making PWM speed controller for Electric bike

Søren

Joined Sep 2, 2006
472
Hi,

You could use a 555 but never with 1 MOSFET. Will need some work to figure out what could handle 1200W and 27A + heavy gauge soldering.
The MOSFET won't have to handle 1200W. That's the point of using PWM, the transistor is either completely ON, dissipating little power (R*I^2 eg. 0.002*27^2=1.5W), or completely OFF most of the time.
The transistions needs to be speedy however, so it has to be driven with eg. 3A..5A to overcome the input capacitance to get a low average dissipation.

You can get MOSFETs handling more than 100A.
The driver is the critical point here.
 

Amberwolf

Joined May 2, 2008
28
-For back EMF, you can use diode across motor and MOSFET has the internal reverse diode for protection.
-Overheating of MOSFETs is probably gonna be a drive problem. And there is the usual heating which has to be taken care of by heatsinks and cooling fans.
-Overvoltage/Undervoltage/MOSFET failure - these can all be protected against with appropriate circuit.
All true; my point was only that these all add cost and complexity to the controller and it's design process, which are part of the reason that good controllers cost so much. ;)

Each individual thing is simple enough to protect against. When in combination, sometimes the protection for one thing can degrade the performance of one or more of the others, requiring compromises between various things or more complex design to work around that.

One thing that is done more and more these days to work around these issues is to control everything in software, and have most of the protection circuitry actually just be feedback from the various circuits into the microcontroller, which then decides which things to prioritize moment-by-moment, to protect the controller, the motor, the batteries, and the operator. The order of priority depends on who designed the controller. :)

I didn't even go into the topic of regenerative braking (regen), but controllers that can do that add more complexity. The ones that choose to not do regen once batteries are at full charge (to prevent overcharge of the batteries and potential damage, depending on chemistry) add even more complexity than that, though most of that would be in software (since they're probably microcontroller-based).


The part that bothers me most is that I barely know *anything* about motor controllers, yet I already feel like I've had a couple of years of education on them after my research and my attempts at making my own. Even with all that, my two most recent attempts this summer other than the 2QD were unsuccesful even on the test bench.

The first, based on the uc2638 by Unitrode/TI, did not work at all. I was too frustrated to even troubleshoot it beyond the basics, and went to my backup plan, based on the MC33033, which can be used as BLDC or PMDC brushed controller, with the MOSFETs driven by a pair of NCP51xx chips (forgot which ones).

That one actually worked for a moment, using a radiator fan motor (with no load) I'd used for one of my early ebike motorization schemes. Then the bottom (drive) MOSFET of the half-bridge died shorted, and apparently took out the drive chip *and* the MC33033, because just about every pin on all chips was at Vcc. There could not have been any overheating, because there was not enough current involved, and no heatsink was even a measurable amount above room temperature. So it was probably a voltage feedback, but I don't get how it could happen.

Except for PCB layout, I can't figure out how it could have died in that fashion, because the Fairchild MOSFETs I'd used had some built-in spike/etc. protection, I'd used very tiny 1/8W 10ohm gate resistors that *should* have burned open upon a current draw that would damage the driver chip, and the driver chip itself is designed to protect the stuff behind it in the chain. Yet it did not. The freewheel MOSFET on the top of the half-bridge was not affected, and I used it successfully in my later bench tests of the 2QD controller.

However, all of both of these controllers were hand-wired on copper-clad solderable breadboard, rather than a properly-designed PCB (another thing I didn't cover), which would probably have prevented whatever radiated energy and induced currents that ended up destroying the MC33033-based one.

So even using advanced technology designed to protect itself and other components around it in the controller does not ensure it will actually do so. One must also design the circuit board they're installed on with the proper layout, routing of traces, and widths (and weights!) of copper to allow proper current flow and heat dissipation. Placement of components can be critical to keep components sensitive to EMI away from the traces that carry large currents, and to keep all the large filter capacitors as close as possible to the devices they're filtering. And to keep their leads as short as possible, and traces as thick as possible.

There are an awful lot of considerations for making a controller that are not obvious until you've blown a few things up. :) Unless you first study all the things other people have done and failed at and then been successful at, of those that have documented them publicly in one way or another.


