Hi,
Two questions:
(1) I have a chassis with UMS heatsinks and a few boards that aren't designed around that specification. Is there any disadvantage to mounting MOSFETS on the heatsink and then connecting the pins to the board with some short wires instead of soldering the MOSFET directly to the board to enable me to use the predrilled and tapped holes on the UMS heatsinks with designs that don't adhere to those specifications?
(2) Could someone recommend a good thermal adhesive to bond some sheet aluminium to a pair of heatsinks for a DIY chassis project?
Thanks
Two questions:
(1) I have a chassis with UMS heatsinks and a few boards that aren't designed around that specification. Is there any disadvantage to mounting MOSFETS on the heatsink and then connecting the pins to the board with some short wires instead of soldering the MOSFET directly to the board to enable me to use the predrilled and tapped holes on the UMS heatsinks with designs that don't adhere to those specifications?
(2) Could someone recommend a good thermal adhesive to bond some sheet aluminium to a pair of heatsinks for a DIY chassis project?
Thanks
Hi,
regarding (2) ... I have often used this with good results:
http://www.arcticsilver.com/arctic_silver_thermal_adhesive.htm
Unfortunately this product seems to be discontinued. If you can get it, that is really good stuff.
Kerafol Keratherm-Bond 100 RT may be an alternative adhesive for the application. But I have not tested it yet.
regarding (2) ... I have often used this with good results:
http://www.arcticsilver.com/arctic_silver_thermal_adhesive.htm
Unfortunately this product seems to be discontinued. If you can get it, that is really good stuff.
Kerafol Keratherm-Bond 100 RT may be an alternative adhesive for the application. But I have not tested it yet.
Last edited:
No idea on #2 but as to extending leads...
MOSFET's are high frequency devices and you may encounter stability issues. A lot depends on what the application is such as an audio amp, regulator, Class D amp etc.
You may well need (probably essential) to include gate stopper resistors mounted right on the FET gate lead itself. These are typically around 100 ohm.
MOSFET's are high frequency devices and you may encounter stability issues. A lot depends on what the application is such as an audio amp, regulator, Class D amp etc.
You may well need (probably essential) to include gate stopper resistors mounted right on the FET gate lead itself. These are typically around 100 ohm.
You have a problem making new holes?
Not too difficult, you can always make a few away from the working area for practice.
Taps cost pennies, less than 2 UK for a set, buy carbon steel, enough for Aluminum.
Or use nuts, that works too.
That is a better solution.
Not too difficult, you can always make a few away from the working area for practice.
Taps cost pennies, less than 2 UK for a set, buy carbon steel, enough for Aluminum.
Or use nuts, that works too.
That is a better solution.
I do that all the time - it sure makes assembling the amp to the heatsink a lot easier by decoupling the process for the MOSFETs from the amp PCB. Just keep the leads shorter - 6in or so. And include snubber resistors as close as possible to the legs of the MOSFETs.
Here are some examples of my amps with this method:
It lets me remove the amp PCB for adjustments and replace it in 2 minutes of all wires are on quick connects. No soldered flying leads.
Here are some examples of my amps with this method:
It lets me remove the amp PCB for adjustments and replace it in 2 minutes of all wires are on quick connects. No soldered flying leads.
Yes, gate stoppers too. I have provision for both gate stopper, snubbers (RC between drain and gate), and protection Zener diodes on my little helper boards.
the helper boards also serve to clamp the mosfet to the heatsink with even pressure and eliminate mechanical stress on the mosfet pins to the flying leads.
the helper boards also serve to clamp the mosfet to the heatsink with even pressure and eliminate mechanical stress on the mosfet pins to the flying leads.
Hi XRK.
How would I go around my problem, in order to spread 6 mosfets per chanel more from each other in order to avoid heat spot with Aleph 5. This is the actual schematic and PCB where you can clearly see that 6 mosfets are to close to each other. https://www.kk-pcb.net/aleph-5.html
My amp case is Nelson's clone case from AliExpress with enough heat sink's power to cool them down. Would adding your helper boards with additional electronic parts change the way Aleph is working? Thank you in advance for your help.
Attachments
You can use the mosfet snubber boards and use the on board snubber resistor. Don’t populate the one on the PCB. They will allow you to mount the mosfet anywhere you wish for best heat distribution.
Thank you XRK. So basically to move and install 221Ohm resistor from Aleph's PCB on to your mosfet helper board as gate resistor of IRFP244 and to place jumper instead of them on Aleph's board and NOT to install any other parts on your helper board?
Thank you again.
Thank you again.
I really dig the flexibility of having the mosfets on molex detachable flying leads. Want to use that configuration for all builds going forward. Gonna order more of those helper boards as soon as @xrk971's Etsy store is back from vaca.
Of course, most amp boards don't have provision for molex connectors at the mosfet positions, but i suppose I can wire direct to the amp boards with molex connectors on the helper boards (unless someone has a better solution for this).
Of course, most amp boards don't have provision for molex connectors at the mosfet positions, but i suppose I can wire direct to the amp boards with molex connectors on the helper boards (unless someone has a better solution for this).
Seems to me that another advantage to detachable flying leads is being able to reuse output MOSFETs across amps. In fact, rather than use, for example, IRFP for some amps and IXYS for others, why not spend the money on the higher power MOSFETs and then reuse them across everything?
Or even better, reuse the whole chassis/mosfets/trafos/psu. Only swap out different flavors of amp boards assuming they are all similar +/-24v rails.
That's exactly what I'm now setting up, except the PSU is a separate chassis, and I've tried to make it easy to swap trafos for different rail voltages.
I'm interested in how you got those RC snubber values, R=47ohms and C=220pF. If there's a resource you could point me to, I sure would appreciate it. I did search online but only found examples from the world of high-speed switching, which is a bit different. Based on first principles, though, I think it goes something like this:
Start with the resonant frequency of the stray inductance from the flying lead wires and the parasitic capacitance in the MOSFET using the old 1/sqrt(LC) formula. Using the standard formula for the inductance of a wire loop (and estimating the geometry of 5 inch long flying leads) you get about 100nH of stray inductance. Looking at the datasheet for an IRFP240 you get 130pF as the parasitic capacitance between gate and drain (Crss). Plugging into the 1/sqrt(LC) formula, you get a resonant frequency of 44 Mhz.
Presumably, you then want the 1/(RC) frequency to be small relative to that resonant frequency. I don't know where you got the 220pF value (though it's roughly the same size as the parasitic capacitance, so maybe that's it), but if I plug your values into 1/(RC), I get a frequency of 15 Mhz. That's about a third of the resonant frequency, so given the 220pF capacitance, choosing 47ohms for the resistor seems reasonable.
Am I in the right ballpark?