Compact Field Oriented Controller, ASI + Grin, limited run

[justin_le]

How about making bottom plate a bit thicker (if its not thick enough) to drill out couple 6mm holes (for DIY modders) for water loop,

We're trying to make everything as small and thin as possible, as you can always bolt on extra metal if you want more mass and heat dissipation but there is no need to impose that on everyone by including a thicker heatsink in the design. That said, we do have a pair of holes in the design model which we didn't drill out in the above prototype. These go right through either end of the heatsink block and are intended for cable ties in order to quickly affix the controller to a bike tube. There's nothing stopping someone from instead jamming a copper pipe through this to invoke some liquid cooling, or possibly threading the hole and attaching pipe fittings.

I like the anodised heatsinks on both sides of the fet. But i wounder how will you keep the screw from conducting to the mosfet housing ? its very easy for the screw thread to break through the anodiseing hence the screw will be alive and hence if it touches the inside of the fet hole it will conduct. Normal to220 bushes won't fit unless their is room above the fet?

Here's the cutaway view showing how that is done. The insulating shoulder washer is on the bottom piece and goes right through the mosfet body, so the screw can't ever touch the fet.

Is it necessary to somehow isolate the USB connection that goes from a BAC500+ to a computer USB port?

If you are powering up the controller from a battery pack, then no need at all since the battery is floating. But if you are using a power supply to power up the controller when you have it computer connected, then it is possible to have grounding issue if the power supply output isn't isolated. And it's possible (known from experience!) to fry the communications port if the power supply is sitting at some high voltage relative to the computer when you first plug in the USB cable. Best practice would be to either have the BAC controller powered from a battery pack, or use a laptop computer which is not plugged into the wall, and then you'll never have to worry.

 

[madin88]

thats a well thought out, awesome compact design^^

 

[agniusm]

Makes more sense now. When these will be available? Any time schedule?

 

[justin_le]

When these will be available? Any time schedule?

Time schedules are for giant companies with lots of experience in a familiar domain where it is possible to forecast and plan. Just follow any kickstarter campaign if you want to see the pointlessness, futility, and broken promises that accompany release time schedules for new products developed by small groups with tons of stuff on their plate and unknowns to sort out

It'll be ready when and if it is ready, and I'll keep the interesting development updates posted here for those who'd like to follow.

 

[riba2233]

Well said

 

[agniusm]

I thought Grin grew to something bigger over this time

 

[flathill]

Nice little controller but I think you are making a big mistake by potting the whole thing is epoxy.

I totally understand, technically, this is the most robust solution as far as vibration and moisture protection go, but I still think it is a design cop-out as now you just made a throw away brick. You can't salvage parts from a broken controller. You can't repair a broken controller. You can't hack a controller.

You are propagating everything that is wrong with today's wasteful culture. You are simply taking the easy route which is only the best route in the short term and for you bottom line. Rant off.

 

[MrDude_1]

Nice little controller but I think you are making a big mistake by potting the whole thing is epoxy.

I totally understand, technically, this is the most robust solution as far as vibration and moisture protection go, but I still think it is a design cop-out as now you just made a throw away brick. You can't salvage parts from a broken controller. You can't repair a broken controller. You can't hack a controller.

You are propagating everything that is wrong with today's wasteful culture. You are simply taking the easy route which is only the best route in the short term and for you bottom line. Rant off.

Clearly it is YOU with the issue... And your issue is, You don't know how to remove potting epoxy. :lol:
Seriously, don't propagate today's wasteful culture just because its hard to remove... It can still be done to fix or mod a controller... but If you want to do that, I would just ask them for one that wasn't potted yet.

 

[speedmd]

Sometime back I recall seeing a study that compared reliability of thick soft conformable coating to solid potted assemblies that showed the coated ones just as reliable in life testing and field work trials. Have not kept up with the trends of late. Imagine that heat transfer is another story.

 

[riba2233]

Nice little controller but I think you are making a big mistake by potting the whole thing is epoxy.

I totally understand, technically, this is the most robust solution as far as vibration and moisture protection go, but I still think it is a design cop-out as now you just made a throw away brick. You can't salvage parts from a broken controller. You can't repair a broken controller. You can't hack a controller.

You are propagating everything that is wrong with today's wasteful culture. You are simply taking the easy route which is only the best route in the short term and for you bottom line. Rant off.

