markz
100 TW
He did fly fast on his setup, but he was always pedaling. I was just on a casual cruise that day.
Cephalotus said:
The end of BionX?
- Thread starter justin_le
- Start date
Cephalotus
10 kW
- Joined
- Jun 18, 2012
- Messages
- 755
markz said:
I can not even imagine sitting on a bicycle and not moving my feet. Why in the hell should I do that? A standard bike saddle is only comfortable if you move your legs, otherwise your butt will hurt very soon.
I assume the amount of movement people are comfortable with this is somehow different between Europe and North America.
But maybe this is a reason for the lack of success for BionX in the American market?
no it's notAWD said:
liveforphysics
100 TW
Awesomely through thermal analysis report Justin!!!
I love to pedal and do about half my cycling on pedal-only bikes and half on ebikes.
Due to terrain requiring flat pedal position for ground clearance over things, I personally am not a fan of pedal assist over throttle, because exactly when you need a big pulse of torque is often when you also need maximum ground clearance riding over an obstacle.
I like the D-series bionix and felt like it was a huge step in the right direction for hubmotor design. The moment a company chooses to make a closed system they have chosen their own ensured obsolescence through creating product that is designed not to be inclusive.
There will be only open-format, open-development, open-source companies and products and brands that some people once remembered. Hopefully now someone will run with the good design aspects of the bionix engineering efforts and leave the flaws of being a closed system so good designs have the opportunity to shine!
Are you saying the Falco guys are the same Bionix guys?
Cephalotus said:
I love to pedal and do about half my cycling on pedal-only bikes and half on ebikes.
Due to terrain requiring flat pedal position for ground clearance over things, I personally am not a fan of pedal assist over throttle, because exactly when you need a big pulse of torque is often when you also need maximum ground clearance riding over an obstacle.
I like the D-series bionix and felt like it was a huge step in the right direction for hubmotor design. The moment a company chooses to make a closed system they have chosen their own ensured obsolescence through creating product that is designed not to be inclusive.
There will be only open-format, open-development, open-source companies and products and brands that some people once remembered. Hopefully now someone will run with the good design aspects of the bionix engineering efforts and leave the flaws of being a closed system so good designs have the opportunity to shine!
Toorbough ULL-Zeveigh said:
Are you saying the Falco guys are the same Bionix guys?
kevinscargobike
1 W
- Joined
- Nov 1, 2017
- Messages
- 64
I asked elsewhere and I think it got missed - the software on the Cycle Analyst isn’t open source, is it?
Thank you Cephalotus for the detailed writeup of all your experiences good and bad. It's enlightening to hear from someone with so much firsthand usage and hacking of these systems to get the most from them. Also very cool to see the thermal image captures that really show the heat coming through right at the perimeter like that. On my own tests I had 3 individual point IR sensors and could see more of a temperature difference than on a metal shell (about 3oC), although not as pronounced as what it looks like in your photo. Were those taken after the motor had stop spinning and rested for a few minutes?
I honestly never tried in on an actual bike so I can't say what would be meant by "working well". When I did the dyno characterization I used a 6 fet Grinfineon controller in sensorless mode OK, but I had some troubles at the higher power levels with the larger 12 fet controller. For doing the wind-tunnel thermal characterization I used the BAC500+ controller, which is the precursor to the Phaserunner but with an all metal enclosure. In that case I wasn't loading the motor, just injecting field weakening current to add heat.
I was hoping to maybe borrow a high speed camera to view into the acrylic window while the motor was spinning but didn't get to that.
I'll be posting on this in the appropriate thread on motor cooling soon, but tests so far running at 120 °C for weeks and weeks in a sealed hub (MXUS in this case) have shown no reduction yet. I think most people who've had it dry out had tiny leakage in the motor side cover seal, as opposed to evaporation.
Cephalotus said:
I honestly never tried in on an actual bike so I can't say what would be meant by "working well". When I did the dyno characterization I used a 6 fet Grinfineon controller in sensorless mode OK, but I had some troubles at the higher power levels with the larger 12 fet controller. For doing the wind-tunnel thermal characterization I used the BAC500+ controller, which is the precursor to the Phaserunner but with an all metal enclosure. In that case I wasn't loading the motor, just injecting field weakening current to add heat.
