Alien Gents Special II - Improvments and Modifications

[jbond]

I decided to dig out the oscilloscope today and have a look at the drive signal on one of the motor phases. What I found was interesting and not what I was expecting.

As I mentioned above the three motor phases are driven by a pulse width modulated signal and what I was expecting to see was a noticeable variation in the duty cycle of the signal and perhaps variation in the pulse amplitude, however, between the three modes there was only a small variation in duty cycle and the amplitude was the same between modes.

I checked at full speed in each mode and found going from low to medium to high the duty cycle was: 50, 55 and 60%. The pulse amplitude measured 40volts, the same as the battery output for each mode, no variation. The big variation was in the switching frequency which I calculated as 111, 161 and 208Hz.

So it would appear the switching frequency is the main parameter that affects speed. I'll post some 'scope pictures later, maybe tomorrow.

Presumably you did this with the wheel off the floor under no load. So the controller will be running the motor up to it's speed limit and then cutting the power to maintain that speed so there's just enough power to overcome friction at that speed. It's presumably doing this by cutting the duty cycle so that the current drops to just above zero. I'd expect a smaller duty cycle than you found to do this, but maybe the back emf and speed means torque has already dropped off considerably so ~50% duty cycle is needed.

Everything I've read about parameterdesigner (Endless-sphere.com • View topic - Keywin E-bike Lab Parameter Designer Software Manual ver 1.2) is that the three speed limits can be set from 0-120%. 0-100% seems to be a percentage of the free running speed affected by the motor winding and some global limit. 100-120% adds in some advance on the phase angle to provide a kind of turbo boost. I can't find much about default values but I suspect the controller comes with something like 60%-100%-120% Which then corresponds with your 10-15-19mph. Thinking about this, changing the DC and Phase current limit is going to give more or less acceleration and hill climbing ability but not change the speeds. So going to a 20A controller from a 15A controller with everything else the same will give more power, but not more top speed.

Two things I'm still trying to find out.
1) A way of getting the stock values out so there's a base working setting you can get back to. The software at the moment doesn't seem to allow this, so it may involve emails to ecrazyman to find out the stock values.
2) What the 100% speed limit actually means in terms of switching frequency and if that can be increased. Without that I suspect there's no more speed to be had from a given geared hub motor regardless of how much power we feed into it.

For reference, if I lift the rear wheel on the Aurora and apply full throttle I get the same max speeds in each mode as when I'm actually riding. Not too surprising. But again, pointing at a switching frequency limit.

 

[NRG]

Presumably you did this with the wheel off the floor under no load. So the controller will be running the motor up to it's speed limit and then cutting the power to maintain that speed so there's just enough power to overcome friction at that speed.

Yes, off the ground. There's no ramp up and cut when applying battery voltage with he throttle wide open, the motor phases are fed with battery voltage and switched at X frequency with Y duty cycle. The only ramp / change in duty and switching frequency is when the throttle is progressively used.

It's presumably doing this by cutting the duty cycle so that the current drops to just above zero. I'd expect a smaller duty cycle than you found to do this, but maybe the back emf and speed means torque has already dropped off considerably so ~50% duty cycle is needed.

I think Andrew mentioned this in the Bafang thread the speed limits due to the back EMF.

Everything I've read about parameterdesigner (Endless-sphere.com • View topic - Keywin E-bike Lab Parameter Designer Software Manual ver 1.2) is that the three speed limits can be set from 0-120%. 0-100% seems to be a percentage of the free running speed affected by the motor winding and some global limit. 100-120% adds in some advance on the phase angle to provide a kind of turbo boost. I can't find much about default values but I suspect the controller comes with something like 60%-100%-120% Which then corresponds with your 10-15-19mph. Thinking about this, changing the DC and Phase current limit is going to give more or less acceleration and hill climbing ability but not change the speeds. So going to a 20A controller from a 15A controller with everything else the same will give more power, but not more top speed.

Correct, climbing has improved but top speed is the same. I will know more when I get the cable but reading the ES threads I think the speed setting is more like 50-75-100%, I'm not sure if 120% is being used....I'll know in time.

Two things I'm still trying to find out.
1) A way of getting the stock values out so there's a base working setting you can get back to. The software at the moment doesn't seem to allow this, so it may involve emails to ecrazyman to find out the stock values.
2) What the 100% speed limit actually means in terms of switching frequency and if that can be increased. Without that I suspect there's no more speed to be had from a given geared hub motor regardless of how much power we feed into it.

Yes, a good point, I'll have to ask what the default parameters are. We know these motors are wound with different RPM in mind for wheel size vs voltage so the adjustable parameters available to us may be limited in their scope. However, raising voltage does increase RPM and that may be the only practical route to take.

