FEELER: Ultra-High Current Test Load

[benmwv]

SMD components? That's pretty crazy if you ask me.

I found these (MBRB4030T4G ON Semiconductor | MBRB4030T4GOSCT-ND | DigiKey) on digikey that have 0.55 Vf at 40A. 45x0.55= 24.75W There is no way 24.75W x 12 diode (for up to 6v) = 297W could be dissipated by a little board with surface mount components. And probably add a couple tens of watts more for the current sense resistor. Even if you heatsink them it would have to be through the plastic or through the board, either way it won't be good enough to dissipate all that heat.

I'd get 12 of these (MBRB4030T4G ON Semiconductor | MBRB4030T4GOSCT-ND | DigiKey) and use 3 standard square CPU heatsinks like this (Newegg.com - Rosewill RCX-Z80-AL 80mm Sleeve CPU Cooler) with 4 holes drilled in them and mount 4 of the diodes on each.

 

[Wolfman29]

The way I see it, if they are designed to withstand 40A at .55Vf and are SMD... they have to be heatsinkable =p

Also - the current sense resistor only dissipates half a watt

But I may as well build one and find out!

 

[benmwv]

Just because they can handle 40A doesn't mean they can handle it all day. They will be past the max rated temp in a few seconds I'm sure. D2PAK simply can't be heatsinked well enough to dissipate that much power.

So you are going to use a 0.00025 ohm reistor? I assumed it would have to be more like .01 or .005 since you said it would be accurate down to around 1A.

 

[Cyparagon]

I don't know a huge deal about SMD stuff, but I'm pretty sure much of the sinking takes place because of the large ground plane. If you try to put more than one component that dissipates 20W into the plane, I suspect things will start melting. You'd need to put the sinking over the top and clamp down every diode.

 

[Wolfman29]

Hmm. I may have mis-stated myself... I'll check on it later. But this is an "on-the-backburner" idea.