There's a controller for EV cars/etc called the Soliton1 by EVNetics that was recently demonstrated in public successfully, by Qer and Tesseract over on the DIYElectricCar forums (in the New Controller Prototype thread). It's development has been ongoing for quite a while now (by already-experienced people), and it is still in it's final beta stages. It's quite the kickass controller, micro-controlled with updatable software and even a web-based interface for getting data back from it, changing parameters, etc. Yet they've still had troubles getting everything working as desired, some in hardware and some in software.

It can all come together in the end, but it can certainly take a while to do even with the necessary knowledge already in your head. If you don't know electronics well yet, and then have to not only learn electronics as you go but also all the design considerations of PWM motor control, it's a long and tough road, depending on what you expect out of the controller when it's done. ;)

I certainly don't want to discourage *anyone* from ever trying to do this, but I do want them to know what they're in for before they head down that road, because I know already how frustrating it can be!
 

Amberwolf

Joined May 2, 2008
28
Søren;164031 said:
The MOSFET won't have to handle 1200W. That's the point of using PWM, the transistor is either completely ON, dissipating little power (R*I^2 eg. 0.002*27^2=1.5W), or completely OFF most of the time.
The transistions needs to be speedy however, so it has to be driven with eg. 3A..5A to overcome the input capacitance to get a low average dissipation.
Exactly. One of the first things I learned about MOSFETs, the hard way. ;)

You can get MOSFETs handling more than 100A.
The driver is the critical point here.
Which is why I used those JFETs to drive my NTY100N10's. :) Those latter MOSFETs are capable of 100V at 123A. There are even heftier ones out there now.

However, there are advantages to using multiple smaller MOSFETs.

--the RDSon of each one will be paralled with all the others to make the total RDSon much lower, reducing the total power dissipation on any one of them and the whole pack.

--the failure of a single one for whatever reason won't kill the whole controller (assuming you have separate gate drive for each one, or at least a fusible link or fusible resistor at the gate for each one, in case of gate-short).

--smaller MOSFETs are significantly cheaper to buy than larger ones, so that enough smaller ones to cost the same as one large one will handle even *more* power in total than that single large one, usually.

--if you're recycling parts (like I am), smaller MOSFETs are much easier to find in junk equipment, such as UPSes, power supplies, audio amplifiers, and so on.

--driving the gates of smaller MOSFETs can be easier, even if you have a lot of them, because you don't need as large a capacity of a gate driver (although you might need multiple smaller gate drivers if you use a LOT of MOSFETs).

Of course, using larger MOSFETs instead means the circuit is simpler to build and probably physically smaller. :)
 

Thread Starter

CaptainPrice

Joined Aug 12, 2009
49
You've given me a lot to think about (& read)

Do you have any schematics to a working controller?
I'm thinking I might just use a 500w 24v motor instead.


anyways thanks, and rec'd -great stuff!!
 

Amberwolf

Joined May 2, 2008
28
The only schematics I have to hand are of the Scoot'N'Go prior to my modifications. They're hand drawn by someone else, located by a post on another forum somewhere quite a while back. I didn't check their accuracy, either, as I was only looking for the basic idea of what it was, and after that I just started poking my meter around it till I figured it out. :)

I never did draw up my own schematic for it's original or updated versions, since by the time I would have gotten around to that it had died on me for the last time (it is literally falling apart in places from soldering/desoldering plus the original corrosion and burned trace/component/PCB damage).

I'll post it if you want it, though.

The other schematics I've used are all found in the appnotes for the chips I mentioned, at the TI and OnSemi sites, with various attempted modifications based on the necessary adaptations to my motors and batteries and available recycled parts and whatnot (which I just pencilled in on various scattered printouts of the appnotes). I am pretty disorganized. :(

Oh, and the schematics on the 4QD tech site. I can't post those here because they are 4QD's property, so you'd need to go there to get them. BTW, I messed up the link earlier, and it is http://4qdtec.com, not http://4qdtec.co.uk. I always forget that. :oops:

I checked just now, and there is a very very simple controller that would work probably as well as a basic 555 controller, as well as the earliest 2QD version schematic there on the PWM part one discussion. The improved versions including the issue 8 that is nearly identical to the final version 9 (close enough to build mine from) are all in the members-only data area.
If you want to join to get all the other info on there (which I think is worth if you're pursuing designing and building your own controller from scratch), go here: http://www.4qdtec.com/Contact/Join.html

I haven't seen any new info there other than the corrections and updates (and changes to things in response to questions I had for Richard in discussions with him during corrections exchanges), but I haven't looked at most of it in a while, so there might be new stuff still being added. (EDIT: I guess my membership has lapsed since I last dropped by, too...trying to decide if I want to re-up to get access to anything I didn't get back when I did have it, now that I have a working controller).