It's not electronically conductive but I hope that the PCB's would have a standard conformal coating on them anyways. The reason that the test stopped at 16mL is because after that the fluid fill was so much that it started leaking out of cracks in the motor side cover from when I had disassembled it and cracked the casing while prying it open.
I was hoping to maybe borrow a high speed camera to view into the acrylic window while the motor was spinning but didn't get to that.
I'll be posting on this in the appropriate thread on motor cooling soon, but tests so far running at 120 °C for weeks and weeks in a sealed hub (MXUS in this case) have shown no reduction yet. I think most people who've had it dry out had tiny leakage in the motor side cover seal, as opposed to evaporation.
liveforphysics said:
Thanks Luke! It's kindof for naught now, although at least the principles and conclusions should translate over for other manufacturers of DD hub motors should we approach them.
goodgnus said:
Thanks sharing that experience too from the dealer side. We've never been a dealer or customer of BionX but most of the shops I've spoken with have stories like this. Various failures and issues along the way but they looked after things responsibly and the customers really liked the setup. And for shops the installation was generally seamless.
I would think that there is enough of an install base and enough brand familiarity that there would be serious interest in a firm taking over the operation and providing continuity to the product line. The atmosphere at the Vancouver Bike Show this weekend from some of the dealers and OEM users of BionX was along the lines of "I guess we'll wait and see". Apparently they are actively looking for a buyer over the next 90 days before things would go more into liquidation mode.
The only BionX bike I've had of course is gutted of the internal controls and running an external controller. We brought out the strain gauge signal from the axle sensor with plans to amplify that and hook it up to a CA3 as an example of 'opening' an obsoleted BionX system and preserving the same torque control mode. Unfortunately it's just a single gauge on the axle rather than a half bridge and so our existing strain amp circuitry wouldn't work very well and we never finished that.
A more concise report: http://www.bike-eu.com/home/nieuws/2018/3/insolvent-bionx-canada-talks-to-investors-eu-service-center-continues-10133076
" Last August BionX reported that it was to receive a forward-looking funding of CAD 3 million (€ 2m) from the Canadian government to develop a smaller, more powerful lightweight propulsion system to improve its existing e-bike systems."
Grin AnyAxle cough, cough and no CAD 3 mil funding....
" Last August BionX reported that it was to receive a forward-looking funding of CAD 3 million (€ 2m) from the Canadian government to develop a smaller, more powerful lightweight propulsion system to improve its existing e-bike systems."
Grin AnyAxle cough, cough and no CAD 3 mil funding....
wturber
1 MW
justin_le said:
A few years back we used a simple hall sensor circuit and a magnet on the lathe chuck to fire a strobe in sync with the rotation of some brass pieces that were being turned after they had been split down the center . The idea was to be able to give the impression of looking inside to show the machine tools working in a way that simply wasn't visible to the nekkid eye.
I think a similar approach might work with your window if you were to ever want to look at your Statorade working in another motor. The hall sensor was remarkably consistent in generating the signal to fire the strobe. No high speed camera required. Most standard video cameras (heck, maybe your smart phone) would probably do OK. There are issues with getting a rough sync with the video camera. You will end up with a partially exposed frames as the camera and strobe drift in and out of sync with each other. A software hack for changing frame rates on the Panasonic GH1 we used gave enough control so that we could shoot for 20-30 seconds before things would drift far enough out of sync to mess up the image. Getting long clean sections was important for our purposes, but for your purposes I'd assume that just getting a few seconds would be more than enough to see what was going on. Also, you probably have some flexibility in adjusting the speed of the wheel. We didn't have that kind of flexibility with the lathe since these machining operations were pre-programmed.
[youtube]Ea1GzQt3c5Q[/youtube]
Cephalotus said:
They're certainly nice systems, but the comparison to Apple is apt - if you couldn't repair Apple products yourself.
The older 37V systems are rock solid. I've only seen a few failed BMS boards on those, and it's almost always obvious physical damage (something shorted out and really cooked the board). And, yes, the consoles look good - won't argue there. A bit dated after 15 years, but still slick. Those systems should last a very long time.
The 48V systems are a "Sell when nearing the end of warranty" type trainwreck. They're nice, when they work, but for some reason BionX screwed up, badly, with the BMS boards, and a ton of the BMS boards fried the output transistors. It's a beautifully modular setup inside the battery pack - so it shouldn't be hard to replace the board (no soldering required), but BionX wouldn't sell battery parts, even to authorized dealers. Only full packs. I imagine they lost an awful lot of money doing that.
