I had a very similar issue and it ended up being my jfets. They came from a trusted source but somewhere w me handling and installing them I jacked one set up. I had two more sets on hand, installed them and was good to go.
Pass DIY Addict
Joined 2000
Paid Member
I am about to start construction on my Aleph-J and would like some input concerning heatsink configuration. Originally, I was planning on using two of these sinks per channel. I was going to line up the pairs of output Fets with the center line on the sink as shown below on the left.
But then I started doing some calculations and came to the conclusion that two of these 5" x 5" x 5" sinks are likely to provide insufficient cooling. So now I'm planning to use three of them side by side with a heat spreader that will span each of the three sinks. I was thinking of using a 2" x 0.25" bar of aluminum as the spreader.
So here is my question: When I mount three sinks side by side, do I arrange them so they are touching each other (like in the second image), or should I provide a bit of "breathing room" between each of the three sinks (as shown in the first image)?
So which is it: space between or no space between?
But then I started doing some calculations and came to the conclusion that two of these 5" x 5" x 5" sinks are likely to provide insufficient cooling. So now I'm planning to use three of them side by side with a heat spreader that will span each of the three sinks. I was thinking of using a 2" x 0.25" bar of aluminum as the spreader.
So here is my question: When I mount three sinks side by side, do I arrange them so they are touching each other (like in the second image), or should I provide a bit of "breathing room" between each of the three sinks (as shown in the first image)?
So which is it: space between or no space between?
Attachments
The antistatic wrist strap and a good, solid ground connection is a must when working with MOSFET's / JFET's.
A few decades back, I worked with the above every day, for 2 straight years. I know from experience that those transistors can get damaged quite easily by electrostatic discharge - their sensitive capacitances get damaged, slightly..., but can't be checked for damage easily...with a tester, they show that they've been damaged once you soldered them on a PCB...and then they just don't "transconduct" properly... tricky... so, better to be safe than sorry.
Pass DIY Addict
Joined 2000
Paid Member
Fo sho. It was my first build so I had a lot of growing pains. I bought two sets of boards and soldered all 4 cause I knew I’d mess up somewhere.
Pass DIY Addict
Joined 2000
Paid Member
I am aware that Alephs run hot and was also wondering just how hot this might get with two sinks and what temperatures I am willing to accept. I'm remembering Nelson indicating at some point that the mosfets sound better when they're good and hot. With two sinks they just might just skip Blimey Hot, Crikey Hot, and Bloody Hot and go straight to the X*?@! = 60c category...
Last edited:
Eric,
Have a look at Transistor Dissipation and Turbo Projects to give you a better idea of where the limits lie.
Terry
Have a look at Transistor Dissipation and Turbo Projects to give you a better idea of where the limits lie.
Terry
Pass DIY Addict
Joined 2000
Paid Member
Ah - I've seen those formulas before, thanks for the link! My problem is that I have too many unknowns to complete the calculations with any real accuracy. I feel like I'm squarely in the trial-and-error phase. Guess I need to populate my board and do some drilling and tapping... This will take me a few days. Thanks for the input, guys!
I was in the same situation with unknown rating heatsink (8x11x3.5), and it worked OK. I believe 2 of 5x5x5 are fine. Aluminum Oxide Ceramic Pad helped to keep the FET lower temp.
4180G Aavid | Mouser
4180G Aavid | Mouser
Pass DIY Addict
Joined 2000
Paid Member
Hmmm, there seems to be a few choices here in terms of configuration which opens the opportunity for lots of experimentation. First, I need to prep the heat sink surfaces. They were milled/machined using a rough process that has left behind a pattern of ridges that can both be seen and felt with a finger nail. These need to go so I have a smooth and flat surface to work with. Wet sanding aluminum really makes a mess...
Next up is mounting interface material. Plasnu is recommending the Aavid AlOx Ceramic pad. It looks to have good numbers in the data sheet. I've been using Bergquist K-10 pads in a number of my builds, and the DIYAudio store stocks Keratherm pads. Perhaps I'll compare the performance of each here by measuring current draw of the mosfets and resulting sink temps.
The final thing I'll try is raw aluminum vs with a thin coat of black paint on them. I have been looking and asking around for months and I have been unable to find anyone nearby that can do anodizing work...