There are also a lot of other things there, from LED battery level meters, to audio stuff, to car bulb flashers with failure indications, various electronics theory sections, industrial controls, etc. Not just the circuits, but explanations of why they are designed the way they are, which is much more valuable to me. :) I probably learned more from that site than I did from all the teachers combined during electronics school way back when. (to be fair, I was studying to be a repair tech, not an engineer, but still....)

Richard Torrens is a clever designer, and so far one of the better resources for controllers and information about them (via that site) that I've found in any one place in my ~2-year quest. I'm sure there are lots of places I have yet to even find, some of which are probably better, but of the ones I know, I'd recommend his first.


**************************
EDIT: the motor size isn't really important, as any controller schematic you do find could be modified to accommodate a larger power output section, as long as the voltage involved can be tolerated by the rest of the components or protected against/regulated as needed.

But keep in mind that at least some places have legal limits on how big a motor you can have on your ebike, if it is to remain a bicycle-class vehicle. Some, like Arizona, don't specify a limit for electric motors but do for gasoline (it says 49cc and below is still bicycle-class, but above that is motorized cycle or moped or motorcycle depending on how far above, and those all require license/registration/insurance, where a bicycle does not). Most of the ones I've seen that do have a limit give it as 1HP or 750W. Some localities (county/parish/town/city/etc) have their own limits that may be stricter than the state's limit. Most also limit the speed at which you can go, too. Some limit it only when the motor is being used, some limit it just because the motor exists on the bike, and some don't say which way they mean it so it's up to an officer at the time of any traffic stop, which has made life difficult for some cyclists. That limit is often 15 to 20 MPH. There are also I think some states which ban the use of electric bikes, and/or mangled laws by "accident" thus making them illegal for road use (I think this happened in Hawaii when they changed the laws to cover Segways some years ago, I don't know if it was ever fixed).
**************************
 
Last edited:

Thread Starter

CaptainPrice

Joined Aug 12, 2009
49
I found this controller at monsterscooterparts which seems to be a ok deal there a bunch under the razor scooter controller section that seem to be rated for pretty high wattage. The one on the link looks to be 24v 700 watt controller.

I live in LA and I looked up the rules and it seemed to be like their considered bikes with a 20mph motorized speed limit and 750 watt motor limit mandated by federal law. Although If I went with the 1200 watt motor I don't know how they would ever tell without destroying my bike and sending it to some company to test it.

Thanks for all the stuff on making a controller I'm not sure if I'll make one this time but either way I find learning about them interesting so by the next time I need one I'll be ready. I kind of need to get it up and running pretty soon.


Next I got to figure all the gearing.

Where do you buy all your parts from? (batteries, motors, spockets, throttles)
 

Thread Starter

CaptainPrice

Joined Aug 12, 2009
49
Another question

If I was going to use a 600 watt motor rated at 24v and 36 amps could I use a controller rated for 24v and 30 amps?

If it is a 600 watt motor how can [24*36= 864] 24v times 36amps??

What am I missing? confused about????
 

Thread Starter

CaptainPrice

Joined Aug 12, 2009
49
Even if the battery was at 20v it would still only have to draw 30A to get to 600watts....????

note the controller cuts out at 20.5 volts
do any of the cheaper controllers cut out safely when over drawing amps?
 

Amberwolf

Joined May 2, 2008
28
I found this controller at monsterscooterparts which seems to be a ok deal there a bunch under the razor scooter controller section that seem to be rated for pretty high wattage. The one on the link looks to be 24v 700 watt controller.
Keep in mind that all controllers I've ever seen except for the Soliton1 (being developed by Qer and Tesseract over on DIYElectricCar forums) are rated at their *peak* power output, which most can only do for a few seconds, if even that. Some can do it for minutes, *if* you keep them cool, which their own heatsinks are usually inadequate for.


