Overvolting - as easy as it sounds?

[Jeremy]

Sod the non technical members, post to your hearts content Jeremy...if its too technical those that don't understand or care can skip the post...I've read a bit about it on ES but to save sifting through reams of posts please could you explain here, thanks!

OK, here goes.

Motor speed is controlled by varying the motor voltage using something called pulse width modulation (PWM). PWM is just switching the battery voltage on and off very quickly (around 15,000 times per second for a typical ebike controller). The voltage at the motor is proportional to the ratio of the PWM on time to the PWM off time. So, if the ratio was 50% (on for half the time, off for half the time) then the voltage at the motor, and the motor speed, will be 50%.

There are two potential problems with PWM:

1 - At 0% (zero throttle) the controller isn't switching on or off at all, it is just sat switched off, in effect, so the PWM circuitry has to detect zero throttle and stop switching altogether. This is easy enough for the controller to do, as it can just internally cut the power, in effect.

2 - At 100% (full throttle) the PWM also has to stop, but in this case it has to switch fully on. This is harder to do, because you want a nice smooth transition from nearly full throttle (say, 99%) where the controller is switching power on 99% of the time and off 1% of the time, to the point where it's 100% on. Equally, you want a nice smooth transition back from this as the throttle is closed. This switch from 99% to 100% is termed the switch from PWM to block commutation. Block commutation essentially means that the three phase drive to the motor is now in "blocks" of solid applied voltage, rather than PWM sliced.

One slight snag with this transition to block commutation is that switching the PWM off (and allowing full voltage) has the effect of advancing the timing of the three phase waveform that drives the motor. In essence, PWM adds a tiny delay to the three phase motor voltage when it's operating and block commutation makes the motor run slightly faster (incrementally) than the small change in applied voltage would suggest.

When you select greater than 100% in the programming software for a Xiechang controller, then what happens is that the throttle position where this switch from PWM to block commutation occurs moves down a bit. This can mean that the controller can switch to block commutation at a slightly lower throttle setting and the timing of the three phase waveform applied to the motor over the remaining part of the throttle to full can be advanced. If the timing is over-advanced then the motor will draw more current for very little extra speed. This is very motor dependent, some big direct drive hub motors with a high inductance respond fairly well to this timing advance, others respond rather badly.

Because this setting is rather crude in the way it operates, it can produce variable results. On smaller, lower inductance, motors settings above 100% rarely give good results. If you want to play around with this then the best way to do it is to hook up a meter and measure the supply current to the controller at various programmed maximum speed settings. Measuring the no-load current (and listening to the noise the motor makes at the transition from 99% PWM to block commutation) will give a good idea as to how effective any setting is. If you go too far you will see the no-load current spike up over the last tiny bit of throttle movement, which is a good indication that you need to reduce the % maximum a bit.

 

[NRG]

Excellent, thank you Jeremy.

 

[KirstinS]

The more I read the less I know it seems. One day I'll get past the hump on the ebike learning curve !

However, like nrg I am ever grateful to those , such as jeremy, who can explain complex stuff in simple terms. It's quite a skill

 

[Deleted member 4366]

Think mine is the short and fat version, looks more complicated than I thought in there, I guess reconfiguring isn't an option. Its either a new battery or higher RPM hub motor???? I'll compare prices

thanks

I've got a code 10 front 350w BPM that's done test miles - probably about 20 miles total. £100 plus postage if you want it.
It's sensorless and has a three pin connector near the motor, and the remaining wire is only long enough to reach the middle of the frame triangle, so might need extending.Don't ask me why it''s like that. It's what BMSBattery sent me.

 

[Deleted member 4366]

The three speed switch on the KU series controllers is a mystery to me. I''m sure that there must be a timing element to it somehow. The last one I tried was a 350w CST/KU123 that would only do about 18mph road speed. Saneaglle and I lifted it on the hoist and joined the high-speed position and it span up to 33mph. It didn't seem to change the power on the road. It felt the same as before but now effectively without speed limit. Unfortunatelly, I didn't get a chance to measure power consumption. I was a bit worried about whether there was any detrimental effect, so I switched back to normal operation, and then my 500w arrived and I changed motors. On BPM/KU93 or KU123, I normally get about 20% increase in speed from the high position. On smaller motors, sometimes nothing, somtimes a little bit.

 

[Jeremy]

It depends a lot on where you buy the controller from, unfortunately. The vendor/middleman programmes these things, as Wuxi Xichen Micro-Electronic Co send them out to dealers unprogrammed. It's the luck of the draw as to what settings have actually been programmed in to any controller, unless you happen to know for sure what the particular vendor is doing.

For example, Ed Lyen tends to programme any Xiechang's that he sells that don't have the three speed switch to 100% at Speed 2 (the mid-speed, which is the controller default with no switch connected). Sheiji Ge used to always programme those he sent out to 66% at Speed 2, but now will set them to 100% on request, I believe.

