blaen99

iTrader: (6)

Joe's driving a fixie now? wtf Joe, you are supposed to troll the hipsters, not join them! /notsrs.jpg

Joe Perez

iTrader: (8)

Bryce

iTrader: (24)

Well there's your problem.

triple88a

iTrader: (2)

I don't know, I don't think that's suppose to be burnt like that.

rleete

iTrader: (21)

Why you let the magic smoke out?

Man, I was just looking at making something like this. I figured you already did all the hard stuff, I just gotta copy that. Now I have to wait for revisions!

shuiend

iTrader: (24)

It seems to me that electronics are trolling Joe.

Joe Perez

iTrader: (8)

Yes, I believe they are.

Thing is, this was supposed to have been a happy post. I was going to put up this image:



and say "look, I found a friend at Home Depot this afternoon," and then you'd all make gay jokes about how we probably met in the tool section, or while looking for a hardened rod, or something equally silly. And I'd have a witty comeback, and we'd all rejoice in the fact that enlightenment is spreading through the land.


(It is also ironic that, despite all of the public EV charging stations which we have here in SoCal, this is the first time I have ever seen anyone actually charging an EV in public, and not even at an officially-sanctioned charging station at that. If you look at the curb just behind his front fork, you can see that he's brought his charger and has plugged it into an electrical outlet that rather conveniently happens to be near the bike rack.)


Unfortunately, about 10 minutes after that picture was taken, the BMS in the top of the battery smoked. Again. And lemme tell you, a 200 lb fatass pumping a 90 lb bike uphill with a dead electric motor dragging on the rear axle (remember, brushless DC motors try to act like generators when they're off) kinda sucks.

This battery has been plagued with problems. I think I documented the fuse issue a few pages back, and I can deal with changing those roadside, but this is now the second time in 9 months that the battery's internal circuitry has simply exploded while riding. I've been keeping in contact with the company regularly each time I have an issue, and they always say "we're investigating a fix" or "we're working on a solution" or something like that. They're extremely courteous and responsive, it's just that it's obvious they're not electrical engineers.

Last night I sent the picture, along with a brief description, to the owner and asked him to call me in the morning. He sent an email response within about 30 minutes which said, essentially "These batteries just don't work at 22 amps. I suggest you send the battery back to us for repair, and we'll also swap out your 22 amp controller for a 15 amp unit."

WTF?

First off, 22 amps is child's play for a 10Ah LiMn battery. The problem isn't the battery itself, it's the stupid circuit you designed to regulate it.

Second, you've been selling this exact kit, with this exact battery and this exact 22A controller, as a complete turnkey package, for two years now. And it's not like I selected the "hot rod" package. You only sell one rear-wheel, direct-drive lithium kit, and this is it.

Third, I live in SoCal. You live in SoCal. We have hills and mountains here, which you obviously know, as your marketing literature is mellifluous in its praise of how your EV drive kit "smooths out the hills" and how "I'm in charge, to work the hills myself, or let the bike work them for me." At 22 amps, this thing is just barely adequate for the 15% grade that I have to climb to get to my apartment if I'm coming from the south, so what makes you think that decreasing the power of the system by more than 30% is going to make me happy?


Basically, I think we're at the point where I'm going to say that this would be a good time for them to generously offer me a full refund. And if he balks, I may expound on how this would be a good time for them to generously offer me a full refund, in exchange for which I will generously shut up and disappear, as opposed to describing in considerable detail the many problems I have had all over the various DIY EV forums (endless-sphere and others) and how, in terms of both cost-per-mile and downtime-per-mile, this bicycle has been the single most expensive vehicle I have ever owned.


Well, I'm pretty much done with AmpedBikes, but I'm far from done with e-biking. There are a number of different companies serving the market, and it looks like I simply made the mistake of picking one with good marketing and a slick-looking but poorly-designed product.

I shall return.

Jeff_Ciesielski

iTrader: (10)

LOLWUT?

Joe Perez

iTrader: (8)

Yeah, I know. This is actually pretty typical of Chinese-made power electronics.