I live in LA and I looked up the rules and it seemed to be like their considered bikes with a 20mph motorized speed limit and 750 watt motor limit mandated by federal law. Although If I went with the 1200 watt motor I don't know how they would ever tell without destroying my bike and sending it to some company to test it.
Well, they wouldn't know just by looking at it unless its' labelled either with wattage or with data that allows calculation of that wattage. But it could be tested by putting the system on a roller dyno and find out pretty easy, without taking it off the bike, at least in theory.


Thanks for all the stuff on making a controller I'm not sure if I'll make one this time but either way I find learning about them interesting so by the next time I need one I'll be ready. I kind of need to get it up and running pretty soon.
A very simple one can be made very quickly. But it will have many potential failure modes that lurk, waiting for Murphy to whistle for them to jump out at you at the worst possible moment. ;) It might never even fail. That's unlikely, but possible.


Next I got to figure all the gearing.
Are you going to do as I have, and run the motor thru the bike's pedal drivetrain so you can shift gears with it? Or are you going to directly drive the wheel (in one of several common ways)?

If you run it thru the pedal drivetrain, then if you put the input sprocket for the motor on the pedal cranks, then make sure the motor's RPM at that sprocket is around 90-100RPM (the same rate at which one typically pedals). If you pedal slower or faster than that, match that speed.

If you put the input sprocket from the motor on the rear wheel itself, then you have more freedom to come up with an output stage for the motor, basd on what final wheel RPM you wnat to have at max speed.

Where do you buy all your parts from? (batteries, motors, spockets, throttles)
I don't, unless I have no other choice. I get them off other things, including old bikes, that people throw away or give away. Many things have come from the local Phoenix AZ area lists for http://freecycle.org. There are probably several lists in the LA area. Just sign up for them, and watch for OFFER items of the right kind that are close enough to you to be worth going after.

Sometimes even a WANT listing is helpful, because if they are asking for a new treadmill or microwave or lawnmower or bike or whatever, there is a very good chance they have a broken one they're about to throw away that you could have for free to use for parts. Just email them and ask. Usually, they'll reply and usually no one else has even thought to ask, nor did they think to offer it! So they'll give it to you. Has worked for me a few times to get various parts.

The problem with this method is that it takes a long time, usually several months or more, to get all the bits you need. I don't have a lot of choice, given the lack of work I seem able to be hired for around here, and that my current job keeps cutting hours all the time. I have to take what I can get in materials, and figure out ways to use it (hence the crazy monstrosity I ride).

I have gone to junkyards and bought some things for a few dollars, like my pancake motors used on the first successful ebike.

I have also been given a number of things, including the old wheelchair motors and batteries I'm using now, by readers of my project blog who found it an interesting enough project to contribute to. I certainly never expected that when I started the blog, but it definitely helps. Sometimes people offer things but never reply when I try to take them up on it, which is frustrating, but hey, it didn't cost me anything either way, so I can't really complain.

Police auctions are a good way to get bikes to rebuild or salvage for parts, since a lot of them are stolen, taken in as evidence, and never claimed by their owners. Most police departments auction off items like that. I've never managed to get any that way, but I've bid on a wide range of bike types trying to get some.

Thrift stores and yard sales are a great place to find unexpectedly perfect items, for almost nothing (sometimes at yard sales they have a pile of things that no one wants that they plan to throw out, figuring they'd never sell, and often if you are willing to take the whole pile they'll give it to you so they don't have to deal with it).
 

Amberwolf

Joined May 2, 2008
28
my reply was too long, so I had to post in two halves:

If I was going to use a 600 watt motor rated at 24v and 36 amps could I use a controller rated for 24v and 30 amps?

If it is a 600 watt motor how can [24*36= 864] 24v times 36amps??

What am I missing? confused about????
Well, the power rating on a motor doesn't necessarily match with any electronics formula applied to it's voltage and current ratings, or it's winding resistance, etc. It simply means it's the maximum power that motor can dissipate safely before burning up. Usually this is a peak power, just like the controllers are rated at, and the actual *continuous* power, which is what's really important in a lightweight EV like an ebike, is almost never stated on the motors. Treadmill motors are an exception, as are some other continuous-duty motors, and are often marked specifically with that continuous-duty rating as well as sometimes a peak or short-term rating.

