Single Ended Tokin SIT THF-51S Common Drain Mu Follower Amplifier, 45W?

Good evening All,

I ran a temp. test tonight.

I set the AC to 26C and let the room stabilize, I then place a thermocouple in between the center two bulk capacitors in the center of the PS filter. This is also roughly in the center of the amp. Of course the thermocouple read 26C at the beginning of the test. I then ran the amps and looked for the temps. To stabilize. At about the 3.5hour mark it hit a peak of 44C, and then bounced between 43-44 for the remaining couple of hours.

This makes me less worried about need to actively vent the chassis.

So this is about 18C rise about ambient, can I assume if the room temp is higher, that it would still rise roughly 18C or is this not typically a valid assumption?

G
 
Thanks ZM!

for those of you interested I measured the SIT temps after 6 hours at it’s base, the temp. was 76C. The MOSFET was a bit hotter at 80C, measured at the washer for the screw holding it down.

I have the SIT directly mounted to the heat sink with just thermo compound, and the MOSFET is using a Keratherm pad.

G
 
ZM,

Sure, if I can run it to equilibrium tonight I will try to post them.

All these measurements were taken with a thermocouple touching the area I described, but I haven’t necessarily calibrated the thermocouple. I don’t currently have a IR thermometer.

G
 
ZM,

Here’s what I measured tonight:
SIT
case top to base of case 74C
base of case to heatsink 68C
heat sink just above the SIT 58C
top of the heat sink 51C (not on the fin)

MOSFET
large washer on the screw 80C
inside the hex of the screw 76C
case to heat sink interface 57C (keratherm prevents touching the heat sink)
Heat sink just above MOSFET 51C
top of heat sink 46C

tonight the SIT was a bit cooler at 74C than what I measured yesterday.

honestly I’m not totally confident in these measurements as the tip of the thermocouple seemed to give different readings depending on how it was placed on the surface. I recorded the highest temperature that I could measure above, and I could repeat it if done in the same way.

G
 
Thanks ZM

remeasured with thermal paste.

All the numbers were within a degree.

one thing I noticed on the MOSFET case, if the thermocouple was touching the case and the keratherm, it was reading 57C, but if I was just touching the MOSFET case and not the keratherm it was reading 71C.

I wonder if I should try switching to mica? It seems like such a big difference near the interface.

G
 
Keratherm 86/82 is better than Mica

anyway, think of some sort of Babysitter - your hardware will last longer...... or simply - you'll avoid any surprises

I'm not fretting about SIT - sturdy bugger, but we always need to think of tiny mosfet

going back through your posts, it seems that you have approx. 30V*2A5 at mosfet ....... which is 70W

ouch

try obtaining matched pairs of these, and replace that single one with a pair

change is trivial, and you'll get half dissipation per TO247 case

post your exact schematic and we can go in necessary details
 
ZM,

Thanks as always!

I ordered two fan controllers today, they look similar in function to the one Citation was suggesting, so I think I will try a babysitter.

Maybe something as simple as getting taller feet and mounting 1 fan on the bottom of the chassis for each active, a 120mm fan would blow cool air through the fins of the heat sink and the lower internal vents of the chassis. Short term I saw a drop in temps just having a house fan pointed at them.

Frankly, I'm not super convinced of the Keratherm as seeing 80C on the washer and only 51C on the heat sink right next to it seems a huge difference. I'm thinking to pull out one of the MOSFET and replace it with split mica (I typically use my finger nail to split the mica into multiple thinner layers) and thermal paste and then measuring again to see if I can get better heat transfer at that interface.

Long term I still think I would be happy with no moving parts (no fans). What sort of temperatures would be safer?

Thanks for the offer to modify my schematic, I thought I would take the first stab at it to prove to myself that I am learning. You can just review it for correctness based on what you taught me before:

1688375150869.png


From a parts on hand perspective I don't have additional of these IXTH, so buying more and then buying enough to get two matched sets might be a lot of parts at $5 a piece it could get expensive quickly. The DS for these parts has these at 460W Pd at a case temp of 25C, obviously I am no where near that, also from a chart in the DC for a case temp. of 80C the max current is more than -12A. I just trying to understand how much head room I have here.

My pocketbook preferred option if I need to go multiple MOSFETs would be my original idea of parallel IRFP9240 as I have a number of these already matched, but Ben mentioned it might not be rated for DC... I think I could just make a small board to keep the source resistors organized. Maybe using 3 of these on the 70W of dissipation would be less 25W each.

Any thoughts or comments on this strategy?

G
 
Listening impressions are subjective. Different people will have different opinions and preferences. The sound will also depend on the rest of the system.

Objectively, the SissySIT has less than one half of the power output. The SissySIT also has more than two times the harmonic distortion at 1W.
 
I would say that all of the SIT follower amps sound great. Choke loading will dissipate less energy but mu follower follower will have higher power output.

I have a variety of single ended follower SIT amps, each with different power output and energy dissipation for the different seasons of a year. The THF-51S mu folllower follower for winter, 2SK180 choke loaded follower for autumn and spring, and 2SJ28 choke loaded follower for summer. My speakers are 103dB sensitive so I am usually listening to the first Watt, but since I have electrical heat in my home, the energy dissipated by the higher powered amps is not wasted.
 
Ben, ZM,

I decided to spin a new board based on Ben's comments to shorten the wiring, as well as more cleanly implement the 2X IRFP9240 in the Mu follower following ZM's advice. Additionally, I combined the FE2022 (using parallel 2sk209) onto the new board to make things even cleaner. I ended up getting everything to fit on on a double layer 10X10cm board, so the fee from JLCPCB was only $2 plus shipping. This cut the wire length in about half and is a bit cleaner.

This weekend I rebuilt both channels and listened for about 4 hours before taking the temps., the MOSFETs are now reading 60C on the washer, without active cooling. So this is working really well. The temps for both MOSFETs are the same, and the heatsink is noticeably cooler.

Still sounds great to my ears, I can't notice a difference, but the cooler running MOSFETs are recommended by ZM for better reliability. (for those who have not read the whole thread, I wasn't using the MOSFET that Ben recommended, which I believe will have no problem with this heavy dissipation.)

I haven't confirmed the amps with REW yet, but will do later and post if I find anything interesting.

G