It’s so odd to see a hot Class A amp without big external fins. 🙂
With +/-24v rails and 1.93amp (0.33R on R131) you are putting out about 100w each heatsink. Nice to see a single fan can keep them all cool.
With +/-24v rails and 1.93amp (0.33R on R131) you are putting out about 100w each heatsink. Nice to see a single fan can keep them all cool.
Pcgab,
I like the music selection, another Metal aficionado 😀
Those heat sinks are working great running with 1.8A bias.
Agree!
That's actually the main reason why I started to think that devices won't get good thermal contact to heatsinks. Don't believe that they reach even 50c, I might have to get an IR thermometer.
Gab and Vunce, we share same taste what comes to music 😉
ALPHA 20 Builds now gone Live:
- xrk971 (2 pairs plus 1 prototype p2p)
- Vunce
- Pcgab
- Juntuin
- Alibear (how is the DC offset drift issue coming along?)
- who is next?
- xrk971 (2 pairs plus 1 prototype p2p)
- Vunce
- Pcgab
- Juntuin
- Alibear (how is the DC offset drift issue coming along?)
- who is next?
Absolutely. Long time metalhead here!
Alpha has been running for ~5hrs at this point.
Ambient is 21-22.2c.
Heat sinks are 35-37c, a few spots near 38, but most of the spots I check are closer to 36c.
Mosfets are 34-35c.
So, living in South Texas, most of our year is hot as hell. We keep the house near 77-78f ( ~25c ) during the day, and ~22c at night during the summers. The alpha will be sitting in a very well ventilated spot in the living room, and under 2 sets of ceiling fans that are on most of the time during the summer.
I think these heatsinks will be just fine for keeping the 20w Alpha cool.
Now my mind is on to the BB version. Is there any chance that passive cooling could be used in the BB? My concern, and I'm certain a layman here, is that simply using a larger heatsink would have diminishing returns, as there are only 2 output devices per channel. Maybe it doesn't matter as much as I'm thinking it does.
I seem to recall reading that placing the dissipating devices (mosfets in this case) near the bottom 1/3rd of the heatsink makes efficient use of the heatsink in a vertical orientation. So perhaps a similar configuration for the BB on an appropriately sized heatsink would work.
A quick look on heatsinkusa it would seem the SSR style heatsinks have good capability:
http://www.heatsinkusa.com/5-000-wide-extruded-aluminum-heatsink-for-ssrs/
~1.70 C/W/3"
Maybe two long sections of those profiles per channel, one per output device?
I think X did some calculations based on the C/W/3" rating of the heatsink profile I'm using now, I'll see if I can find that post and run the numbers for the BB.
It may still be simply too much heatsink to be practical though. Too large, expensive, etc.
Well, after re-reading X's post:
Aksa Lender P-mos Hybrid Aleph (ALPHA) Amplifier
I think I may have been looking at the C/W/3" rating backwards...
Which means something like this may be better suited:
http://www.heatsinkusa.com/8-000-wide-serrated-fin-extruded-aluminum-heatsink/
Likely still not terribly possible, I need to do the math...
Aksa Lender P-mos Hybrid Aleph (ALPHA) Amplifier
I think I may have been looking at the C/W/3" rating backwards...
Which means something like this may be better suited:
http://www.heatsinkusa.com/8-000-wide-serrated-fin-extruded-aluminum-heatsink/
Likely still not terribly possible, I need to do the math...
Does this look right?
http://www.heatsinkusa.com/8-000-wide-serrated-fin-extruded-aluminum-heatsink/
150w per channel (this may be inaccurate)
150 * 0.66C/W/3" = 99
100 / 30 = 3.33333 (30c being the rise I can tolerate in an average ambient environment of ~25c)
3.34 x 3 = 9.999999" of the above profile.
So 10-11" of that profile might work for a single channel?
That's ~$100 USD without shipping, per channel... hmmm... suddenly less interesting, though not nearly as large as I initially thought.
http://www.heatsinkusa.com/8-000-wide-serrated-fin-extruded-aluminum-heatsink/
150w per channel (this may be inaccurate)
150 * 0.66C/W/3" = 99
100 / 30 = 3.33333 (30c being the rise I can tolerate in an average ambient environment of ~25c)
3.34 x 3 = 9.999999" of the above profile.
So 10-11" of that profile might work for a single channel?
That's ~$100 USD without shipping, per channel... hmmm... suddenly less interesting, though not nearly as large as I initially thought.
One channel up and running on test loudspeaker, no speaker protection board so am not going to chance connecting to Spendors. Shortened supply leads and generally tidied up wiring and offset is steady at 12-15mV, may see if I can match input transistors a bit better. Can't really give any meaningful comment about sound quality, but sounds good through test speaker.
