Mosfet connector

Use one of those 0.100" pitch molex connectors.

I believe a TO220 has the same pitch.

Get the header and Housing, plus pins. Check the current rating too.

Cheers

"Martin Riddle" wrote in news:i51kru$j6t$ snipped-for-privacy@news.eternal-september.org:

Thanks Martin I'll definetly be using something like that for the board part.

I've been looking at Digi and Newark for something I could use to connect to the FETS (TO-3PN package) though. I know there must be something that is used I just dont know the name of it. Soldering wire directly to the PINS looks messy ;-)

is

Oh, there are sockets, all right; soldering to them is scarcely less messy, and they cost (with the insulator kit and loose hardware) more than some FETs... The sockets have to match the current you're pulling, of course, which is NOT easy for big MOSFETs. 18 gage wire, you said? The fusing current is over 80A.

allelectronics.com part number CON-243

Oops, I had asssumed you were using TO-220.

The DigiKey ED7636 aka Mouser 575-361012 is a strip of connector pins with 0.10 spacing that accepts FET pins (by MillMax).

TO-3 sockets i think are made by Kobiconn or Keystone.

whit3rd wrote in news: snipped-for-privacy@v8g2000yqe.googlegroups.com:

I guess I'll just have to solder and use heatshrink. I was hopeing some audio guru might have had a suggestion but it seems most are to pricey.

It will be 60A max 10A countinous each through 6 FETS. The 18AWG interconnects will be small runs (one inch max) the combined total will return through a bus bar.

Its for an active load I'm putting together. Im after as wide an input voltage as possible. I just tested 2 of the fets at 200VDC 130W on a relativly small heatsink/with fan in 27c AMBIENT and the heatsink case was only hitting 70 degrees or so. I have two large bricks (Thermal Resistance:0.08°C/W) I'm putting 6 on one for now good for 300W countinous to 50C ambient or so.If I ever need more power handling I can always use more fets or add the other heatsink but for now 300W is good enough for genaral purpose use.

60A will let me test down to 5V 300W PSU'S After that FET rdson is the liminting factor 0.2ohms at 25C,

Hammy wrote in news:Xns9DDEC888B39FAHammyhamsterca@69.16.185.247:

watch out for thermal problems. On the 2200 TEK scope line,we had to remove the Molex connector to the switching pre-reg FET because it overheated and charred. not a real high current circuit,either. This was a TEK mod to be done on all scopes that came into the service center.

--
Jim Yanik
jyanik
at
localnet
dot com

On a sunny day (Wed, 25 Aug 2010 11:47:45 -0500) it happened Jim Yanik wrote in :

Yep, solder those connections at 10A or higher.

You could run blank PCB lengths as bus connections? Top picture shows this technique with transistors and emitter resistors:

Are you using source resistors? I found it fairly easy to desolder the MOSFETs from PCB and resistors, used source (emitter) resistors to jump the blank PCB, gate (base) resistors to a common drive bus.

Bottom of page show a smaller heatsink with three pass transistors, this was for a 50V at 3A bench supply regulator that I prototyped but didn't finish by properly boxing it.

You could switch in power resistors to share the load at higher voltages?

Grant.

I agree with you, never met a reliable socket on flat pak power device.

Grant.

Grant wrote in news:a56b7616t4njg59hh6ropdkfc124bhtts6@

4ax.com: [snip]

I have some metal bars from some clamps I overstressed, I almost threw them out. I'm glad I kept them now. I should be able to drill and bolt wires to it witout any difficulty.

Yep I'm useing source resistors and individual opamps per fet the opamp worse case offset is 150uV so I can get away with small resistors. I havent completly decided yet but I'm thinking maybe 2 x 0.2 ohm 2W 1% thin film and a 0.01 ohm

2W 1% thin film for the high current. I tested it out to 130W on a breadboard and I can go down to a couple of mV for my feedback signal.

There is'nt going to be much PCB other then the board for the opamps and the power supply for the opamps etc. The current will all be going through bars and

18AWG stranded wire.Well and the fets too.;-)

I thought of that too but I dont think its nesscessary.I'll only be useing the fets at 55W max. Also switches for 200VDC rated for 5A or better arent cheap. I'll never be drawing 5A at 200VDC but I will at say 15VDC so the switch would get expensive.I'm useing an 8mOhm fet to switch between sense reistors 40 cents per; cheaper then any toggle.I'm also trying to keep it as small as possible.

The heatsink I used for a quick test with two FQA24N50 is about the same size as the one in your pic the P4 CPU Intel stock model one.

I've designed in high side P-channel switches for next active load, all saturated though. My top voltage is much lower than what you're designing for, only 32V. They switch in extra resistors to allow the top 5 or 6 V to work without too much strain. All hand-waving until I hook the thing up :)

Okay. Hope it goes well for you.

Yeah that size should be good for ~75W with fan.

Grant.

I've looked at Molex connectors, too, and decided the 'floppy-drive' type were inadequate for a TO220 at full stretch. When you move up to the 5mm spacing, there are inexpensive screw connectors rated 16A. I did try to find slim copper tube to act as a ferrule and bridge pin to cable, but my experiments turned out badly...

Given the trace lengths, do those 'FETs need a ferrite bead and/or small but local cap' on the gate pin to prevent instability ??

Also, if you are mounting fans, fit two: I've had cooling fans bind and seize. Okay, it is rare, but this PC nearly died as a result...

--------------------------------------- Posted through

Try wrapping a strand of wire around a mandrel (the shank of a suitable diameter drill bit works) then cut the spiral into little tube sections. Make it suitable diameter to take the wire lead and the transistor pin, and you can easily solder the assembly up, neatly. The tube-as- ferrule can be done, too, but it takes fluxing the interior of the tube, and benefits from an inspection hole... not easy to work up on the bench.

I put one in to be safe ;)

Temperature sensors are cheap. You can make a nice little PWM fan controller with LM324 and have opamp left over for high temp shutdown.

Grant.

On a sunny day (Fri, 27 Aug 2010 12:36:27 +1000) it happened Grant wrote in :

Even better than that, I replaced my washing machine dual thermostats + control

I do admit I used relays for the heater element and motor. Was a lot cheaper then a new one :-) Could not get the parts for that model anymore anyways, The PIC is glued against the tank, seems to detect temperature very well.

But do I want to glue a PIC to a heatsink for fan control? Hmm, they have the industrial temp range, and one must allow for the MOSFET junction to case... Could be done? Drive a 2N7000 in the fan power return lead.

Sometimes I'd like to sit a heatsink in water to define max temperature that way. Except water wanders away all by itself :(

Grant.