Driving FET Gate, is 2.5" too far from the driver?

[zombiess]

I've designed an add on board for a EB312 driver, but one of my concerns is I have almost 2.5" from the gate driver transistor to the FET gate on one of the phases which is a long run but I'm having trouble making it much shorter. I've read that it's best practice to keep the gate as close as possible to the driver in order to prevent ringing in the gate, but I've also seen some crazy stuff work.

I've read that the higher the inductance of the trace and the higher the frequency the more likely this problem is to occur. Since I can't place the driver near the FET I'm using a larger trace than what's on the stock board to help keep the inductance lower. The stock board uses a 0.030 trace and I'm using .050" as the smallest, but all long runs .080".

Any recommendations, does anyone thing this is going to be a problem? I'm going to have 4 of these boards produced for $100 for testing purposes.

 

[Njay]

Use twisted pair cable, that's what I've seen a semi manufacturer (don't remember if it was IR or Texas) and bigmoose advise.

 

[ZOMGVTEK]

I'm waiting on some copper blocks to show up, and then I will be doing a similar build.

Twisted pair appears to work just fine. I'm going to scope it with stupid long twisted pair and compare to a shorter run, along with messing with ferrite beads on the gate legs of the FET's to see if I like what it does.
I did see a few people do really dumb things that somehow work. It's probably OK to have a few inches of wire going to the gates.

 

[zombiess]

Thanks for the tips, I'll be adding some additional holes to add twisted pair if the circuit board causes ringing on the gate. I don't think I'm going to have problems, but I know there are many much smarter than my elementary knowledge of power control on here.

Best part about my setup is it's using almost all off the shelf components which makes my setup easy to replicate if it works. I hope to have it up and running next week with some pictures. Circuit boards are going off to be manufactured tonight so I should have them Wed or Thurs.

 

[crossbreak]

if you want to drive more FETs and longer wires you have to decrease the output resistance of your gate driver. ATM i'm thinking about doing this with an op-amp. Don't know if this could work

the question is, if you use twisted pair, how long will the fets travel through their linear region till they completely shut on?
the capacity of the cable has to be loaded.it would be nice to know how the gate capacity compares to the cable capacity.

all just thoughts though...

 

[texaspyro]

It all depends upon your gate drive configuration and the matching of the driver impedance to the gate capacitance. My welder drives 18 big FETs spread out over around 12 inches. There is as series matching resistor at the gate driver and a small diameter coax to the middle of the FET bank (also worked just as well driving the close end of the gate string).

 

[zombiess]

Ok, so stupid question time. With a twisted pair, one wire goes from the gate driver on the board to the gate, where does the other wire in the twisted pair go to, source pin and source return?

I'm not understanding the driver section of the board.

Could someone please tell me where a twisted pair hookup would go to on the board and the FET? I'd like to get this wrapped up tonight and get my boards produced this week.

here are some mega sized shots of the drivers sections of the EB312 board, view the image to see the whole thing.

 

[rhitee05]

For the relatively slow switching speeds you get with one of those boards, probably not a big deal. I'd suggest laying out your board so that you can place an extra series resistor just in case it's needed. The best would be to put a separate resistor for each FET, as this is more effective to damp out ringing between FETs without impacting switching speed so much. If you were to use twisted-pair, it would be connected to the FET gate and source at one end, and the driver output and source return on the other end. Either way, don't get yourself carried away and worry about the switching speed too much or you'll just cause yourself lots of problems.

 

[zombiess]

For the relatively slow switching speeds you get with one of those boards, probably not a big deal. I'd suggest laying out your board so that you can place an extra series resistor just in case it's needed. The best would be to put a separate resistor for each FET, as this is more effective to damp out ringing between FETs without impacting switching speed so much. If you were to use twisted-pair, it would be connected to the FET gate and source at one end, and the driver output and source return on the other end. Either way, don't get yourself carried away and worry about the switching speed too much or you'll just cause yourself lots of problems.

Ok, one end goes to Gate and the Source of the FET and the other goes to the gate driver and the source return. Where in my pictures is the source return? I'd like to know for future reference.

 

[rhitee05]

I'm not familiar with the layout of those boards so I can't really help you there. If you were using a discrete gate driver, it would have a pin that's supposed to connect to the source of the FET. Best bet is probably just connect them to the gate and source pins of the original FET mounting holes.

 

[tedcs]

... here are some mega sized shots of the drivers sections of the EB312 board...

I'm working toward getting a clue - the images are helpful!
Thanx