 

[Gdread]

First thing to try is the 3 speed switch connector, which can give a significant increase. You don't need a switch - just shove a bit of wire up the middle hole, and bridge it in turn one side and then the other. One way will make it slower and the other faster..

Thanks for this tip, just used it to great effect on my green bike kit.

 

[jackhandy]

Thanks Jeremy - I actually understood that

Now, how does that work with brushed motors?.... I was kidding.......I Was Kidding!

Appropos of nuffin - While reading Jeremy's post, I was watching Last of the Summer Wine in which Sid was trying to get a new doorbell wired-in, "helped" by the 3 "Experts".... The juxtaposition was surreal

 

[newguy]

Jeremy I understood every word

So my e bike stopped working, so i think i need a new controller, will start a new topic for that however

just as well i love my bike or it would be in a river right about now!

 

[dmcgoldrick]

Jeremy I understood every word

So my e bike stopped working, so i think i need a new controller, will start a new topic for that however

just as well i love my bike or it would be in a river right about now!

hi newguy........what bike do you have ??? can you ride the bike on the flat at 20-22mph ??? if not and you are relatively fit, then think about putting your kit on a bike that you can ride faster which will probably be better geared for top speed, lighter and 'quick' tyres. your motor would then also probably pull you along quicker ...... if you are wanting assist to 25mph then you need to step up to german s class level but then should not use in any public places of course.......i can ride my converted moulton unassisted at over 20 mph on flat easily........ just has a small motor to assist on the uphills......
regards

 

[piotrmacheta]

Just my 2p worth. I run an ezee motor (350rpm in 20") wheel with an ezee 36V battery in series with a 4s Lipo (used in electric helicopters etc) bought from ebay. It is true that you have to be careful with connections and charge the batteries separately but it works fine as a set up to get more speed and power. As long as the cheap Lipo batteries have a lower capacity to your big main battery then if you accidentally discharge the batteries too far then it won't damage your main battery.
I agree that the best way to get more speed is to go for 48v system with 300 or 350rpm motor.

 

[newguy]

The big problem is the motor, rated for low rpm, but high torque. I am relatively fit, if I get the money together i will be considering a higher RPM motor to assist to higher speeds 22mph+ would great if possible

 

[Jeremy]

The big problem is the motor, rated for low rpm, but high torque. I am relatively fit, if I get the money together i will be considering a higher RPM motor to assist to higher speeds 22mph+ would great if possible

Worth noting that the motor torque/rpm/voltage relationship is often misunderstood. If you have, say, a motor that is rated to spin at 328rpm at 36V, then upping the voltage to 48V will increase the maximum rpm by roughly 30%, but if you keep the same controller current rating then the torque at speed will be pretty much the same.

Similarly, for a given type of motor, then the different winds (slow, medium, fast) will all give the same maximum torque, only the rpm for a given voltage will vary. The reason is that the torque is a function of the number of wire turns around each pole of the motor multiplied by the current, so the more turns you have the less current you need for a given torque. However, more turns means more wire resistance, so more heat, which means that the maximum current is lower for a slow wind motor than it is for a fast wind motor.

The net result is that for a given motor, irrespective of whether it's a slow or fast wind the maximum continuous torque stays about the same.

The exception is for motors that have an alternative internal gear reduction ratio, like the Q100, which is, I think, the only motor I know of that has either a single or two stage internal gear reduction, depending on voltage and rpm.

Generally, faster wind motors can give better performance, even at the low rpm, high torque, end of the performance band, because the lower winding resistance tends to allow the controller to deliver more current at that end of the power curve. A look at the results from the simulator at ebikes.ca Hub Motor and Ebike Simulator with show this point as a change in gradient of the torque curve - the change happens at the point where the motor winding resistance and back EMF combined stop being the major current limiting means and the controller current limit kicks in. If you run through a few combinations of motor and controller current limit you'll see how this changes.

 

[newguy]

I do have trouble understanding the whole relationship, but nothing a bit of bed time reading wont solve,

......and deeper down the rabbit hole...

 

[Jeremy]

I do have trouble understanding the whole relationship, but nothing a bit of bed time reading wont solve,

......and deeper down the rabbit hole...

TBH, I struggled with it for a while.

The simplest way to sum it up is to use the crude rule of thumb that says that the torque a motor can produce is proportional to the amount of copper in it.

You can either have lots of thin wires in the windings (which gives you maximum torque at a low current) or a few fat wires (which gives you maximum torque at a high current). In both cases the maximum torque that the motor can continuously deliver is the same, and determined by how hot it gets.

Both the "lot of thin wires" and the "few fat wires" windings have about the same amount of copper, as that's fixed by the size of the motor and the dimensions of the slots in the stator.