Tried huge passive heatsink 350 x 150 x 90 mm for one channel and within an hour or so it was too hot for comfort, so have now reverted to CPU coolers with Papst fans. Much cooler heatsink although output devices are too hot to hold finger on. I am using the silicon ? rubbery insulators, might try mica to see if that lowers temperature. Any suggestions?
At the moment I still have 22K feedback resistor, using a Quad pre amp the volume control is set very low for normal listening levels so I may reduce it to 15K or lower.
Thanks for all the support and advice so far, I'm sure there will be more questions to come.
Regards
Alan
Tried huge passive heatsink 350 x 150 x 90 mm for one channel and within an hour or so it was too hot for comfort, so have now reverted to CPU coolers with Papst fans. Much cooler heatsink although output devices are too hot to hold finger on. I am using the silicon ? rubbery insulators, might try mica to see if that lowers temperature. Any suggestions?
At the moment I still have 22K feedback resistor, using a Quad pre amp the volume control is set very low for normal listening levels so I may reduce it to 15K or lower.
Thanks for all the support and advice so far, I'm sure there will be more questions to come.
Regards
Alan
Attachments
I think despite the sink being large enough the bigger concern is the tremendous heat flux through the mounting pad. It is 100W over 4cm^2. Getting that hotspot spread over a larger area is important. So even without fans you would want to use say, 6 heatpipes to spread that over a larger area (similar to how 6 discrete MOSFETs would be spread out). 100W over a small pad is kind of like a soldering iron tip. We all know what happens to electronics at the end of a 100W soldering iron tip.
Congratulations on first sound Alibear! 
I wonder how I am getting 0-1mV DC offset on 4 amp boards now. It might just be lucky confluence of all the voltage drops between V1/V2/V3 and the diode and the P channel MOSFET. 15mV is not a big deal with speakers though as long as it is stable.
I ordered the aluminum oxide spacers which are 2x better than Keratherm. Will include 4 per pair of amp boards for $1 in case anyone wants it. I had to buy 100 of them. I got TO264 size so should work with all actives.
I wonder how I am getting 0-1mV DC offset on 4 amp boards now. It might just be lucky confluence of all the voltage drops between V1/V2/V3 and the diode and the P channel MOSFET. 15mV is not a big deal with speakers though as long as it is stable.
I ordered the aluminum oxide spacers which are 2x better than Keratherm. Will include 4 per pair of amp boards for $1 in case anyone wants it. I had to buy 100 of them. I got TO264 size so should work with all actives.
Exactly, I thought about heat spreader bars. I have some tellurium copper bars, ~3mm thick, 18mm wide. Though that would likely still not be enough to spread the heat load out enough to let the heatsinks work passively.
I was just thinking about using a finned heatsink with a nice thick bottom as a mosfet clamping device. Any thoughts on that?
As the famous TV ad salesman said about his chicken rotisserie oven, “Set it and forget it!” Adjustment-free Class A with DC-coupled outputs. The setting here is simply choice of R131 to set bias current.
Now if we could also roast a chicken with all that Class A heat...
Now if we could also roast a chicken with all that Class A heat...
Last edited:
Congrats Alibear!
The nursery is full of newborn ALPHA's🙂
In my setup, reducing the resistor helped with finer volume control.
I am running two 8.2k in series for R113 at present for 16.4k. Good trade off with lower feedback and decreased gain for finer preamp volume knob control.
Congratulations to Alibear, Gab and Juntuin!! More babies in nursery!
Alan, I'm a bit concerned that a heatsink of 350x150x90 did not cool the module; this is worrying. Were the fins vertical, permitted vertical convection along the rising fins?
I hope we get a few more snippets of the sound quality!
HD
Alan, I'm a bit concerned that a heatsink of 350x150x90 did not cool the module; this is worrying. Were the fins vertical, permitted vertical convection along the rising fins?
I hope we get a few more snippets of the sound quality!
HD
Last edited:
I think that would help some, but still not properly deal with the issue of spreading the intense heat load over a larger surface of the main heatsink.
There may be some better 'heat spreaders' to mount the mosfet to, that then mounts to the main heatsink to help. I've not looked much, tbh.
Cheers,
Gable
You really can’t beat a heat pipe for max flux. It’s way more thermal conductive than solid copper. The thermodynamic transport of heat via evaporation of water and recycling it via condensation is an amazing process. 300W/cm^2 can be dissipated with heat pipes.
I've been listening to classical stream (KMFA) all afternoon, and it is sounding lovely.
I'm going to work on a proper chassis this weekend, so I can put the alpha in my bedroom system with some very nice 2 way monitors. They really prefer a bit more power, but my listening position is not too far away, so I think it will be good for more analytical listening.
If that works out, the alpha will move to the living room and power one to two 'ways' on my to be built synergy horns.
Cheers,
Gable