Well maybe it would be wasteful to let controller get destroyed by moisture or sea salt in the air? Any way, if controller is reliable, and I believe it is, epoxy is a great way to keep it out from bad outside conditions.

 

[flathill]

I know epoxy is the sure way to solve the moisture/vibration problem, that is not my argument. The harder path is to never let the controller be destroyed by moisture or sea salt in the air, while still being repairable/salvageable.

An analogy would be welding your hub motor axle to your bicycle frame to make sure you hub motor does not come loose. Sure that is the best/safest/robustest way to do it but that is not the point.

Shoot, even if this controller enclosure used a hermetically sealed enclosure, that would be better than potting it in epoxy. At least you could cut/force the enclosure open. Satellites use potted hermetically sealed electronic packages all over the place, but at systems systems level the electronics are hermetically sealed but not potted. If you want to consider a controller a throw away electronics "package", that is fine by me, I just don't agree with you.

A better solution would be to make the wiring pass through seal potted in epoxy, and use a softer removable material to pot the electronics inside. But then you would increase costs and need an enclosure. This design is cheaper and easier, but not the best in the long-term.

As a bonus soft silicone potting would improve reliability as it is much more thermally conductive than epoxy. A thin metal enclosure would also simply look better while again providing better cooling as a bonus.

"LORD Thermoset SC-320 thermally conductive silicone encapsulant is a two-component system designed to provide excellent thermal conductivity for electrical/electronic encapsulating applications, while retaining desirable properties associated with silicones."

 

[okashira]

I think this would fit the bill and be removable in a pinch
http://www.masterbond.com/newsrelease/super-gel-9

 

[liveforphysics]

I know epoxy is the sure way to solve the moisture/vibration problem, that is not my argument. The harder path is to never let the controller be destroyed by moisture or sea salt in the air, while still being repairable/salvageable.

An analogy would be welding your hub motor axle to your bicycle frame to make sure you hub motor does not come loose. Sure that is the best/safest/robustest way to do it but that is not the point.

Shoot, even if this controller enclosure used a hermetically sealed enclosure, that would be better than potting it in epoxy. At least you could cut/force the enclosure open. Satellites use potted hermetically sealed electronic packages all over the place, but at systems systems level the electronics are hermetically sealed but not potted. If you want to consider a controller a throw away electronics "package", that is fine by me, I just don't agree with you.

A better solution would be to make the wiring pass through seal potted in epoxy, and use a softer removable material to pot the electronics inside. But then you would increase costs and need an enclosure. This design is cheaper and easier, but not the best in the long-term.

As a bonus soft silicone potting would improve reliability as it is much more thermally conductive than epoxy. A thin metal enclosure would also simply look better while again providing better cooling as a bonus.

"LORD Thermoset SC-320 thermally conductive silicone encapsulant is a two-component system designed to provide excellent thermal conductivity for electrical/electronic encapsulating applications, while retaining desirable properties associated with silicones."

A hermetic seal is cost-prohibitive for a controller, and just the cost of the hermetic pass-through connectors would exceed the controller cost.

Properly ruggidized electronics are potted. Properly designed electronics have an extremely low failure rate, particularly when well potted to defend against vibration and humidity.

 

[flathill]

You are missing the point. With the cast enclosure the design is limited to hard epoxy which is almost impossible to remove. The hard epoxy serves also serves as the enclosure shell

I fully support potting the electronics, but use either a silicone potting compounds that cant be dissolved or using a soft hybrid epoxy than can be cut and peeled away. These requires a separate hard shell which increases cost. The cast design seems clever but it is really a design cop-out

 

[justin_le]

Nice little controller but I think you are making a big mistake by potting the whole thing is epoxy.

I totally understand, technically, this is the most robust solution as far as vibration and moisture protection go, but I still think it is a design cop-out as now you just made a throw away brick. You can't salvage parts from a broken controller. You can't repair a broken controller. You can't hack a controller.

Well maybe it would be wasteful to let controller get destroyed by moisture or sea salt in the air? Any way, if controller is reliable, and I believe it is, epoxy is a great way to keep it out from bad outside conditions.

