I apologize if I have managed to overlook it.
I am building a fuel pump driver board to run about 40 amps. This will have 5 or 6 mosfets that will need a heat sink. This will be in a cars harsh dusty environment. I haven't seeing designs that show heat sinks in an enclosure with a fan. I would really like to prevent all the dust/moisture from getting to the board. Just wondering if anyone had an suggestions or ideas. My initial thought was to us potting to half fill the case leaving a channel through the case with the exposed heat sink fines. Then just blow the fan through that channel.

Thoughts?

 

That is exactly how I have seen potted power modules.

 

Careful design of the Circuit can reduce the dissipated Heat down to an extremely low level.

There are Cast-Aluminum-Boxes readily available that would make a perfectly adequate Heat-Sink,
either Potted, or with an "O-Ring" gasketed, screw-down-Lid.
They are also available with Flanges for Screw-Mounting.

Are You doing just-plain "On-Off", or "PWM-Control" ?

Do You have a Schematic to share with us ?
.
.
.

 

A 480W fuel pump?
and 12V MOSFETs are available with Rds(on) <1mΩ, so 40A will only dissipate 1.6W.

 

Race-Car Fuel-Pumps can be pretty serious pieces.
Is this a Race-Car application ?
.
.
.

 

Yes this is for a race application driving 2 walbro 340 series pumps. I'm just in the prototyping phase. I'll try to post a diagram later. I don't have one with all the mosfets added. Just a single for simulation.

Ian0 do you have a part number? I had some extra irf3205 on hand. They are 8m and I'm getting a pretty high calculation which is why I was going to split the load over several mosfets. As well my ambient temps will most likely be as high as 60+ c. If I could get the rds down that would help a ton. I'm still new to this so sourcing the best parts has been a studious eventure.

Lowq was cases are you reference. I have spent the last 2 evenings trying to find a case that would mount mosfet and keep the air channel separate from the main board.
This will be a sensor (0-5v ref) driven pwm module I'm trying to build. From what I have read I am trying to keep it to 1k hz.

I greatly appreciate all the help.

 

 

This only a 10 amp motor as that is what everycircuit maxes out at. Again this was just to see if the pwm concept would work with my ref voltage

 

I apologize if I have managed to overlook it.
I am building a fuel pump driver board to run about 40 amps. This will have 5 or 6 mosfets that will need a heat sink. This will be in a cars harsh dusty environment. I haven't seeing designs that show heat sinks in an enclosure with a fan. I would really like to prevent all the dust/moisture from getting to the board. Just wondering if anyone had an suggestions or ideas. My initial thought was to us potting to half fill the case leaving a channel through the case with the exposed heat sink fines. Then just blow the fan through that channel.

Thoughts?

Welcome to AAC.

While potting might offer some advantages, such a vibration resistance, the first thing I would be is put a conformal coating on the board. There are many options, and I would certainly be looking for something that offers chemical resistance against petroleum by-products.

I would also echo and emphasize @LowQCab’s advice about using an aluminum enclosure. Even if it was not sufficient on it’s own, you can thermally couple your MOSFETs to the box from the inside, and thermally couple additional heat sinking on the outside.

I would try to provide sufficient passive cooling to avoid the need for anything active. As soon as you include a fan, you are adding in a possible point of failure and a potential maintenance requirement. Since the pump electronics should only be heating when there is a high fuel demand, this seems to suggest the car will be in motion when the maximum cooling is required.

If this is the case, routing some air from under the car, or somewhere else, delivered by the forward motion of it, would seem preferable to trying to keep a fan running in that environment. I believe the combination of a conformal coating and a good, sealed enclosure would allow for maintenance of the board if needed while potting would make it very diffcult.

Pay particular attention to the quality of the passthroughs for signals and power to ensure they are well sealed and rugged enough. Connectorizing them with appropriate automotive grade connectors, like the Deustch range, is probably best.