As an example, the wheelchair/powerchair motors I have right now are listed at 24V/3.6A, which if you multiplied them is around 85Watts. Pitiful. Now, any motor like that would be terrible to move a load, but it turns out that 3.6A is only the *unloaded* current draw. It can easily draw *much* higher currents, more than 5 or 6 times that, probably more, under full load (up to stall current). In reality, they're probably about 350W motors, able to dissipate that much power (including waste heat). You yourself can probably generate 100W of power when riding; 150W if you're athletic. So they do a fair job of helping me along, even with all the extra weight (just not enough help for hills and the like).

Even the 650W 4-pole motor I also have from a powerchair is still listed as 24V/3.6A, but it obviously can handle a lot more power than that, as it operates perfectly normally over long periods when it's hot enough to be extremely uncomfortable to hold.

I would prefer if all motors had rating plates that included their torque ratings, RPM/V, etc, but these are rarely found outside of industrial motors of various types. Sometimes a manufacturer of a motor will provide you this information, but more often they either don't know or don't care and won't tell you even if they did. ;)


Even if the battery was at 20v it would still only have to draw 30A to get to 600watts....????

note the controller cuts out at 20.5 volts
do any of the cheaper controllers cut out safely when over drawing amps?
It doesn't necessarily work that way. Remember, these are almost certainly *peak* ratings, and you would not want to run them at their peak power levels very long, or you risk damage to the motor or controller.

Most controllers sold for the purpose of EVs have a low-voltage-cutoff to prevent damage to the battery pack. It depends on what kind of battery they were designed to be used with as to what the specific voltage is, and most are probably not adjustable. For lead acid, around 10V per battery is as far as I would ever want to take them down. Even that will shorten their expected lifespan. That's probably the chemistry expected by that controller (and many of the cheap ones, since lead-acid are also very cheap batteries).

As to whether they cut out during overcurrent, you would have to check their specs or ask the manufacturer. If they do, then you need to know if that overcurrent is adjustable so you can protect the *motor* with it, or if it is only designed to protect the controller itself against meltdown/MOSFET failure.

The 2QD has some protection against meltdown, in that if the voltage across the motor drive MOSFET (when it is in conduction phase) is greater than a certain amount, it means too much current is flowing thru it, and the PWM is cut back automatically to keep from burning it out.
 

Thread Starter

CaptainPrice

Joined Aug 12, 2009
49
I'm still a little confused.

so the motor thats 600 watts and says 24v 36amps, even if that is peak does that mean the motor would peak at the 600watts or at 36 amps even though its over 600watts?


So would it be safe to use a peak 30 amp controller with the motor? Cause if it did get that high it would already be over the motor peaking point????



If not do you think I could open up the controller and replace the mosfets? or could I simply attach an appropriate sized mosfet on the leads to the motor??
 

Amberwolf

Joined May 2, 2008
28
so the motor thats 600 watts and says 24v 36amps, even if that is peak does that mean the motor would peak at the 600watts or at 36 amps even though its over 600watts?
What it means is that it was (probably) designed and rated for a peak of 600 watts output power at whatever voltage and current combination you choose to use it at.

So if you use it at 24V, then depending on it's internal resistance and the load you put on it, it will draw a certain amount of current. At some point that may be more than the 36A it is rated for, and if kept up for more than a certain time period (seconds to minutes depending on the motor size and design, with smaller motors being more susceptible) it may burn the motor windings or damage the brushes, etc.

Depending on the labelling, it may also mean that when used at that rated voltage with a controller that is limited at the rated current, the motor may draw that much current when such a load is placed on it that it actually stalls (stops spinning). More current than that and the motor then begins to be damaged from overheating. Motors aren't labelled consistently, especially with cheap ones, so all you really have is a rule of thumb. Much bigger motors can be run beyond their rated limits for much longer.

You can run most brushed motors at much higher voltages than they are labelled at; for instance many EVs use forklift motors that were rated at 48V, but they use 144V or higher battery packs. The voltage problem usually arises because of the brushes--arcing always occurs between brush and commutator, but at some point in raising the voltage, especially at high loads and thus high currents, that arcing begins to be very destructive of the brushes and commutator faces.

You can also run most motors above their rated currents, but how much more depends on their design, their size, how much cooling you apply to the motor and how effective that cooling is. In an enclosed motor with no ventilation, adding fins on the outside is about all you can do easily, and it is not very effective at ebike speeds. Adding fans to cool the motor helps some, but unless you can force the air through the inside of the motor, to cool the armature windings, it doesn't do enough to cool the most easily damaged part (since it is the source of the heat generation and is only covered by what amounts to plastic, which will melt or burn above a certain temperature). THere are other more exotic cooling methods but they're not appropriate for every motor, and they can get heavy and expensive and waste power.


