 120%).
For now, voltages are alot closer to reasonable with the VVR knob set to about 3 o'clock. Although most my amps I run the VVR around 1/3.

20250408_065546.jpg

Clean channel worked...
Although, I see I forgot to tie together pins 1 and 8 on the EL34s. Just fixed that.

Anyways the main problem is the gain channel isn't giving me anything... Especially if the clean channel volume is down.
I rewired the inputs with cliff jacks... No dice.
I also switched the cathode resister/cap grounding for V1a to reflect the ceriatone layout (through the input jacks).

My suspicions today are maybe the 4m7 resister I subbed for the 5m1 (don't have one) isn't high enough to bias that triode properly or VVR doesn't allow that style of biasing to work at all. What's it called? Grid leak bias?

[bcmatt]

If that style of V1A biasing is not going to work with VVR, does anyone know where this alternate schematic came from:

Hotcat preamp.JPG

[dcribbs1412]

Link to a hot cat style build I did years ago.
May have info that might help.
Its a great circuit, nice cleans and monster gain. Recommend checking out the stray cat layout from ceriatone for info also
https://ampgarage.com/forum/viewtopic.php?f=6&t=15875

D

[bcmatt]

Link to a hot cat style build I did years ago.
May have info that might help.
Its a great circuit, nice cleans and monster gain. Recommend checking out the stray cat layout from ceriatone for info also
https://ampgarage.com/forum/viewtopic.php?f=6&t=15875

D

Thanks for responding. I did scour that thread a bit last night when I was trying to resolve my issues, but I'll take a closer look again.
Yes, the ceriatone Stray Cat layout was what I relied on pretty heavily as a trusted source that ought to work. I tended to trust that layout over the actual Hot Cat Schematic that has been online forever. I get the feeling that the Stray Cat was originally made from that old schematic but has been tweaked over the years to resolve mistakes and issues.

[bcmatt]

I added a 1M resistor in series with my 4M7 Grid leak bias resister in case it was just not enough... but that didn't fix it. But then I realized that my gain pot was not grounded. That's probably the issue... we shall see.

On another note, I am running tests with the VVR voltage set to give me close to schematic voltages at somewhere around 85% on the pot knob.
The weirdest thing that I am trying to understand when checking the cathode bias on these EL34s:
With 370 Volts on the Plates
Using 470 ohm Bias Resistors (even measured them to confirm while in the circuit)

I'm getting 20 Volts on one cathode but 40 Volts on the other. So with everything supposedly the same, one tube is biasing at 56% and the other is at 106%. I pulled these tubes from my Express where they were pretty well matched last I checked...
Could this be that one of my Cathode bypass caps is faulty, or do you think I fried one of the tubes?
I tried another set of EL34s and it is doing the same thing. I'll check those bypass caps.