And, yes, they moved from German components to Chinese components. I expect that has something to do with it.
They built "throwaway systems." And most people, burned once by that, never would consider another BionX. Their solution to an I2C system with a weak battery was, "Well, we might be able to give you a loyalty discount on a brand new $2000 system..." - because they didn't believe in maintaining their older systems or keeping long term customers happy. It's a very short sighted move, IMO.
I suppose I have some of my summer projects cut out for me...
itselectric
1 kW
The real reason that BionX filed bankruptcy:
• It is not because proprietary system
• It is not because lace of market share or cost too much
The reason as the following events:
• Magna is one of larger auto parts company in Canada, owned by Frank Stronach, major partner GM, FORD, BMW, etc…
• Magna wanted get into e-car components / motor for auto industries, such as GM and Ford
• Magna purchased Bionx to gain electric motor technologies in 2008
• There was a project arrangement between Bionx and GM to develop some type electric car motor
• Bionx was unable to deliver the project on term
• GM fined BionX for 4M+ penalty for breach of contract
• BionX was unable to pay for penalty, and this is one of the major reason why BionX went under
• There are will be 60 – 90 days to find a new buyer/investor, until then everything up in the air
This is almost a bit like repeat story when Lee Iacocca wanted sell e-bike in the early 1997 through car dealership and WaveCrest in 2006 wanted to get into e-car market by entering into e-bike market first
• It is not because proprietary system
• It is not because lace of market share or cost too much
The reason as the following events:
• Magna is one of larger auto parts company in Canada, owned by Frank Stronach, major partner GM, FORD, BMW, etc…
• Magna wanted get into e-car components / motor for auto industries, such as GM and Ford
• Magna purchased Bionx to gain electric motor technologies in 2008
• There was a project arrangement between Bionx and GM to develop some type electric car motor
• Bionx was unable to deliver the project on term
• GM fined BionX for 4M+ penalty for breach of contract
• BionX was unable to pay for penalty, and this is one of the major reason why BionX went under
• There are will be 60 – 90 days to find a new buyer/investor, until then everything up in the air
This is almost a bit like repeat story when Lee Iacocca wanted sell e-bike in the early 1997 through car dealership and WaveCrest in 2006 wanted to get into e-car market by entering into e-bike market first
itselectric said:
Thanks, that's the first I've heard of this particular detail and it's certainly a different insight. It's funny because when news broke of Magna taking over BionX, we all wondered what great things their automotive expertise might bring to the ebike market. And instead, they were looking to see what great things this acquired ebike expertise might bring to their automotive market. That's a funny twist if it's true!
I agree though that the fact that the kit was proprietary or on the expensive side wouldn't at all be reasons for this development. Those are reasons why DIY users would be disinclined to want the system, but play almost no role in the commercial viability as a consumer product sold through bike stores. I don't imagine many customers asking the bike shop "oh, but is it proprietary? Can I upgrade the motor controller or replace the battery with another brand?"
kevinscargobike
1 W
- Joined
- Nov 1, 2017
- Messages
- 64
justin_le said:
I think that question gets more relevant to people once their first bike is 4 or 5 years old and the battery is dead.
I will say though, this world of separate bike components is really really hard to penetrate, even if you’re technically inclined. The time between “there’s such a thing as a bottom bracket torque sensor” to actually understanding enough to order a system was very long for me.
Odd as it sounds, it was the Grin factory tour video that electricbikereview.com did that finally started to build the whole picture for me. I’m loathe to admit it, but that kind of marketing really is valuable. I’d love to see Grin do some YouTube videos that go through what a kit is like on a bike, and have as much explanation and demo as the fun torque sensing unicycle video.
Also, as a programmer and not an EE or ME, I’d love it if the Cycle Analyst was open source, so that I could do interesting things at the software level, since I don’t have the skills or inclination to do interesting things electrically or mechanically. Non-proprietary batteries are great, but non-proprietary software is even better, and is very much the un-Bionx thing to do.
I'd guess the main reason the CA is not open-source is that if it was, the first thing that would happen is some other companies around the world would immediately take that code and make their own half-price versions of the CA, and most people would no longer buy the "real" CA.