I used those 12V pancake radiator fan motors at 36V, and put a heck of a lot more load on them than they were ever designed for, yet they still show no problems with the windings. One of them had a shaft break, but I think that was from repeated pothole encounters plus my mounting methods. They got so hot that if I poured water on the outside of them after a while, it'd just flash into steam. Yet they still worked. I'm sure it shortened their life, and it probably even reduced the effectiveness of their magnets (since those are heat sensitive at some point, too). (For motors that were designed to be in constant high airflow, they did a dang good job at temperatures and loads much much higher than they were ever expected to take in their original application).

On the flip side, I had a 24V 250W motor designed for the ScootNGo, and that thing was so poorly made that even running it *at* 24V was problematic. On the ebike, I set it up as just an assist motor, and even so it actually burned out windings because I was trying to let it do more of the work. It took less than an hour's total use, over days of use like this, to die, but die it did. And that was simply because the load on it was too high, making it draw too much current, and overheating inside.

So...you can exceed a motor's specs, but for how long and what happens when you do depends on the motor's design and quality of workmanship, as well as how much overrating they put on it.



So would it be safe to use a peak 30 amp controller with the motor? Cause if it did get that high it would already be over the motor peaking point????
Since the controller's peak power probably is rated for much shorter than the motor's, most likely it would be fine. Just remember that with cheap controllers, they might not have a current limiter so they'll supply as much as the motor and load on it demands, which could be much higher than 30amps when starting it up from a complete stop on the bike. If so, the controller might not take a lot of that. If it has a built-in current limiter and other protections, it should be fine, but if it doesn't specify that it has these protections and what their limits are, I would assume for safety's sake that it probably doesn't.

The motor can most likely take more momentary abuse than the controller, since the motor has more mass to absorb heat and then dissipate it over time. Most controllers barely have a heatsink, and so don't have much tolerance for exceeding their limits before damage occurs, assuming they have no protections in place.



If not do you think I could open up the controller and replace the mosfets? or could I simply attach an appropriate sized mosfet on the leads to the motor??
You shouldn't put extra MOSFETs out past the original ones, as then if the controller does have any current limiting or other protections, they won't work because it won't be able to sense what's going on at the motor.

You could always replace the original MOSFETs, but replacing the heatsink with a bigger one will also usually be necessary. Sometimes you may find that the original MOSFETs are rated higher than the controller is, simply because the heatsink is too small to deal with the power wasted as heat!

Also, at higher currents, the capactor(s) in the power section of the controller are less effective, unless you add more of them in parallel, too. They're there to limit ripple of the motor voltage. Without them, all the current comes directly from the battery, and heats up the wires from the controller to the battery more, as well as the batteries themselves.

If the batteries are small enough, then depending on their chemistry they may not be able to supply instantaneous currents at full voltage for some motor situations, and that's what the capacitors are there to help with (since they can supply most of their energy instantaneously for low-ESR versions).



In practice, most of these things are not that big a deal, because you probably won't be hitting peak powers very often for very long. Most of the time those high currents are only at startup from a stop or during acceleration. Once you're running, they drop as the motor spins faster. If you have a motor designed for at least 300-400W, and a controller and batteries capable of supplying it power as it needs it, then on a typical plain bicycle with an average weight rider it'd probably help you pedal at decent speeds with a little effort of your own for most situations.

Hills would be another thing, if long or steep, and trying to go really fast might be more demanding, too. Without any pedalling, you'd probably need a bigger motor (600W or more), assuming the motor does not go thru the drivetrain but instead is directly connected to the wheel.

The reason that a human outputting maybe 100-150W of power can be "better" than a motor with 300-600W of power is because of that drivetrain. If you shift to low gears when starting up or hill climbing, your speed drops but it is easier to pedal, so you are still using the same amount of power as when you reach the flats and speed up by shifting to higher gears.

The motor can be a smaller motor and do the same things, if it also runs thru the drivetrain. That's how I can use that small 300-350W wheelchair motor to run my 120+ pound bike plus my 150 pound weight plus 10-20 pounds of stuff in the cargo pods, and ride without any pedalling effort at up to 15MPH. Granted, the motor gets awfully hot that way, and even the 31Ah batteries don't last very long, but it does work, simply by shifting thru gears as needed to let the motor stay in it's higher speed ranges and be more efficient.

