MitchJi
10 MW
Hi,
Justin was kind enough to answer a couple of questions via email, and give me permission to post his replies:
http://endless-sphere.com/forums/viewtopic.php?f=6&t=49583&p=739436&#p738781
Justin was kind enough to answer a couple of questions via email, and give me permission to post his replies:
http://endless-sphere.com/forums/viewtopic.php?f=6&t=49583&p=739436&#p738781
justin_le said:The past few days have been really awesome but before going into that I'll first want to finish the project build details.
One of the more exciting pieces of gear being field tested is a high power modded 6-mosfet sinusoidal field oriented motor controller (FOC) that we've been developing with Accelerated Systems of Waterloo, Ontario. They have a nicely compact stock controller in production called the BAC 500 that we were looking to carry as a sophisticated higher-end ebike controller. FOC controllers have current sensors on the phase leads rather than a shunt on the overall battery current, and as a result they can run 3-phase hub motors super smooth by controlling the individual phase currents for almost no torque ripple. Plus it has proportional regenerative braking, dual sensored and sensorless operating modes etc. There are hundreds of parameters available to tweak through ASI's "bacdoor" software which made it a dream project controller to play with.
However, in our testing they would get hot very fast at the power levels that we tend to deal with. Just a minute or two before hitting controller thermal rollback. Part of that was due to the the sinusoidal drive having more switching losses than a traditional trapezoidal drive, but the mosfets were also high resistance 10mOhm devices, the internal bus capacitors were pretty minimal too and generated heat from the ripple currents. And finally, the tiny enclosures provided little heat capacity to absorb energy and relatively little surface area from which to get rid of it. So for ~250W setups they'd be fine, but not the 500-1000W range we needed to run.
So we had ASI do a board revision with 2.5 mOhm 100V mosfets and modified gate drive circuitry, and promised we would look after the cable harness and enclosure details. For the enclosure design, we wanted to make room for additional bus capacitance, a precision shunt for the Cycle Analyst connector, and most importantly come up with a form factor that would fit nicely to the tubes of a bicycle. We also wanted it to be fairly manufacturable by CNC so that it would be conceivable to do small scale production runs in-house if this really worked out
Here it is with the ASI controller circuitry inside. You can see the extra room on the right which provides room to fit an external bus capacitor and the CA shunt (neither of which were wired up yet in this photo):
And the finished lid we also wanted to seal tight around the cable ports so that we could make it all watertight. The white LED hole at the top is also a breathable membrane vent, so that as the controller heats and cools the pressures can equalize without risk of drawing water through the wiring.
And on the underside, we had a channel machined down the middle so that it self-aligns when zip tied to any tubular object, allowing it to attach easily anywhere on the bike frame:
------------------------------------------------------------------------------------
nicobie said:That thing is tiny! Any idea of the watts your version can handle?
I've got mine set to 60A phase current limit and 40A battery current limit, so it's able to do 1400-1600 watts off the line OK, but on any substantial hill it will hit thermal rollback towards the top and drop back to more like 1000 watts. It really depends a lot on the extent of air cooling available. We were initially doing all of our testing with these on our bench dynamo with the controller sitting in still air, and kept being disappointed by the results even with all the power tweaks with better fets etc. But once we put a fan nearby for even a modest 10-15 kph steady airflow over the controller the thermal dissipation increased by nearly a factor of 4. There is a tendency for people to want to tuck their controller inside a bag or chassis or somewhere out of site, so a key goal in our enclosure design was to make it look attractive with inline wiring, so you are more inclined to mount it easily on the frame of the bike where it is most likely to get exposure to airflow while moving. Here's where mine is at under the downtube, although mounted right on the front stem would be most ideal:
Mitch-email-to-Justin said:When do you expect this Controller to be available?
What are the chances of you producing a version that can handle something like 36-80V, and something like 60-100A sustained? If you don't intend to do that how hard would it be to do that ...?
Justin-reply-via-email said:...mostly it just has to do with time and resources and trying to do too many projects at once so we've shelved that to put focus on a few other things that were further along.
That would just require a 12 mosfet model. We've actually been sitting on the 50pcs of 6-mosfet high power modded BAC PCB's for 2 months now but have been too backlogged with other projects to machine the enclosures for them, but once we do then we'll see what are plans are with it.
We can also get sinusoidal firmware on the infineon controllers for a lot cheaper, but that's not the same as a true FOC control and they don't run in a sensorless mode that way, while the ASI units work great sensored and sensorless even at very high eRPMs once they are tuned appropriately.