Intense Brushless DC Stage Brings More E-Bike Power

[ZombieSS] wrote in to share the latest iteration of his new open hardware Brushless DC motor high power output stage posted on the Endless-sphere forums. The thread is a gold mine of useful information on designing, building and debugging high power electronics and the whole thing is worth reading. This includes the story of issues he faced with common mode noise picked up by the probe leads on his Rigol, which sidetracked the project for a while.

ebikeWe’ve covered various ebike hacks before, but the guys at endless sphere appear to be developing a number of solid open hardware designs in this area. This includes the Lebowski controller which [ZombieSS] used in conjunction with his design.

He has installed the controller and output stage on his electric bike, and you can see one of the first test runs in the video below. We look forward to hearing more from the awesome hackers at Endless sphere!

17 thoughts on “Intense Brushless DC Stage Brings More E-Bike Power

  1. I’d go with IGBTs or larger fets. I mean they do make stuff to fit the bill and it would likely be cheaper and easier in the long run. It’s not like controlling motors in anything new. I cringe when I see them trying to pass crazy amounts of current through circuitboards and globbing on massive amounts of solder to try to carry all that current. Why not get IGBTs or fets that are designed with large tabs that are designed to handle massive current in the first place? Cost? Well kids, you just shot that argument dead when you spent hundreds of dollars on bus bars and solder!

    1. Which IGBT or FET would you take? Tell which specific part you’d take. Bus bars and solder cost almost nothing…not hundrets! Why would you take an IGBT if the voltage is below 150V and the switching frequency is 20kHz+? What about scaleability? Switching losses?

      1. Pick up a mouser or digikey catalog and look it up yourself. I do this stuff ALL DAY for a living… And what parts I use are need to know, and you don’t need to know…

        1. Oh, and as far as the hundreds comment, TIME = MONEY. Well at least to some of us. And spending time trying to parallel fets and globbing on solder seems like a waste of time when you can buy parts that would work just as well without all the fuss . And “What about scaleability”? If you design a fet driver board, it should probably drive just about any fet. So all you have to do is choose the fet you need and how to cool it. And paralleling fets is probably hurting your switching losses because they are not going to all switch at the exact same time anyway…

          1. switching time should be so short that it doesn’t matter. IGBTs would be terrible choice at such low voltage,
            the FETs used are ~5mOhm, even with several hundred amps per FET you won’t get near the ~2V Vce of an IGBT

          2. Haha look up the parts myself? I You won’t be able to provide any part… you do this all day long? Then show evidence, I want to see a single FET that is able to handle the current! A Fet driver for any FET…and IGBT probably as well, huh? …like fonz and syntroniks said… IGBTs are a terible choice here. If you can do it better, why don’t you give constructive advice?

  2. eyes bleeding, you should check out zombiess’ post on endless sphere, its a very very good layout for a 3-phase inverter of eBike power level (i forget now, is this thing 5kW rated continuous?). all his work is open source so you can get to see any detail you want including why things were done the way they were. with your experience, your input would be valuable.

    some other comments, just to keep the fire blaze’;n ..
    IGBT vs. Mosfet at low voltage? i think for most DIY eBike application probably the Mosfet with a few in parallel is the way to go (and yes there are ways to get the current to share mostly equally so you can effectively parallel). but there are also times when the IGBT should be used, even at low voltage. there are things the IGBt offers that the mosfet does not, most notably failure mode and high ruggedness. if you are desiging to a customer spec or a certain high rel application its possible you will have no choice but to use IGBT. also keep in mind that not all mosfets or IGBTs are good for H-bridge topology.. some perform better than others (switching noise & short-circuit withstand which are really important features of the switch).

    for losses, its not Rds_on that matters most… the losses have to be evaluated in the application which is dc/ac at 20khz switching. the design should be such that the switching loses are about the same as conduction losses (or 60/40 split is more realistic for mosfet). you need to know how to read the datasheet, how to match this datasheet to your application and how to do the iterative math to make an educated guess of the losses. zombiess has done all this, i’m sure he’d share his math with you if u ask.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.