The three caps next to the heat sink will get cooked. They also obstruct airflow to the fins. This pcb needs to be 1.5cm longer.
If you put them on the bottom side of the board with the heat sink on top, the caps will be much cooler. And the heat sink will get more air. The diodes too...
If you put them on the bottom side of the board with the heat sink on top, the caps will be much cooler. And the heat sink will get more air. The diodes too...
These are the load specifications. Just 1.25A per board so I don't think the caps need to be mounted at the other side of the board. It all depends on the voltage you will be feeding them and since it is made with a 1.3V dropout voltage LDO regulator you make it quite efficient.
I would use 105 degrees rated caps just in case as stuff will be warm anyway and of course the highest heatsink that fits in the casing like the 50.8 mm or even the 63.5 mm.
This is what the designer says:
http://www.pmillett.com/DC_filament_supply.htm
Last edited:
Better to go big on the heatsink than to run really hot IME. I’ve seen so many power supply faults in 40 years working in the lab instrumentation industry due to inadequate heat sinking or failure to account for all operating conditions. After you put your DIY gear into service, circumstances can occur that cause higher than expected ambient temperatures. For instance putting a hot running Class A amp in your stand under your front end can heat up the entire neighbourhood and cause issues. I’ve seen that last one in my own equipment rack. If it’s running too hot to touch and you have space, installing a bigger sink would be my advice. Oh, and there is no mechanism for convection within a solid because it’s a solid and unlike liquids and gases it’s molecules cannot move in response to changes in density. The convection happens in the air around the heatsink.
I read through that a few times. He says if you are clever you can mount them to the chassis but I'm not that smart!
My heat sinks are 55mm tall.
I am planning on feeding them with a 6.3V@8A. I have Hammond 266N12 on my shelf and was thinking that would do.
The caps he specified are 85° but I think I'll place an additional order for some 105° caps instead.
Thanks everybody for your opinions and help!
Yes 105 degrees rated caps and the 63.5 mm tall heatsinks for TO220 (same series as in the BOM, just choose the tallest) when they fit. Of course less hot = better. Chassis mount is possible but it will look ugly as the PCB was not designed for that AND you'll need to use insulation stuff for the regulator. It will be most user unfriendly when you need to take stuff apart. Besides that you'll need to drill extra holes that may be in the visible area and if not less holes is always better IMO. I would definitely put the small caps C3 and C4 at the other side of the board as they are really too close to the regulator. Don't make the mistake by reversing them when you do. Peufeu is right that the board is simply too small.
In all cases measure the voltage under load both AC and DC after rectification/filtering so before the regulator. Also measure output voltage, then calculate the power loss in heat. If you have a tempmeter then measure the parts after a few minutes, then after 15 minutes, 30 minutes etc.
In all cases measure the voltage under load both AC and DC after rectification/filtering so before the regulator. Also measure output voltage, then calculate the power loss in heat. If you have a tempmeter then measure the parts after a few minutes, then after 15 minutes, 30 minutes etc.
Last edited: