MOSFET B+ reducer

[pjd3]

Hello,

I'd like to check in here for a sanity check on my understanding (or mis-understanding) of how I may safely go about fastening and heatsinking my 20 volt Mosfet zener B+ reducer to my amp/chassis.

This is a 35ish watt AB763 with 6L6WGB NOS tube pair. It has a Blackface Bassman PT which will be making for a B+ of around 460vdc. I was told by Mike at KCANOS that the 6L6WGB's will be OK up to about 470vdc. Since 460vdc is getting very close to the recommended Plate V maximum, I thought it could be good to knock of around 20 volt or so to perhaps get some more life out of the pair.

I'm wondering if I was to mount the tab hole of the Mosfet at the base of my established "power star grounding point, directly to the chassis" if that would suffice as both a reasonable heatsink and also a clean electrical connection to the HT CT so to avoid any ground loop noise causing conditions. I see that in both R.G"s circuit and Aikens circuit that the Drain of the mosfet is connected to the Mosfet metal tab plate then directly connected to the amps HT CT. I believe I usually connect the HT CT, the bias ground, the filter cap negative and earth ground to this "power star ground" that would have the mosfet mounted to. I hope I was clear on that!
Have I missed any critical concepts to this?

Thanks for your time on this,
Best
Phil D

[xtian]

Sounds like you're clear, Phil. In stress testing this circuit, I used Zeners to drop as much as 80v without issue, using chassis as heat sink.

[GAStan]

Just an FYI- should you ever need to electrically insulate such a device from the chassis/heat sink it is possible. Search "mica insulator" and "insulating washer" at Mouser. Apply heat sink compound to both sides of mica insulator. It will be electrically isolated yet still thermally conductive. This was common on older power transistors in solid state power supplies.

[pjd3]

Thanks guys for commenting on this.

Now, I think it is worth bolting the Mosfet directly to the chassis with some good thermal grease right next to the power ground. One thing I can say is that this amp will never be pushed to heavy volumes as it will be played clean in small to medium rooms and miced. Ha, no-one ever lets me get very loud, ever. It occurs to me that the Mosfet sees the full HT voltage regardless of whether I'm generating a signal or not but, I think this is worth a try. I may even be able to grab a nice Fluke meter with bi-metal thermistor wire from work to measure the temperature at the chassis sight. And if it never gets close to the 150 degrees or so "fry temperature" then I'm probably good to go.

If for any reason, things aren't looking or sounding too good I'll simply take the insulation/drain wire to ground/good heatsink approach. Hey, I'm only asking for 20 volts. Lets see what happens.

Thanks very much for coming by. As usual, I'm probably playing it more safe than necessary.

Best,

Phil D

[TUBEDUDE]

Just an FYI- should you ever need to electrically insulate such a device from the chassis/heat sink it is possible. Search "mica insulator" and "insulating washer" at Mouser. Apply heat sink compound to both sides of mica insulator. It will be electrically isolated yet still thermally conductive. This was common on older power transistors in solid state power supplies.

Screenshot_20231109-171122_Brave.jpg

There are thin silicone pads that are electrical insulators and are more thermally conductive than the mica/past combo. And they avoid the mess. (Don't get any of that paste on clothing, it's there forever.)

Screenshot_20231109-171211_Brave.jpg

Screenshot_20231109-171122_Brave.jpg

Screenshot_20231109-171323_Samsung Notes.jpg

[R.G.]

Now, I think it is worth bolting the Mosfet directly to the chassis with some good thermal grease right next to the power ground.

Good choice. You might want to look at Aavid's max Clip products: https://www.mouser.com/datasheet/2/2/Aa ... 578917.pdf These require one screw hole to hold down the device and do not require insulating that screw. The spring clip puts a near-optimum force on the transistor. Sizes for TO220, TO218, TO247, etc. After a lifetime of drilling and deburring and insulating power device mounting screws, I love these things. They're even great for the "FP" or "full pack" transistors that are completely coated in 2500V of insulating epoxy.

It occurs to me that the Mosfet sees the full HT voltage regardless of whether I'm generating a signal or not but, I think this is worth a try.

It sees the full HT if it's in the high side after the rectifiers. Tying the device between ground and the CT is more benign. But pick a MOSFET good for over 600V anyway.