For reference, if I lift the rear wheel on the Aurora and apply full throttle I get the same max speeds in each mode as when I'm actually riding. Not too surprising. But again, pointing at a switching frequency limit.

Interesting, thanks for the data point.

 

[jbond]

Yes, off the ground. There's no ramp up and cut when applying battery voltage with he throttle wide open, the motor phases are fed with battery voltage and switched at X frequency with Y duty cycle. The only ramp / change in duty and switching frequency is when the throttle is progressively used.

I think Andrew mentioned this in the Bafang thread the speed limits due to the back EMF.

You see I think this is where it gets interesting. The switching frequency has to be in phase and at the same speed as the motor is running so the resultant speed comes from both sides. If the controller was unlimited in freq, it would just keep on speeding up the motor, staying in sync with it, until the motor reached it's back emf limit. And as others have said, this is ultimately set by the input voltage. But in low mode, we clearly haven't reached that limit, you're finding that it's still feeding 40v but the no load speed has dropped. So the controller must be limiting the frequency and/or dropping the duty cycle so that the motor doesn't accelerate any faster against drag. I *think* that if it just limited the frequency, the motor would naturally sync with that frequency. If it started to over-run, the phase between poles and magnets would pull it back.

I have this suspicion that the noload maximum speed of the SB hub motors is actually something like 30mph at 36v, the actual 18-22mph max we see is a combination of controller freq and amp limits and hence either the motor just can't generate enough power to go any faster, and/or the controller is limiting the speed in the same way that it limits the speed to 10mph in low mode. And I guess whether we can release this extra performance is what we're trying to find out.

 

[NRG]

Yes as I noted above the Freq and duty cycle changes in between the three modes.

Saw your post on ES, the two controllers I have are 116 based so the documents will be useful. I've asked E-Crazyman if he has a default parameter file. Also found this link which could prove useful:

Google Translate

 

[NRG]

Controller fun aside I've now got the lights working with a simple wiring loom switched via the handle bar switch. Also fitted a rigid seat post from Halfords, something I should have done a long time ago the ride is now much better.

The speed switch is really useful and the bike far less frantic to ride but with improved hill climbing ability. It's a tricky upgrade but a really worthwhile one, it has transformed the ride-ability of the bike and highly recommended.

 

[John L]

Controller fun aside I've now got the lights working with a simple wiring loom switched via the handle bar switch.

Nice one! From your previous experience with the lights - are they a big drain on the battery?

John

 

[NRG]

I'll need to check John, they are LED front and rear so I doubt they draw too much power...

 

[John L]

I'll need to check John, they are LED front and rear so I doubt they draw too much power...

After asking the question I thought it really depends on what kind of light we're talking about. I guess leds won't take that much power. Are they the alien lights?

John

 

[NRG]

Yes, the standard fit ones on the GSII. I supplement them with much more powerful add on lights but I wanted to retain their use. I'll see if I can measure them later today.

 

[lessped]

having trouble with noisy front brakes on our wisper...cleaned them as per Fleccs instruction and things where much improved. but after the lay up over Christmas squeaking like crazy again maybe new pads then.......

I always find a little rub down with some fine emery cloth seems to take the squeel out of the brake block's

 

[jbond]

Another question I'm afraid. Do you know what motor this bike has?

 

[NRG]

Front SWXB from Bafang.

CHEAPBUTGOOD——swxb series -Suzhou Bafang Electric Motor Science-Technology Co.,Ltd

 

[wurly]

I 've been reading this with interest.
I can say that my tongxin controller gives a frequency of 90hz. It doesn't offer any other modes. The same controller runs a bafang8fun hub at 15mph. It will also run my other 260rpm hub at 23mph(in a 26" wheel).

Considering the rpm of the motors before they get geared down to drive the wheel. I reckon' they are spinning pretty fast already. But selecting a different mode(frequency), the motors can cope with lower speeds ok.

But, if it is possible to program a controller to for higher frequency, would the motor be able to give power and torque in the same way?

What i am asking is, to get the motor to drive the wheel faster, the motor itself will have to spin considerably faster for you to notice a speed increase because of the gearing...i wouldn't have thought it possible if the motor is wound for a certain rpm???

I'm another person who would like to increase the speed of my hubs without going for higher volts or selecting motors for smaller diameter wheels.
I think i need to hook my motors and controllers up to the 'scope again and see what i can find. i'll report back.......

 

[NRG]

Any input most welcome Wurly, I'm still waiting for the programming cable to arrive....

 

[wurly]

I've corrected the Tongxin controller frequency after looking up my previous notes when i tested it.