If I tried to direct drive the bike with that motor at only one gear ratio, I'd probably be able to walk faster than it could go ;) assuming it was geared to be able to start from a complete stop.
 

Thread Starter

CaptainPrice

Joined Aug 12, 2009
49
ThankS!!

about How many amps would you draw going up a steep hill at full throttle with no pedal assist?... would going half throttle solve the problem?
 

Amberwolf

Joined May 2, 2008
28
about How many amps would you draw going up a steep hill at full throttle with no pedal assist?... would going half throttle solve the problem?
There are calculators (including one on the http://4QD.co.uk page to help you with product selection) that will help you determine the current draw in various situations. There is no way for me to answer that question directly.

Going half throttle would indeed help with heating, as you would only be applying half the voltage of your pack across the motor, so would only be putting half the power thru it.

However, it will also make it take at least twice as long to get up the hill, so the heating would be happening for much longer than otherwise, so it is sort of six-of-one, half-dozen of the other, sort of problem.

I'd guess that as long as it is a short enough hill that you would not normally reach overheat temperatures on it at max throttle, then you would be better off running at max rather than half. But if it is part of a series of hills, you would be far better off having a bigger motor (and/or controller) and/or a better cooling solution (for either/both). If most of your travel is on flattish ground both before and after the hill, there would probably be time enough to dissipate the heat buildup to prevent damage to the motor (or controller). Otherwise, repetitive or continuous heating of them at or beyond their ratings will eventually cause enough wear and tear that something will fail earlier than expected.

How much earlier? Unknowable without knowing the *real* limits of the devices, quality of workmanship and materials, etc. You'd be testing those real limits. :)
 

Thread Starter

CaptainPrice

Joined Aug 12, 2009
49
What size, rating capacitor is ideal for use with that 600watt motor?

also were did you get your 31 amp batteries? freecycle? did you get them free ot recycled like your other stuff?


So Im thinking Ill probably open up a speed controller and put a huge heatsink and fan on it. do you think another cap would also help or not really?

thanks
 

Amberwolf

Joined May 2, 2008
28
What size, rating capacitor is ideal for use with that 600watt motor?
That depends on your controller design, batteries (what "C" rating they have), cabling from batteries to controller (length, gauge, connectors, etc), acceptable ripple voltage, actual load motor places on controller, etc.

Practically, however, put the biggest low-ESR caps you can as physically close to the MOSFETs as you can, with the power runs from battery to MOSFETs to motor as short as physically possible, with the appropriate gauge wire for the current you expect it to draw.



also were did you get your 31 amp batteries? freecycle? did you get them free ot recycled like your other stuff?
Actually, those came from someone that is following my project blog, from a wheelchair or a powerchair. I expect that the person (or someone they know or help) upgraded or repaired their chair(s) and I got the handmedowns, given the used condition of the parts.

I've also seen them on Freecycle once here, but I was not the person picked to get them. Craigslist might be a good place to look for used ones, but remember that large-capacity SLAs like this are HEAVY.

My 31Ah's (U1) are 23 pounds EACH. It takes at least two to power a bike motor system, for 24V. at 46 pounds, and it isn't efficient until at least 3 of them for 36V, at 69 pounds!

Also, see my most recent couple of project blog posts about the UPS batteries I'm working with right now. Sometimes you can find old server UPS units that have had a recent battery change but the UPS blew up, possibly leaving it with good batteries. Many companies with server rooms and backups like this dont' save or repair these, they just buy new ones. It's cheaper, usually. So if you can find one you might be able to get it for nothing.

So Im thinking Ill probably open up a speed controller and put a huge heatsink and fan on it.
Most likely the MOSFETs are already bolted to a heatsink that has an external face on it. If that face is flat, then you can simply polish it a little to make it as smooth and flat as possible (see the heatsink lapping articles for computer overclockers), then do the same to your add on heatsink, and bolt them together, with your fan on that.

