Paul & Sabrina's cheap DIY 144v motor controller

[jerpelea]

Originally Posted by williamson

I clicked on "pictures and albums" but then I didn't know what to do. I want to take the 8 files (schematics) from my flash drive and transfer them to the box. I told you I don't know the first thing about what I'm doing!!

you can zip the pdf files and upload them to multiupload.com

thanks

[ElectricZX2]

Originally Posted by CamLight

With the heat sinks used in my 400W electronic loads, using four TO-264 MOSFETs, the sink-to-ambient thermal resistance (of my heat sink and fan combo) was increased approximately 0.04C/W over the bare aluminum number. This is at least 10x better than any other material that can be used.

Don't underestimate the power of being thin.
Type III anodizing, while being a VERY efficient insulator (electrically and thermally) is so thin when applied to a heat sink that it really doesn't matter. Yea, if you increase the thickness you can probably block a blowtorch, but when thin, it allows enough heat to flow vs. its dielectric strength (ability to electricall insulate) that there's just about no other material that can match it. Certainly not any thermal pad, mica, kapton tape, etc. They might be able to if they were 1/2 mil thick...but they're not.

Do you know how thick the Type III lining was for those ICE piston? I suspect it's more for reduced friction and wear-resistance than heat-flow blocking...unless a very thick layer was applied.

Sounds like you've done your homework CamLight...I like those numbers. I do agree with you though to go as thin as possible while still being effective as an electrical insulator. In the Piston example the coating was much thicker than 1/2 a mil, and you're talking about primarily radiant heat transfer rather than conduction, which might account for a significant difference.
At 1/2 a mil are you really talking about Type II coating at that point? I thought the primary difference between Type II and Type III anodizing was the thickness...with a "hard coat" starting at around 2 mils.

[MPaulHolmes]

Well, PCBCart doesn't take ExpressPCB anymore. Makes me mad! I downloaded KiCAD, and I'm finding it VERY awkward to work with. It was written for another language I think. It's free though. I'm really starting to become a strong believer in the old adage "you get what you pay for". The Advanced Circuits software is pretty clean and easy to learn and free, but then you have to pay 50 gagillion dollars to get the boards. Maybe I'll hawk my car and get Eagle.

[jackbauer]

What about "hacking" the output files from The Advanced Circuits software into something useable?

[MPaulHolmes]

That's a good idea.

[jackbauer]

Originally Posted by MPaulHolmes

That's a good idea.

Thats why i'm Jack Bauer !

[CamLight]

Originally Posted by ElectricZX2

Sounds like you've done your homework CamLight...I like those numbers. I do agree with you though to go as thin as possible while still being effective as an electrical insulator. In the Piston example the coating was much thicker than 1/2 a mil, and you're talking about primarily radiant heat transfer rather than conduction, which might account for a significant difference.
At 1/2 a mil are you really talking about Type II coating at that point? I thought the primary difference between Type II and Type III anodizing was the thickness...with a "hard coat" starting at around 2 mils.

I don't know the small details about the differences (as I've only used Type III myself) but both Type II and Type III are offered in an overlapping range of thicknesses. But, it's a different process for both of them. Type III is a colder (below 40F IIRC) process that uses a higher current setting than Type II (which is room temp). It's also a darker solution so it's harder to use dyes to achieve different colors.

I've seen hardcoat offered down to 0.1mil but have only looked at vendors who have experience doing heat sink related anodizing. It's critical to go thin for this. For hardness (and not electrical insulating) purposes, I'm sure that the low end for the coating is much thicker as the mechanical properties of the coating are then more important.

Another reason to use a vendor who has experience with heat sink coating is that you might need to remove a clear coat of Type II anodizing from the heat sink. Many are plated this way and you can't apply the hardcoat until it's removed and you have bare metal. But, it's critical that the vendor not overetch the surface when removing the standard (Type II) anodizing!!! This leaves a matte surface that, while feeling pretty smooth and is still perfectly flat, is now a LOUSY conductor of heat from the sink to the MOSFET (or whatever). The matte surface is comprised of a much larger number of peaks and deep valleys that significanly reduce the surface area of metal actually touching when the MOSFET and sink are pressed together. Thermal compound fills in these low spots, but it's nothing like the results you get when two pieces of metal with mirror surfaces are used.

I got zapped this way on a batch of heat sinks two years ago. The thermal performance of my fan/sink combination dropped 20% just because of the over-etching of the surface.

[steiner]

Paul,

Have you tried FreePCB with PCBcart.com? That is the softare that I used with them. It was fairly easy to use but didn't have very many automated steps. It generates a generic gerber file that they were able to import directly.

The software from Advanced Circuits is pretty good for free but like you said it forces you to be tied to them for your PCB needs.

Rick

[jyanof]

Originally Posted by CamLight

Hi Joe,
Thanks for the links, I'll check them out.

Sounds like you have a good cooling system already in place...no need for heat pipes.

I'm a bit sleep-deprived right now but isn't the heat generated by the static resistance of a MOSFET proportional to the square of the current? That is, going from 200A to 400A would increase the power (heat) that needs to be dissipated by four times (Power = Current-squared times Resistance)?

Yah, the fan/heatsink seems to work well. I had thought about using heat pipes for an IGBT based power section. But, this is a whole other discussion

As for mosfet heat generation, it depends (mathematically at least) which term - resistive or switching - dominates the heat generation. The formula for switching losses is linear with respect to current, whereas the resistive losses are proportional to current squared as you mentioned. In our case, with the 20 ohm gate resistors in the controller and the really low Rds_on characteristic of the mosfet, the switching term greatly dominates the heat generation of the mosfet. Thus, the heat generation appears to be roughly linear with current (as calculated).

[Tweety]

Originally Posted by MPaulHolmes

Well, PCBCart doesn't take ExpressPCB anymore. Makes me mad! I downloaded KiCAD, and I'm finding it VERY awkward to work with. It was written for another language I think. It's free though. I'm really starting to become a strong believer in the old adage "you get what you pay for". The Advanced Circuits software is pretty clean and easy to learn and free, but then you have to pay 50 gagillion dollars to get the boards. Maybe I'll hawk my car and get Eagle.

Seriously, what needs do you have that the free version doesn't fulfill? I have spent hours learning Eagle... If you hit a snag, let me know what and I'll help out... Pm me and you can add me to msn...Live support...