I may even be able to grab a nice Fluke meter with bi-metal thermistor wire from work to measure the temperature at the chassis sight. And if it never gets close to the 150 degrees or so "fry temperature" then I'm probably good to go.

Good idea! However, the 150C fry temperature is the chip/die temperature. The outside temp on the case will be lower. It needs a temperature drop across the internal thermal resistance to drive the heat out. TO220s will usually have thermal resistance of 3 to 5 degrees c per watt, so 20W in a 4C/W device will drop 80C getting out to the case. At 150C in the chip, the case will be 70C. Try to set things up so your transistor's case doesn't get over 60C to 70C.

[pjd3]

Thank you all, such great informative and helpful feedback!

I like that clip! It has occurred to me that the screwing down of a 220 could possibly be biasing the "pressing force" in one area where the other end of the 220 may be lifting up compromising the heat transfer. A senior engineer where I work was trying to inspire me to use silicone pads as opposed to any grease, claiming that the grease is not really a very good thermal conductor and is only effective in the very small valleys between two well pressed surfaces. Well, it looks like that clip at Mouser addresses that issue pretty well. I like it. its a deal and order is in. OK, I'm definitely going to at least take my amp into work to use a Fluke/thermistor. The EE director of R&D is getting into electric guitar so, will find interest in this. I take it that I can place the thermistor directly opposite the MOSFET and touching the chassis at that point to get a good "outside the Mosfet case" temperature.

One thing I can say is that this amp with the previous component boards (Rob Rob's Blackvibe 6L6) was quiet as a mouse so, I will be keeping the same layout and grounding as before, just with this new board. It shouldn't take long to discover if there are any issues with the Mosfet reducer. I think its a good idea to use one for this project. I may even place in a switch that allows me to bypass the Mosfet reducer in case good ol' 6l6GC's make it into the amp. Which I"m sure would be nice too.

I will also keep those Mica insulators in mind as I am moving into having voltage regulators in future projects.

Again, thank you for all the good and generous feedback. Awesome.
Best,
Phil D .

[pjd3]

Just curious,

Would it be ridiculous or at minimum, useless to place a heatsink on the opposite side of the chassis where the Mosfet is screwed into? Just had to wonder if even a little heat could be mitigated away from the chassis into the air outside the chassis.

Phil D.

[R.G.]

Would it be ridiculous or at minimum, useless to place a heatsink on the opposite side of the chassis where the Mosfet is screwed into? Just had to wonder if even a little heat could be mitigated away from the chassis into the air outside the chassis.

Not useless or ridiculous, just maybe not as effective as the heat sink might otherwise be.

If you're thinking of putting an actual finned heat sink outside the chassis, it would be far more effective to bolt the heat sink to the chassis and cut a window in the chassis so you could bolt the MOSFET directly to the heat sink.

Heat flows much like current through various resistances. There is an internal thermal resistance of about 1 to 5 degrees C/W from the MOSFET chip to the metal pad on the MOSFET body. From the MOSFET metal pad there is an inevitably thermal resistance between the metal pad and whatever it's bolted to (or to the outside air if it's not bolted to anything). Careful use of grease/mica/kapton or oxides-loaded silicon rubber can get the gap from MOSFET metal pad to the heatsink down to between 3C/W to 0.5C/W if you do a really good job. That gets it into the metal of the heat sink. Inside a solid metal, the thermal resistance is much lower than across any air/grease gap, so the closer you can get your hot power devices to the metal fins thing the better. But the heat always ends up in the outside air, so the lower the resistances in the path to the outside air, the better.

[pjd3]

Thank you that explanation RG.

Of course, the Mosfet would be bolted to inside of the chassis right next to the power ground using the chassis as the main heatsink that is using thermal grease and also, i will be getting those clips that you recommended and like. They seem like a great addition and why not - they certainly are cheap enough, may as well just get them and use them!

I may even use my Micromesh pads to bring the chassis surface down to a highly flat and polished to bring even more surface to surface contact. The heatsink idea on the outside was to be an addition to the recommended method of fastening the Mosfet. But, it sounds like it might not be making a significant improvement to a well chassis/thermal grease fastened Mosfet.

thank you!

Best,
Phil D.