High Current MOSEFT Circuit Resistors

[Silntknight]

Good idea, but the load isn't adjustable. I think I'll just go in for some nice brakes (literally). Go to a junkyard and pick up a caliper/rotor setup and use that. I'd connect one motor as a generator turning the second. The shaft is connected to the rotor, and the caliper is used to provide a load. I can either use it with constant RPM, thus constant voltage and measure current, or I can go constant load and measure voltage based on varying RPM. I like the first for a true stress test and to get real torque/power curves, but I'd use the second for spark adjustment because seeing a change of 1RPM is easier than seeing a couple of mA.

 

[Jaguarjoe]

Dang, I forgot about the adjustable part. Its a key part of the system too.

Here's a pretty good ignition website:

GS4 Inductive Ignition Module | GILL

 

[Silntknight]

Well, I think the setup I came up with should work. I'll either use car brakes (if I can get a complete hydraulic system cheaply) or bike disc brakes. I might even just make my own system depending on what I have available. All I really need is a steel rotor and some really tough rubber with brake pads.

Anyway, now that I have my basic setup, I can go ahead and work out problems even before I build it. Thanks for the link. If my team has enough money, I might buy a small spark module instead of ripping one out of an old car. Do you know if many coils have their amp rating on the outside? I only have 11.5A MOSFETS.

 

[Jaguarjoe]

Auto emergency or parking brakes are mostly cable operated.

I've never seen any coil spec on a coil other than an ocaissional(sp?) "12V" stamped on a case. GM's HEI units used in 10's of millions of cars controls coil current at 4.5- 5 amps. High performance HEI units run at about 7.5 amps. Get an HEI (High Energy Ignition) coil from the junk yard for a buck or 2 and you should be good to go. Look for just about any mid 70's to mid 80's GM car. The coil is inside the distributor. It is not round it's square, IIRC it takes a 1/4" nut driver to remove it. It has a resistance of about 0.5 ohms so you'll need a ballast resistance of about 1 ohm to limit coil current to 7 amps or so.

What's a spark module?

Here's some good MOSFET stuff. Don't miss the heat sink part, you'll probably need that. Any semiconductor dissipating more than 1 watt will need a heat sink:

**broken link removed**

 

[Silntknight]

Thanks! I'll be sure to be on the lookout for that.

Just what Gill called their products, I think. They are like spark coils, but programmable.

I'll be making my own heatsinks. Aluminum plus a TIG welder. I'm good to go. I'm also making the enclosures out of aluminum. For the mainboard and shield, I will also use a small fan for cooling. I realized I haven't posted my latest revision for the PCB layout. Check these out.

Also, if my MOSFETS have a Vgs rating of 30V absolute max, is running a gate voltage of 12V a good idea? The threshold is 3-5V and my uP only puts out 3.3V, which might not be enough. Should I run 5V to the MOSFETS' gates or 12V?
EDIT: I found the answer to my question: "To turn a power MOSFET on, the gate terminal must be set to a voltage at least 10 volts greater than the source terminal (about 4 volts for logic level MOSFETs). This is comfortably above the Vgs(th) parameter. " That article was pretty helpful.

 

[Silntknight]

I think I might need some help with the heatsink requirement. I know that the thermal de-rating of the Rds(on) makes Rds(on) 1.4625 Ohms. I am pushing (estimated) 7.5A through it. The fastest I will be switching is on every 16ms and off 1ms after it turns on. It has a switching time of 170ns. This information gives me 1.4625*7.5^2 = 82.3W. The switching power dissipation is extremely low, so it's negligible. 82.3W seems extraordinarily large for my application. The maximum duty cycle is 1/15 = 6.67%. If I multiply 82.3 by .0667, I get 5.53W of power dissipation. Is what I'm doing reasonable? I really doubt it is possible to dissipate 83W of power from a TO-220 easily.

Also, it's unrelated, but can I use an LED instead of a diode for my DC-DC converter (5V regulator)? I have a 9V input and most LEDs have a 4V drop across them. Otherwise, I can use a 12V source. How good are 9V batteries for this kind of project? I read that they are designed for very low currents. Should I go with different batteries?