The lower wire connection looks ok, the upper one is visually kind of iffy, though there's no sign of discoloration that I'd associate with a poor connection actually causing i2r loss and localized heating. (I'm assuming that the black spooge on it is ejecta deposited there from the failure of the FET, and that prior to the explosion the junction was clean and shiny.)

The resistor across the FET, which if you look closely is between Drain and Source, is most likely a precharge resistor. Probably an afterthought that they added to provide a soft-start function, as this will take care of charging the big capacitors in the controller when the battery is first connected, thus eliminating a huge current surge when the power is initially switched on. Shoulda been integrated into the artwork, but it's a fundamentally sound design.

I'm not defending the workmanship, just sympathizing as a fellow "person who designs electronics."


Also, the black stuff on the red daughterboard is a conformal coating, not damage from the explosion. The physical orientation is a coincidence.


EDIT: Bryce, just wanted to let you know that I got a good chuckle out of your response. I can just picture Adam or Jamie leaning over the battery and saying that.

y8s

iTrader: (8)

You forgot to hook up the doll.







(if only I got here sooner)

Joe Perez

iTrader: (8)

I'm not sure I understand the doll joke, but...

Progress is being made. It took a few days' worth of emails and phone calls, but Amped Bikes has agreed to refund me the full price of the battery. I'd originally asked them for a refund of the entire kit (despite the fact that it's long out of warranty) and I was actually kind of hoping that they'd refuse outright. I was assembling all sorts of documentation in support of a really professional and well-polished small claims case, and was really looking forward to a day in court- it's been nearly a year since my last summons, and I didn't even make it to the box that time. But to be honest, their offer is fair and reasonable. The battery represents about 2/3 of the cost of the system, and I can probably sell off the rest of the parts.

(Sidebar: when you walk into FedEx to drop off a box, make sure it doesn't have Hazmat placards on the outside left over from when the original vendor of the thing you are shipping back shipped it to you in the first place. That freaks them out, even if it's a ground shipment. Regardless of whether the box contains hazmat, you are not a certified hazmat shipper. Cover that ѕhit up and just disclose the contents on the regular shipment form. Maybe put a simple printed sticker on the outside saying "Lithium Battery, 36v 10Ah." But no placards.)


Anyway, I was at a local bike shop a day before the meltdown and happened to spend a few minutes browsing their selection of "nice" MTBs (aluminum frames, disc brakes, decent damped forks, etc) so I've decided to take the plunge and pick up a whole new chassis to start fresh with. Very exciting.

Also narrowing down the selection for the next drivetrain. There's one supplier in China (emissions-free.com) and one supplier in SF (ebikessf.com) that I'm particularly interested in. Both have some rather nice looking motors from MAC and BMS, which have internal gear reduction and freewheel clutches. (Better hill climbing, and less resistance when pedaling with motor off.) I'm leaning towards a 52v, 30A system, which along with the gear-reduced motor should make for a real hill-climbing champion. The supplier in SF is understandably a tad more expensive, and I like the "Triangle" battery that EF.com offers a bit more, but I'm still drawn to dealing with local(ish) vendors.




Anyway, as James May would say: "Good News!" I think I finally figured out why these batteries from Amped Bikes keep failing. It should be mentioned at this point that I am apparently not the only person who has been having this problem, but so far I appear to be the only one who has A: refused to downgrade the system by 30%, and B: actually analyzed the problem intelligently.

Everything you need to know to figure it out is right there in the picture. It came to me this morning as I was enjoying a particularly gratuitous bowel movement, and while I'm going to wait until the refund check clears before I tell the guys at Amped, they're going to feel like complete idiots, assuming they understand it, which is far from a given. Depending on how many hours it takes between the time I get delivery confirmation on the returned battery and the time they email me saying that my refund check is in the mail will determine how snide I am when I post this information on Endless-Sphere. I may even extort them for the information if I can come up with a good angle. (Suggestions welcome.)