Hey Flathill, I sympathize with where you are coming from, but suggest it is premature to say this is a "big mistake" :wink: . Potting isn't exactly a cop-out, it has more than its share of challenges as well. But yeah making an openable and serviceable controller that ALSO has the same level of vibration resistance, environmental ingress resistance, and thermal dissipation capabilties as a potted unit is indeed a massive undertaking. This is what we tried to do in the first run of 50pcs on the start of the thread, and arguably it was a bit of a failure. Even with the gore-vent to alleviate pressure differentials they can still wind up with water ingress, there's little chance they would stand up to the vibrations of being on a skateboard say, and there was absolutely no way for us assemble/manufacture them in this method and come close to breaking even.

To your points about repairing and hacking, the goal is to make a part that requires neither of those. People only hack controllers because lots "economic decisions" in the component selection and assembly of china supplied stock devices so there is easy room for improvement. If a device is already using the top of the line mofsets and high ripple bus capacitors, then there is nothing left to hack from the hardware side that would provide gains. And in terms of repairs, I've personally dealt with 100's and 100's of failed motor controllers over the years and am sick of it. At least 95% of those fall into 3 categories:
a) Water ingress / corrosion damage
b) Vibration damage (wires breaking off the PCB, capacitor legs breaking, mosfet screws scoming loose, )
c) Mosfet failures.

A and B would have never happened had the controllers been potted, and C can be almost completely eliminated with properly robust power electronics engineering. You won't see mosfets randomly fail on a controller that has thermal rollback, pulse by pulse current limiting, a finely tuned gate driver circuit etc. You can take one of these controllers running at full power and then suddenly short all the phase wires together and the device wouldn't suffer a blip. You can have it doing regen at full current then yank out the battery pack and before the voltage has a chance to rise up and damage things all the regen will cease.

Even with a, b, and c, taken care of, there inevitably will still be a few random mystery failures that result in pretty looking bricks. But if we can get this down to << 1%, ideally <0.1%, then overall people will be much more content and less troubled than the current situation of controllers with high single digit failure rates, even if those failures are sometimes repairable.

Properly ruggidized electronics are potted. Properly designed electronics have an extremely low failure rate, particularly when well potted to defend against vibration and humidity.

Precisely. We can learn a lot from automotive here, since that is one application where vibration and exposure are similar to what an ebike can see. In that world, almost everything is potted.

 

[flathill]

Again I support potting, just not the cast shell design which requires a hard epoxy in this particular application. I don't mean it is a big mistake in the sense it is a bad design that will fail. I just mean you are propagating everything that is wrong with today's throw away culture. You can easily add a hard shell which will enable you to use a softer potting compound that can be removed for repair or to salvage parts. It is not like capacitors last forever.
-
With nature or against
U have two choose

 

[amberwolf]

FWIW, I'm probably one of the most likely to hack, mod, fix, etc., various stuff, and I don't see potting as a problem that makes things throwaway.

It's relatively easy to remove many types of harder potting...and it is actually *harder* to remove many of the softer pottings I've run across, as most seem to be silicone-based, and don't dissolve but rather have to be scratched or picked away, risking damage to small components, wires, coil insulations, etc.

I've had quite a few things I could've fixed if the potting was hard, but since it was soft and not soluble in anything I tried, I had to scratch and pick at it. In the process I usually broke things I could *not* fix that had nothng to do with it's original problem (most likely bad capacitors in a number of things, sometimes I suspect blown fuses). If it'd been the normal hard pottings I could've dissolved it away (brake fluid works on quite a few types, as an example; acetone works on some others, etc), and then replaced the necessary parts, and continued to use it (repotting with something if necessary).


If there's an easily-soluble soft potting compound that won't eat components, markings, etc., and the solvent also won't eat them, then go for the soft stuff if you like.


So...given the failure modes usually seen (as Justin points out), I'd say potting this level of quality of device is probably fine.

Heck, if it weren't for the fact that the screen on the CA prevents it from being an easily-pottable module, the CA itself is a really good candidate for potting!

 

[flathill]

Precisely. Some epoxies are easier to remove than other. S

But if you use really soft and sponge e silicone you can simply buy silicone cleaner from any hardware store and dissolve it. You can dissolve epoxy also but it damages pcbs if you dont know the antidote. On some glassy types of epoxy nothing but heat works. At the same time the best potting silicone is an almost order of magnitude more conductive to heat than any epoxy or epoxy hybrid suitable for potting electronics

 

[justin_le]

you are propagating everything that is wrong with today's throw away culture.