The link that Terry provided will be useful to see how much I'm stressing the mosfet junctions. After running a set of calculations, I'm thinking I'll reduce the bias in my giant Aleph-X amps a bit. Present configuration of 9A with 22v rails looks like junction temps are near 123c. If I back the bias down to 8.5A, junction temp should drop to somewhere closer to 113c which will help them last longer.
Should be interesting to see the results, thought it will also take me a while to do.
Next up is mounting interface material. Plasnu is recommending the Aavid AlOx Ceramic pad. It looks to have good numbers in the data sheet. I've been using Bergquist K-10 pads in a number of my builds, and the DIYAudio store stocks Keratherm pads. Perhaps I'll compare the performance of each here by measuring current draw of the mosfets and resulting sink temps.
The final thing I'll try is raw aluminum vs with a thin coat of black paint on them. I have been looking and asking around for months and I have been unable to find anyone nearby that can do anodizing work...
The link that Terry provided will be useful to see how much I'm stressing the mosfet junctions. After running a set of calculations, I'm thinking I'll reduce the bias in my giant Aleph-X amps a bit. Present configuration of 9A with 22v rails looks like junction temps are near 123c. If I back the bias down to 8.5A, junction temp should drop to somewhere closer to 113c which will help them last longer.
Should be interesting to see the results, thought it will also take me a while to do.
Pass DIY Addict
Joined 2000
Paid Member
Pass DIY Addict
Joined 2000
Paid Member
After some playing around with numbers for a while, I think 2 of these sinks might work for my Aleph-J. My data is estimated, but based on the performance of three completed amps:
My M2 amp dissipates 54w of power into 695sq inches of sinks and exhibits as 21c thermal rise. 54w/695sqin = 0.0777. 21c/0.0777 = 270*
My Aleph-X amp dissipates 70w of power into 587sq inches of sinks and exhibits as 33c thermal rise. 70w/587sqin = 0.119. 33c/0.119 = 277*
My A-40 amp dissipates 25w of power into 264sq inches of sinks and exhibits a 25c thermal rise. 25w/264sqin = 0.0947. 25c/0.0947 = 263*
*So, the result of each of these calculations is 270, 277, and 263. This is close enough to me to call it a relative constant (give or take, due to physical dimensions/configuration of the sink).
Projecting the outcome of the Aleph-J thus becomes: 50w/377sqin = 0.133. Multiply the constant ~270 * 0.133 = 36c projected thermal rise. Adding in room temp results in 60c final projected temperature. I think it'll work, but it will be my hottest amp yet...
Maybe this is a formula that others can use to project if sink size is adequate.
My M2 amp dissipates 54w of power into 695sq inches of sinks and exhibits as 21c thermal rise. 54w/695sqin = 0.0777. 21c/0.0777 = 270*
My Aleph-X amp dissipates 70w of power into 587sq inches of sinks and exhibits as 33c thermal rise. 70w/587sqin = 0.119. 33c/0.119 = 277*
My A-40 amp dissipates 25w of power into 264sq inches of sinks and exhibits a 25c thermal rise. 25w/264sqin = 0.0947. 25c/0.0947 = 263*
*So, the result of each of these calculations is 270, 277, and 263. This is close enough to me to call it a relative constant (give or take, due to physical dimensions/configuration of the sink).
Projecting the outcome of the Aleph-J thus becomes: 50w/377sqin = 0.133. Multiply the constant ~270 * 0.133 = 36c projected thermal rise. Adding in room temp results in 60c final projected temperature. I think it'll work, but it will be my hottest amp yet...
Maybe this is a formula that others can use to project if sink size is adequate.
after all these years , in meantime taking variac - xformer - battery of power resistors- heatsink under test , as most effective way for planning things .... yeah - after all all these years , I'm finding eyeball as pretty much sure way of ....... eyeballing
so , I made that one , creeping 100W per side ..... this one is similar in chunkiness ...... that's it !
so , I made that one , creeping 100W per side ..... this one is similar in chunkiness ...... that's it !
Pass DIY Addict
Joined 2000
Paid Member
Pass DIY Addict
Joined 2000
Paid Member