Look at the picture of the blown FET. What is missing from that picture?

A fuсking heat sink!

(I can't believe I didn't notice this before.)

Even when they are operated in a steady-state configuration, all FETs still have some forward drop, and thus, generate some heat. It's a lot less than if they're being used to do high-frequency switching like in the motor controller, but it's still there.

The specific FET which failed in that BMS (IRFB3307) has a perfectly reasonably voltage and current rating (75v at 120A) however it also has a relatively high on resistance (RDS) for devices of its type, at 4.6 to 5.8 milliohms. Assuming the worst case on-resistance of 5.8mΩ, that means that at 22 amps, there's going to be 2.8 watts of waste energy. That may not sound like much, but in a totally enclosed and isolated space, that power is going to cause significant localized heating. Pump 2.8 watts of thermal energy into a device with a physical mass of just a few grams continuously for a full minute, and you're going to seriously stress the part. Repeat that process enough times, and eventually something's gonna fail.

And remember, those resistance figures are only nominal values at 25°C. As temperature increases, RDS increases, which causes power dissipation to increase, which causes temperature to increase more, which causes RDS to increase more, and so on. FETs, you see, are the ******** of the semiconductor world. They're like the girl at the party with the Chinese word for "soup" tattooed on the back of her neck. The drunker she gets, the more she drinks, right up to the point where she does the technicolor yawn all over your nice sofa and then passes out face down in the mess. Once a FET starts to overheat, it tends to rapidly accelerate towards destruction. And there's no dog to lick up the aftermath.

Now, I realize that the FETs are heat-sunk to the PCB, but what is the PCB heat-sunk to? Absolutely nothing. It's simply floating inside an enclosed space atop a plastic bracket, with nearly zero ability to shed heat to the outside environment. Adding mass like that will buffer the FET slightly and lengthen the amount of time it can operate before failure, but as die temp increases, its ability to dissipate power decreases, and eventually the whole assembly is destined to heat up to the point where something goes wrong.

Thermodynamics is a bitch.

I'll bet you €1 that if they heat-sink the FETs to the aluminum enclosure of the battery (or create a finned aluminum upper cap which they mount to, or similar) that they'll stop failing. With a proper heatsink, you could pull 30A through that sucker all day long with no problems (except for that damned 5x20 fuse in the bottom which they really need to replace with either a circuit-breaker or an AGC-series fuse.)

They could also improve the thermal performance by using a different FET with a lower on resistance. The IRFB3077, for example, has about the same current and voltage ratings, however the RDS(on) is much lower, 2.8 to 3.3mΩ, meaning that at 22A it'll only dissipate 1.6 watts in the worst-case scenario- 40% less. It'd still require heatsinking, but it would run slightly more efficiently. The IRFSL3107 is even better (2.5 to 3.0mΩ), and while it's a TO-262 part (and thus lacks the mounting tab of a TO-220), if you're engineering a new mount anyway you could accommodate it with a clamp-style heatsink attachment.


Anyway, very exciting stuff. The Armstrongs aren't even going to be able to dream of catching up to the guy wearing jean shorts and a Hawaiian shirt, riding a heavy MTB with a milk crate on the back.


Sidebar II: Jason, I fully acknowledge that you are more experienced than I in power supplies and MOSFET technology. Nonetheless, if you come in here and start crapping up my thread with a bunch of unrelated exhortations, I swear to the Hindu Maha Kalki, tenth avatar of Vishnu, that I will come into your house as you sleep and murder you with an axe.

curly

iTrader: (25)

What bike are you going with? That was the only part of the above post I understood. But read and enjoyed it all.

My girlfriend and I both have trek's entry level mtb, wheel brakes and front suspension. Very nice, although after a few rides with another couple, said male counter part bought his wife a step up from the ones we have. Better suspension, disc brakes, and bitchin' paint.

Im sure the gear/shifter equipment is better, but i couldnt tell. The suspension is better, probably the difference in calving quality of kyb agx vs. koni shocks. You can definitely tell, but they're both good.