After that, Grin Tech would no longer have the money coming in to fund further CA development, or make and sell them themselves, ending the whole project except for whatever individuals continued to do.
So as much as I'd love to see it open-source (there's at least a handful of things I want to be just *slightly* different than they are), I can understand why it isn't.
After that, Grin Tech would no longer have the money coming in to fund further CA development, or make and sell them themselves, ending the whole project except for whatever individuals continued to do.
So as much as I'd love to see it open-source (there's at least a handful of things I want to be just *slightly* different than they are), I can understand why it isn't.
Syonyk said:
That's another thing that BionX was a bit ahead of the game one. They were one of the very first ebike companies using quality 18650 cells for pack assembly. I remember the first time we had one in our shop for some reason and noticed this. Like most people we thought that was silly and wasteful. Why have like 40 small laptop cells all tabbed together when you could have 10 large capacity EV cells? All the other ebike companies at the time had either LiPo pouch cells or prismatic LiFePO4 in a 1P pack layout, which surely made more sense. Boy did time prove us wrong there.
Those early BionX batteries also didn't have ANY cell by by cell voltage monitoring on them, just pack level voltage monitoring. And yet the cells were so precisely balanced and matched that they never had any balancing issues or need for per-cell voltage cutoffs. That also made them much more robust against BMS circuit failures which in those days was a huge source of battery troubleshooting issues.
So the early BionX lithium packs to their credit lasted much longer and with much higher reliability than anything we had been dealing with. This is in the 2006-2010 period. I remember at one point visiting JVBike who is the main BionX dealer in Vancouver and he was lamenting that the BionX batteries now had a full cell level voltage sensing and he was rolling his eyes about what this was going to mean for reliability. And sure enough endless headaches and battery return issues as a result of the more conventional BMS board with all the consequences of high quiescent current drain killing cells in storage and that.
No kidding there. At the very least if a company is going to discontinue supporting a proprietary interface themselves, they should disclose the spec standard so that 3rd parties could step in and take over the ongoing parts supply. I can't imagine how many I2C based BionX kits with perfectly good motors and controllers on perfectly good bikes will wind up being mothballed for lack of any replacement battery as the primary one wears out.
I take it you battery rebuilds or repairs for BionX kits? Is this on your own fleet or for other people?
kevinscargobike said:
Hey Kevin, this isn't really the correct thread for this, but we're talking firmware (15,000+ lines of 100% assembly code for a dated 8-bit micro) and not software or arduino coding. Send me an email if you are actually an experienced embedded firmware programmer familiar with the mid-range PIC series and want to try stuff, but it's off topic to this thread here.
kevinscargobike
1 W
- Joined
- Nov 1, 2017
- Messages
- 64
Sorry for pushing the issue. I can certainly see how there would be plenty of downsides and few upsides...and nope, I don't personally want to see any assembly code.
justin_le said:
Heh, I would laugh if it wasn't so sad. This literally never happens. Not in our industry, nor in any other transportation-related one, nor computer software, or video games, cameras or any industry. The tendency to take your "intellectual property" with you into your grave is very high. It's very unfortunate.
Yes, this is about to happen with the company that makes the music editing software I use (Cakewalk), since Gibson is shutting them down rather than selling them off. 
All their software will likely become useless when that happens, as when their activation servers are shutdown then no reinstallation of the software will work anymore (well, it'll work for 30 days, then you get to wipe your harddisk and reinstall your OS so you can then reinstall the software, plugins, etc etc because it leaves stuff behind to prevent you reinstalling the "demo" again and again). (yes, there's ways around that, too, but not for the average user that doesn't back stuff up, etc...and some of the people at Cakewalk are working on a way around this, but I doubt Gibson will let it happen).
Sorry for the OT....
All their software will likely become useless when that happens, as when their activation servers are shutdown then no reinstallation of the software will work anymore (well, it'll work for 30 days, then you get to wipe your harddisk and reinstall your OS so you can then reinstall the software, plugins, etc etc because it leaves stuff behind to prevent you reinstalling the "demo" again and again). (yes, there's ways around that, too, but not for the average user that doesn't back stuff up, etc...and some of the people at Cakewalk are working on a way around this, but I doubt Gibson will let it happen).Sorry for the OT....
justin_le said:
Correct, though it's not just the cell balancing and matching. It's been covered here on ES, and I did some analysis as well (https://syonyk.blogspot.com/2016/12/a-tale-of-3-bionx-packs-self-balancing.html) - but the spinel LiMn chemistry they use is "self-balancing enough" (search for Konion on ES if you want more details) to get away with this. The charge efficiency drops a bit as they get nearly full, and evidently this works quite nicely in practice. The only time I've seen voltage horror shows on a BionX pack is if the pack is fully drained (that's my one complaint about the old BMS boards - they have enough of an idle draw to kill a pack in a year or two if it's not charged). The new boards supposedly go into a deep sleep mode, but that's mostly irrelevant as the boards fail regularly and the cells don't seem to be lasting very long anyway.