If you do choose to pot the board, I would encourage you to conformally coat it first for certainty of protection and the possible side effect of making depotting possible, and to use a flexible potting compound for both this reason and for superior shock protection.

 

Lowq was cases are you reference. I have spent the last 2 evenings trying to find a case that would mount mosfet and keep the air channel separate from the main board.

The scheme I have in mind, and I think @LowQCab as well, it so attach the MOSFETs tab directly to the aluminum enclosure, then if needed, add a heat sink to the outside. There would be no air movement inside the enclosure and all the radiation and convection would be happening to the enclosure, which the MOSFETs were made (thermally) part of.

 

The scheme I have in mind, and I think @LowQCab as well, it so attach the MOSFETs tab directly to the aluminum enclosure, then if needed, add a heat sink to the outside. There would be no air movement inside the enclosure and all the radiation and convection would be happening to the enclosure, which the MOSFETs were made (thermally) part of.

I was wondering if that would work. How do you calculate the new c/w?

 

I was wondering if that would work. How do you calculate the new c/w?

It is treated like a circuit, you add the various thermal resistances together. This calculator might help, or did I misunderstand your question?

 

You need a proper MOSFET driver.
A comparator will not have enough output current to switch it quickly enough, and that will result in high losses.

 

You need a proper MOSFET driver.
A comparator will not have enough output current to switch it quickly enough, and that will result in high losses.

I was wondering if I needed to add a driver to it or not. I will work on adding one.

 

It is treated like a circuit, you add the various thermal resistances together. This calculator might help, or did I misunderstand your question?

I tried the site on my phone but i was not getting numbers. I'll try it on my pc later. I was talking about heat sink c/w numbers and enclosure c/w numbers being calculated. If I mount a mosfet to enclosure face with a c/w of 15 and then a heat sink with a c/w of 5 to the exterior of the enclosure what would the new c/w be. Would it be 10?

 

I tried the site on my phone but i was not getting numbers. I'll try it on my pc later. I was talking about heat sink c/w numbers and enclosure c/w numbers being calculated. If I mount a mosfet to enclosure face with a c/w of 15 and then a heat sink with a c/w of 5 to the exterior of the enclosure what would the new c/w be. Would it be 10?

I see what you are asking. You can consider the enclosure and the heating as one thing, and it would have a lower thermal resistance than either alone. You an calculate an approximate RΘ for the enclosure based on the material thickness and exposed surface area, then subtract the RΘ of the heatsink, and add back the RΘ of the junctions of the device to the enclosure and the heatsink to the enclosure.

My expectation is that an estimate will be good enough because you won’t have any problem adding excess capacity to the system. I would make empirical measurements and (because I am me) instrument the controller for temperature and probably electrical things (V and A) for diagnostics.

 

I was wondering if I needed to add a driver to it or not. I will work on adding one.

A complementary emitter follower would probably do the job.
By the way, I hope that really is a comparator and not an op-amp!

 

A complementary emitter follower would probably do the job.
By the way, I hope that really is a comparator and not an op-amp!

In everycircuit program that is a op amp being used as a comparator.
As I said I'm new to this and learning. Greatly appreciate all of the infor

 

In everycircuit program that is a op amp being used as a comparator.
As I said I'm new to this and learning. Greatly appreciate all of the infor

That will make it ten times worse. No wonder it is getting hot.
Very few op-amps make good comparators.
Some op-amps won‘t work at all.
Many op-amps can be used as slow comparators.

 

That will make it ten times worse. No wonder it is getting hot.
Very few op-amps make good comparators.
Some op-amps won‘t work at all.
Many op-amps can be used as slow comparators.

Until 2 weeks ago I didn't know what a mosfet was much less than a comparator or op amp. I have always wired in relays and ran it. After wanting a fuel heating issue I researched pwm and my brain took off running.


I just read up on comparators and have a couple here. I will be shaping it for testing