The brakes are meh. VERY strong, and difficult to modulate. When I first bought my bike I discussed at length with the bike shop guys about the benefits of disc brakes. They don't fade less, stop harder, squeal less, last longer, and certainly aren't cheaper, with more expensive pads, and with traditional brakes, the "disc" doesn't need to be replaced.

I'd relate the feel of the disc brakes to a 6 puck clutch. Drivable yes, but much harder to modulate than an oem clutch.

I've only experienced two different disc brakes, both trek, both cable operated.

skidude

iTrader: (3)

The only argument I have ever understood for disk brakes is that it takes more of a mud puddle to get them wet, so they work better when riding in the mud. That and they look cool I guess.

fooger03

iTrader: (2)

My experience with disc brakes is that they are amazing...

When cycling through actual woods, up and down hills, through thick mud, narrow trails, try not to hit your handlebar on trees sort of stuff, you simply don't run out of stopping power. With old style wheel-clamp brakes, it becomes a real b!tch to slow/stop if you get a little bit of mud on the wheel. That crap gets between the wheel and the brake and friction goes to ****. On a disk brake, if anything does get between the pad and rotor, (less likely from the start simply because of the increased clamping force,) the grime is quickly forced into the slotted brake rotor which cleans the pads almost instantly, and you are granted full stopping power. It is especially useful on super steep downhill sections, (the kind where you're so far back on the bike, your face is in danm near in the seat,) because the brakes are much more "sure footed", and it's so much easier to modulate them when you're not needing to torque your brake levers to near the point of failure.

They do take a little bit of getting used to, though not a lot. Another interesting tidbit is that I've DEFINITELY thrown myself over the handlebars and done a faceplant on a flat sidewalk before. Going from a wheel brake to a disc brake is a lot like going from a super heavy truck using shoes/rotors on all 4 corners in the pre-power brakes era to a modern ZR1. Are there going to be some challenges? Absolutely. Is change good? I'd like to think so!

Braineack

iTrader: (62)

no pics no care.

golftdibrad

When I saw that pic I just kinda assumed it was heat sunk to the metal thing in the background and they were exceeding the design of the component from crappy board/circuit design. But that will do it every time.

rleete

iTrader: (21)

I thought it was odd, because any other time I've seen those components, they are mounted to something, even is it's just a tiny piece of aluminum plate. But, being that electronics are pretty much greek to me, I figured it was best to keep my mouth shut than to risk coming off like a complete fool.

So, could you replace the MOSFET, mounting all three of them to some kind of sink, and prove your point? Why re-engineer the entire system, if you think that's the problem? Or, to put it another way, why spend more money if the fix is relatively cheap?

Joe Perez

iTrader: (8)

Well, the one store nearest me where I've gotten the best vibe overall carries the Giant Revel series, which are reasonably priced in the $500-$700 neighborhood.







I have no idea why I didn't notice it before. Maybe because it's such an obvious thing that I just assumed it had to be there. If you're looking at an engine, and it's suffering from chronic overheating issues, the first question you ask is probably not "did the manufacturer of this engine bother to install a water pump?" It's such a basic thing that the idea that they might have left out the water pump just doesn't even cross your mind, even if you're staring right at the place where the water pump ought to be.



Yeah, I probably could.

However, I've been wanting to replace this system for a while anyway. The performance on hills is pretty lackluster, and I'm itching to try something new. By doing it this way, I save $500, since Amped is taking the old battery off my hands.

y8s

iTrader: (8)

this is not the correct scene but there is, in fact, a doll.



the line comes later when they try to enhace her bosom for baby iron man. below on left.

curly

iTrader: (25)

I think this is what we have? Two years can change a lot. Looks like their website is newly redesigned.



Just don't forget to compare the weights, ------- bike companies can take the ------- time to weigh their frames. I'm so ------- over it. Yes a 21" down tube weighs more than a 20" down tube. TELL US WHAT THE ------- DIFFERENCE IS YOU PIECES OF ---- ----.