Jeremy Harris wrote:I'm already working on a controller, Luke! I've been doing a lot of work recently on building a small, high current, controller. One thing I discovered, when looking through dozens of FET spec sheets, is that there are some surprises, as a big package isn't necessarily the best. Some may have spotted that getting heat out of the FET junction is the big issue that limits the current a single FET can handle. There's nothing much that can be done about the internal thermal resistance, and this tends to be similar for pretty much any package. The big packages offer the means to get heat out of the package more easily, but this can be done another way, I think.
I've found this FET: Yes, it's a small, surface mount package. BUT, look at the ratings and then think about this as a mounting option:
What if the power rails and the phase connections were solid copper bars, connected to internal finned heatsinks?
What if the FETs were soldered directly to the copper bars, providing not only a good electrical connection, but also a near-zero thermal resistance (no insulators needed)?
What if the controller had a small internal fan to keep the "live" heatsinks cool?
I'm thinking of a 12 FET power "board" using these FETs, driven by a 6 FET controller board. I reckon such a controller would comfortably handle well over 200 amps. In fact, in might well do around 150 amps on just 6 FETs. The downside would be the 75V FET Vds limit, but as the motor is rated at 60V max this shouldn't be too big and issue.
Hi Jeremy - Your ideas posted here remind me how people with a similar goal often end up going down the same windy road!
I had looked into (and even bought a few) FETs using the same package. Having the source current go through 5 pins instead of one would indeed remove one of the bottlenecks of these devices. However, sometime latter I decided to stick with the TO-220's, as I found that it was too tricky to design a easy to assemble copper-bar based power stage using these surface mount D2paks.
As for soldering FETs to thick copper bars without stressing them too much *and* not pulling your hair out - good luck. I've decided to go solderless, but I still do use the tabs as an electrical drain connection anyways... it works very well, and I would not go back to solder.
This route led me to my last protoype, which I tested (months ago) at 200A per phase continous for a 12FET power stage (3006's @ 50V input, 60%duty). It was actually the 8AWG output wires that went over 100oC that was the limiting factor, specially the one doing PWM which heated up significantly more than the wire on the phase being held low. Needless to say, I was *very* happy about the results of this last design!
Since then though I have had many distractions and have not touched the controller, and was also discouraged a while by a potential small outfit interested in finishing the controller development together, but backing off after much discussion due to financial constraints (and it's the second time I've wasted time/energy negociating and having the other party back down afterwards). I guess my lesson is to do it myself, and just have fun with the technical stuff I really like doing instead of wasting energy on negociating association conditions and the likes. Sorry for this being off-topic BTW, but Jeremy's controller post caught my attention!
All my ebikes have gone up in flames (with my whole house) --> Many lessons learned (like that insurance companies can be cold hearted a$$holes; and to read and understand your contract before your house burns down...)