If it's already finned, just add your fan to the existing heatsink; it's a lot easier than trying to replace the existing heatsink with a new one, unless you already know for a fact that it's totally inadequate. It's pretty likely that the power dissipation capability of the heatsink plus a fan will be enough to carry away as much or more power than the MOSFETs can actually take, anyway. Just remember the fan will eat away at your battery too, so you should use one that is thermally controlled (with the little green sensor that sticks out into the airstream, if using one from a computer case for 12V), and move it's thermal sensor to the inside of the controller between the MOSFETs.

do you think another cap would also help or not really?
Depends on the ripple voltage you get in the system at load. You'd have to check that under load with the completed bike, using an oscilloscope at various points in the controller and at the motor.

If there's only a volt or two ripple, then it's no worse than mine, which seems to work fine with it. I'm sure it's more efficient if there's less ripple, but I don't have a bigger cap that will fit in the space available without taking the cap off-board, and that is likely to cause as many problems as it fixes.

If you see big dips in voltage at every PWM cycle while the wheel is loaded down (say, by holding the brake lever down but not enough to actually stop the wheel, just to place a significant load on it), then you should first try to fix the power supply issue causing it, by shortening the battery-to-controller wires, making them larger gauge, or fixing whatever bad connector might be in there (lots of times crimped connections are badly made, especially on cheap controllers, and soldering them can help--but only if you're good at soldering, otherwise cold-solder joints can make it worse).
 

Thread Starter

CaptainPrice

Joined Aug 12, 2009
49
Do you think I should replace the current mosfets or no? And if so Would I need to use the same number of mosfets that were originally inside?
Can I just add more to the original setup?
 

Amberwolf

Joined May 2, 2008
28
Do you think I should replace the current mosfets or no?
I can't answer that. You'd need to see what specs the current system you are putting together will need, then see what the controller itself is designed and spec'd for. If it doesn't meet one or more specs, then you could upgrade the parts that are needed to correct that.

Modifying a controller is not quite as hard as designing one, but you still have to know how they work to be sure you do it right so that you don't have problems down the line.

Remember also that if the controller is designed with a current limiter in it, then upgrading the MOSFETs won't change anything about that current limit, with most designs, so you'd have to do that separately or you still won't get any more power out of the controller than before.

And if so Would I need to use the same number of mosfets that were originally inside?
That depends on how you intend to upgrade it. If you're using the same physical size MOSFETs, and intend to just replace the ones that are there with higher capacity ones that fit your system's needs, then you should probably use the same number.



Can I just add more to the original setup?
If you use identical p/n's, then you can just parallel a bunch of them together, after you ensure the gate drive circuits that run them can handle that much current, or that you upgrade that gate drive circuit, too.
 

Søren

Joined Sep 2, 2006
472
Hi,

However, there are advantages to using multiple smaller MOSFETs.

--the RDSon of each one will be paralled with all the others to make the total RDSon much lower, reducing the total power dissipation on any one of them and the whole pack.
The R_ds_on is usually much higher on devices with a lower power handling. Each time I have checked, it just didn't make sense to go with several "lesser" instead of a single one that could handle it all.
Also, the parallelled input capacitance usually gets larger with the current handling equivalent amount of "lesser" devices.


--the failure of a single one for whatever reason won't kill the whole controller (assuming you have separate gate drive for each one, or at least a fusible link or fusible resistor at the gate for each one, in case of gate-short).
The usual failure mode for MOSFETs are D-S shorted, so you'd have to fuse each drain (or source) to keep it running. But... If one blows, the rest have to share the current and if one device died by over-current allready, how long do you think the rest will last? *pop1*, *pop2*, *pop rest* :(


--smaller MOSFETs are significantly cheaper to buy than larger ones, so that enough smaller ones to cost the same as one large one will handle even *more* power in total than that single large one, usually.
But they will have production tolerances, so won't share current completely equally, which might set of the *pop all* in severe cases - there's for and against, the important part is to know all the possibilities and then make an educated desicion.


--if you're recycling parts (like I am), smaller MOSFETs are much easier to find in junk equipment, such as UPSes, power supplies, audio amplifiers, and so on.
I totally agree, and for non-critical stuff, it's fine, but I'd never second-hand the parts in the final design of a high power controller.


--driving the gates of smaller MOSFETs can be easier, even if you have a lot of them, because you don't need as large a capacity of a gate driver (although you might need multiple smaller gate drivers if you use a LOT of MOSFETs).
I have to disagree here, as mentioned before.

But that said, If it works for you, I'm all fine with it :)
 
Top