The "good news," at least on that front, is that the I2C systems (and I think some of the Canbus systems) can be hooked up without the fancy BMS. And, at least on some, can take a rather nice voltage bump.
I think you can feed about 50-60V into the old systems without tripping overvoltage cutoff. You still need the low voltage output for the console, but it will work (just without a battery gauge).Yeah, I do BionX battery rebuilds for most of North America. Not officially, and there's another guy in Ohio who does some, but if you search for BionX battery pack data, I'm the only guy who has torn a bunch of them apart (and extensively documented the process). I've been working on non-BionX projects lately (I can tell you a lot about moving rock with 75 year old tractors), but if BionX is down, I should probably bump the priority on getting a few research bikes put together (I've got... uh... maybe 3 or 5 old systems in various stages of dead laying around) and work on BionX Longevity Projects.
It pisses me off to no end that they go on about how amazingly environmentally friendly their systems are when they build throwaway systems that cannot be repaired or maintained.
wturber
1 MW
Syonyk said:
I suspect that the majority of, or at least a very good sized chunk of spending that goes towards being environmentally friendly is "feel good" spending with minimal consideration for the net impact of the purchase. One of the most environmentally friendly things you can do is to simply not buy something new, but instead keep something old in service. But it is often harder (though not impossible) to make a lot of money doing that.
wturber said:
Oh, certainly. I have this complaint about Tesla as well - they make unmaintainable cars that will be scrap in relatively few years, simply because people can't fix them. It's part of why I put in the effort to support the older BionX systems (and eventually other ebikes as I have time to design rebuilds) - replacing the entire system when all it needs are new cells is beyond idiotic, but it's sure what BionX wants.
As much as I make good money providing aftermarket support for their systems, I really wish they would support the systems directly. A $2k "sustainable transportation" device that's scrap in 5 years is pretty bad. Probably worse than just putting miles on a car, once you figure that they source all their stuff from China now.
liveforphysics
100 TW
All maintenance/repair work I've done on Teslas went more smoothly and was cleaner and easier to do than any ICE car I've worked on. People love to say how they can't work on new vehicles, but myself and many mechanic friends find newer vehicles to be easier to fix/hack/hotrod.
A bionix setup shouldn't be overly difficult to sniff the CAN or I2C lines and at least capture the handshake and/or whatever commands are needed to get the setup to function. Or gut the proprietary controller out of the motor and wire up something that works and/or delivers hotrod power.
A bionix setup shouldn't be overly difficult to sniff the CAN or I2C lines and at least capture the handshake and/or whatever commands are needed to get the setup to function. Or gut the proprietary controller out of the motor and wire up something that works and/or delivers hotrod power.
Syonyk said:
liveforphysics said:
Sure, people I know with Teslas find maintenance seamless because you don't do it. You can't get the service manuals, you can't use an independent shop, and even if you can get the parts, you can't get the software to pair them. So... I mean, if you completely root it, sure, but that's not practical or reasonable for most people.
A bionix setup shouldn't be overly difficult to sniff the CAN or I2C lines and at least capture the handshake and/or whatever commands are needed to get the setup to function. Or gut the proprietary controller out of the motor and wire up something that works and/or delivers hotrod power.
It's not particularly difficult to sniff, and half the work is done, but there are quite a few non-understood commands in the protocol, and it's still needlessly complex compared to a system that requires "Volts & Amps" from the battery to run.
As far as gutting the controller, the vast majority of BionX owners, at least that I deal with, are older and use the assist to let them continue bicycling even if they have a weak knee or hip ("that one hill" comes up a lot if I ask). They're not hotrod systems, and that's not the market for their systems. Though I will say, their "250W" motors seem to put out an awful lot more than 250W...