 

[Jaguarjoe]

1.4 ohms is a lot. Where did that number come from? What's the part # for the MOSFET? Look at this IGBT:

https://www.electro-tech-online.com/custompdfs/2010/11/IS2FISL9V5036S3.pdf

It has a Vces(saturated voltage drop) of about 1.5 volts. P= EI= 1.5 X 7.5= ~11 watts. At 1/15 duty cycle Pavg= ~0.7 watts and its got a built in zener diode.

Are you using a gate driver? The uP alone isn't going to work. 3.3v is way too low to turn a MOSFET or IGBT on.

A lithium 9v battery puts out about 1200ma-hrs. An alkaline puts out 600ma-hrs. A rechargeable NiMh or Nicad puts out 300 or less. Life will be determined by total ma load and what voltage the circuit drops out at. If lithium is too costly, use 6 AA's.

LED's are poor voltage regulators. Their Vf changes with current. For $0.50 you can buy an LM7805 5v regulator. Its old, but it works.

Building PCB's at this stage is a tad premature.

How do you calculate the Rth of a home brew heatsink?

 

[Silntknight]

That number came from derating the Rds(on) to 125°C, as a conservative estimate.

I am using a driver. I was wondering about whether to control the gate with 12V or 5V, but the article answered that question.

I'll probably just go with a bunch of AAs.

I'm not using the LED as a regulator, I am using it as a diode, but with an indicator for power. I remembered just now that I'd just end up blowing the LED anyway because I'd be pushing nearly 1A through it. I have a 7805 5V regulator already.

I'm not building the PCB yet. I am actually waiting for a few parts still. The layout I uploaded is what I designed for etching after I test it out.

For calculating the Rth, you just make a big enough heatsink. The entire enclosure will be part of the heatsink. Also, I would simply use the thermal resistance of the base material. I can get that from the specific heat. Then, a little math with surface area and the conductivity of air and I arrive at the thermal dissipation rating. I know I need 5W dissipated, so I make a heatsink capable of dissipating 10W.

 

[Jaguarjoe]

E= IR= 7.5 x 1.4= `10.5 volts across the MOSFET. With a 12v supply that leaves 1.5 volts for the ignition coil. A 0.5 ohm coil @ 7.5a needs 3.75v to work. 125 degrees rise is insanely hot. That's without the heatsink?

AA cells put out 2500mah+

 

[Silntknight]

Correct, that is assuming the worst case scenario. I'd rather make the heatsink too large than too small. In reality, the Rds(on) .75 or so. That would leave 6-ish for the coil.

Good to know. Each one, I assume, sources that much current?

 

[Jaguarjoe]

Each one what sources that much current?

 

[Silntknight]

I meant batteries. Each AA battery sources 2500mAh (aka 2.5Ah)?

 

[Jaguarjoe]

Those are alkalines. 6 in series will get you 9 volts at 2500mah. That should last for a while.

 

[Silntknight]

Yea, that would do nicely. I'll look around for some Lithium Polymer variants. They should last longer. Then again, batteries are cheap and I have plenty.

 

[Jaguarjoe]

You can buy a bunch of alkalines for the cost of one LiPo.

 

[Silntknight]

That was my thought. Can I string them in parallel to add the capacity instead of voltage?

 

[Jaguarjoe]

Sure, string them in parallel as long as they are the same voltage. You can also go to C or D cells.
Here's something that might keep that hot MOSFET cool:

**broken link removed**

 

[Silntknight]

I was actually thinking about using one of these: Cooler Master V8 CPU Cooler at Xoxide!
or maybe one of these: Ultra ChillTEC Thermoelectric CPU Cooler at Xoxide!

Really, the link you sent me seems pretty good. It might save me some trouble to get that one actually. Building one wouldn't be so bad, but I'll save that for the injector MOSFETS. They run under 1A and have really low duty cycles. I could even get some small MOSFET heatsinks like these ones: Newegg.com - Enzotech MOS-C1 C1100 Forged Copper Heatsinks only.

 

[Jaguarjoe]

That Newegg part looks really nice. The price is too, 10 for 11bucks. 90 and 100 bucks for those other parts are pretty steep. Out of my price range anyhow.
Poke around the All Electronics site. They have Peltier devices, fans, heatsinks and lots of other stuff.

 

[Silntknight]

I think the MOSFET heatsinks should be on my to-buy list. I was joking around with the others, though I could cool my entire system with one of the other two. I would just need to make some heatpipes.
I actually have some Peltier Coolers, but I'm using those for a powered intercooler.

Also, my opto is non-inverting... I have inverting gate drivers...