Here's the proper way to judge that. We do a follow up 5 years from now on everyone who purchased one of these potted controllers and on everyone who purchased a generic but serviceable china controller. Then we see how many from each camp have 'thrown away' their device. That will tell you who here is promoting a disposal parts culture, and I think you know how it would play out.

That said:

You can easily add a hard shell which will enable you to use a softer potting compound

This is likely the longer term plan, but not for the reasons you suggest it. It's vastly easier from a manufacturing/potting perspective to pot into a shell than it is to do a removable mold. But we're doing a run of 50 pieces. The tooling for a rigid plastic shell is about $8,000-12,000. You can see the issue I hope.

 

[justin_le]

I thought Grin grew to something bigger over this time

[youtube]5IaNaQHjIRE[/youtube]

 

[flathill]

you are propagating everything that is wrong with today's throw away culture.

Here's the proper way to judge that. We do a follow up 5 years from now on everyone who purchased one of these potted controllers and on everyone who purchased a generic but serviceable china controller. Then we see how many from each camp have 'thrown away' their device. That will tell you who here is promoting a disposal parts culture, and I think you know how it would play out.

That said:

You can easily add a hard shell which will enable you to use a softer potting compound

This is likely the longer term plan, but not for the reasons you suggest it. It's vastly easier from a manufacturing/potting perspective to pot into a shell than it is to do a removable mold. But we're doing a run of 50 pieces. The tooling for a rigid plastic shell is about $8,000-12,000. You can see the issue I hope.

I'm right with you there buddy. I'm just throwing out a little hyperbole to push you towards Design for Recycling/Reuse. I often find this is not on the minds of many designers/engineers or is last on their list of concerns. Most electronics are just chopped up and then the precious metals are recovered while the rest turns into a black slag. At least in China they recover the chips and components and reuse them. The problem is most are then resold as new and most use very crude methods to extract the components. A few do recover the high value components in an environmentally sound way and resell them as used. These little markets that sell used components for dirt cheap are a DIY dream. Even though electronics recycling has picked up still over 80% ends goes straight into the land (fill). Rant off. Continue on as you were.

 

[Alan B]

I think there is room in the marketplace for both the highly reliable potted electronics and the budget hackable and repairable electronics. I'm looking forward to small reliable controllers.

The potted Meanwell LED power supplies that make such good battery chargers are potted, and come with a 5 year warranty as well as the various ratings including UL. The unpotted budget priced models have much shorter warranties.

I have now experienced both the reliable sine wave controllers, and the low cost hackable and repairable controllers, and I have to say that for many projects I just want something that always works. No hacking or modding required. Every adjustment necessary is in the parameters. The FOC design requires fast current sensors and this facilitates the ability of the controller to detect faults before damage is done and shut down safely. Many motor controllers we use at work are fed with 208V 3 phase power, and they never blow up. When something goes wrong they shut down and display a fault code. If they did not, the failure could be spectacular with the power that is available there.

I look forward to the BAC800. It looks like a very compact and efficient design in the mechanical sense, and I suspect it is even more so in the electronics and firmware.

 

[Arlo1]

Flathill Justin has it right. They are not a throw away as they if designed properly will last 20years +. This is what I have been learning. And since adding proper gate drive control I have not wrecked a single power switch. As Justin said you can do anything you want and not kill the controller. There is no way to increase power as they are run to the limit with thermal cutbacks. Over voltage under voltage protection and all kinds of other awesome stuff. Good job Justin I love this design its the best bang for buck at its size on the market and its going to last a long time!

 

[liveforphysics]

Greatest waste comes from products designed to fail.

Products designed to survive using every design advantage to survive, including embracing the awesome weather/vibration ruggedness advantages of potting will be least wasteful.

 

[spinningmagnets]

I'll have to agree with Justin on this one. There are guys around here who end up needing to buy a new cheap China controller every year. I'd rather have one unfixable/un-cannabalize-able controller that has a 99.9% long life success rate. Failure points?

1. Sky looked good when I went out, started raining before I got home...controller got wet, doesn't work any more.
2. Hit an unexpected pothole very hard, almost wrecked, had to pedal home, controller no longer works.
3. Controller acted janky, opened it up and one FET had a poor connection to the thin heat-sink, and overheated.

I don't anticipate that this controller will "take over" and no other controllers will be able to survive against it, so...if you like another controller, get the other controller. As for me? I'm looking forward to buying one of these.