Maintain current

[ridelong]

All,

If the problem is runaway heat/current, why not just use a PWM with current feedback?

I can design one if it is worth while.

Will post design for free. Just want to screw the screwers.

Any opinions?

[sp1r0]

Sure, I'm up for a good design. Let's mention some spec's first.

1. Needs to cover a frequency from 1Hz-500kHz(maybe 200kHz if higher freq's are a problem with current demands)
2. Duty cycle 10%-90%
3. At theoretical 100%, the power MOSFET needs to conduct 200Amps DC at 12volts, 24 volts max. I know this would mean big heat sinks and possibly fans... I've seen good power MOSFET's capable of dissipating 500W for $35 each.
4. Money not an option. This should be the "best of the best".
5. What have I missed

Thanks for that offer, are you going to build it?
I could help in testing and design.

[sp1r0]

I would be willing to build a circuit around this if you design it:

http://www.st.com/stonline/books/pdf/docs/8220.pdf

[sp1r0]

Yeah I thought so, gotta do it myself...

[ridelong]

sp1r0,

Doodled the first schematic last night.

Went down to the basement to prototype it this afternoon, and am missing some stuff. Found what I need at Digi-key. Will order tomorrow if they are open.

Read your suggestions. First, I want cheap. Was going to design this thing to work only to 10 amp (for my gen). Decided to extend to 30 amp, because that seems like what the high amp guys run.

Frequency will be determined by the feedback loop. In this circuit, the fet will turn on and off due to the current.

Haven't figured out how to transfer a schematic.

[c02cutter]

If designing for use that people will build, design for maximum and let the rest figure it out for settings. Any electrical component needs to be designed on over kill in this environment to allow for someone that does not know to use it. If designing at lower amps, you'll get a lot of people building it and blowing it immediately. So go for the over kill when it comes to components and such. Get the schematic as a working unit, and let them figure it out from there. Most do not really understand an electrical schematic, other wise there would be a bunch of zero's pwm's available. Remember heat is an issue with this kind of device.

[overtaker]

Good advice co2cutter!

[sp1r0]

Hey ridelong, try mouser.com for electronics, I think you will find they are far superior. That FET I recommended a few posts ago is abit pricey at ~$34, and may need a pricey heat sink as well, but as an alternative, could you parallel a few IRFP064's? It's in a TO-220 package and handle's 70A max. What do you think of this shcematic? How would you modify it to run multiple FET's in parallel?

alt.nrg.org: It's in the Plans & Schematic section.

[BIGGUN]

Damn it.....I hate being stupid or at least not knowledgable.

A pwm is my next undertaking and I guess I have to blindly build from threads that I can find on the web. What is going to be different on this design than the ones that are currently offered? Most of the build is over my head but I can understand using a pwm and how it helps in production.

Thanks,
Kevin

[ridelong]

All,
I am going to start with a IRL1404ZPBF-ND From Digi-Key. It is 40 volt/ 75 amp on, but the on resistance is only 0.0031 ohms. At 30 amps, that will waste 900 X 0.0031= 2.8 watts in the device . No massive heatsink. Plus, it is only $3.67.