Tomorrow i will see what my standard 2-mode controllers give with a bafang hub. I don't have a 'scope at home unfortunately.
I noted (unscientifically) that they will reach 18.6mph (unlaced/unloaded with a bike speedo+magnet stuck to the hub) which i worked out to be around 235-240 rpm. Certainly faster than using them with a tongxin controller! it must be something to do with signal frequency....this is why i am hoping your experiments to get the motors running faster work out.

I should be able to get my 260rpm Tongxin hub (currently in my commuter bike) to really fly with a different controller....................

 

[jbond]

My understanding as of today...

Top speed on geared brushless hub motors is a combination of things.
- Motor winding. Various numbers of poles.
- Internal gearing
- Rim diameter
- Voltage
- Power
- Controller limits
- Battery limits

Brushless motors have a natural maximum speed set by the winding and voltage. This is the point where the back EMF equals the input voltage and current-power drops to zero. The input voltage affects this. More volts = more rpm for this no-load speed.

The internal gearing then turns this maximum motor speed into a maximum hub speed. The trade off is in torque. More speed, less torque.

Rim diameter then turns this hub rpm into road speed. Again we have a torque trade off.

The controller limits the available current and also sometimes limits the max frequency the motor can run at.

The battery limits the available current according to the C rating.

And finally, the available power (voltage times available current) may not be enough to reach the speed required.

So what we get right now is hub motors designed for a specific voltage, winding and gearing to work with a specific rim diameter to meet a design speed, often 25kmph to meet the EU legal requirements at 24/36v. They then pick a cheap controller to match.

If we bump the voltage to 48v, fit a 20" wind and gearing hub into a 700c wheel and put a higher current limit on it, we should be able to get a higher top speed and still retain the torque by putting more amps through the motor at lower speeds. And with a 10Ahr battery should be able to do short bursts of 20A or 20*48 = 960W. Which all adds up to 25mph or so and enough hill climbing to wind us up any realistic hill at 10-15mph.

However, if you already have a motor designed for 26" 36v and rated at 250w, just bumping the current limit from 10A to 15A might get you from 17 to 19mph and 400w along with climbing a hill at 12mph instead of 10. The extra power doesn't buy top speed because you're hitting the no-load speed limit not really the power limit.

I'm beginning to think about that first option. Take a strong, light, simple steel frame rigid bike. Fit a small motor like the Tongxin/Nano, Cute85 or Bafang SWXH/SWXK in a 20" wind, laced to 26" or 700c rim . Run it with a 48v10Ahr LiFePo Ping battery and a 48v20A ecrazyman controller.

Gearless hub motors and RC motors take the compromises in opposite directions. Gearless like the MagicPie and 9C motors naturally rotate much slower at wheel speed. So they're noload speed limit is much much higher. But with no gears they have to be large and heavy to get the torque and need and can take much more power. Speed limits are power limits not noload rpm limits. RC motors are the reverse. Small, light and very high revving which then needs lots of gearing to bring it down to wheel speeds. The problem is that the gearing has to be home built as it's not needed on helicopter rotors!

ps. A lot of those figures are guesswork! They give an indication of the principles not real facts.

 

[NRG]

Yep, I'll go along with most of that, I think though you meant lower speed more torque for the geared motors, the Bafang are geared down by 4.3:1 increasing torque.

I read a lot last night on ES including the DB 120% thread and it seems there may be a small gain to be had by using this setting but it really will depend on the controller and tolerances if it is going to be of any use. It seems timing is not altered with this setting just the Duty cycle. 100% dosen't mean 100% duty cycle its somewhat less for practical reasons it seems and this is seen on the 'scope traces I saw on my controller.

It was also reasurring to read that the three way speed switch on the 116 based controllers does limit (actually scale) throttle response and not power as I suspected.

I'm begining to think the only practical way to increase speed is via upping volts, sure, controller parameters will help fine tune things but for all practical purposes volts is the way to go initially. Still, it will be interesting to try this and there may be some benefit with standard 36v battery/controller combos.

 

[wurly]

Mmmmm interesting.......
For my own benefit (and others), i was really interested in the difference between my 12A tongxin controllers and my dual mode 15A controllers. So i took the controller and hub into work today and had a play.

I have also discovered that the dual mode controller also uses different frequencies as NRG had discovered, although it would appear on the lower setting the duty cycle is also reduced. I had assumed that the lower speed setting simply reduced the duty cycle, it didn't dawn on me that the signal frequency was changed!!

This picture shows the waveform on the highest mode. 175hz

And this is the lower setting 98hz(notice the pulses and reshaped waveform)

So how does this affect the speed on my bike?
On the way home using my tongxin controller (90hz) and 260rpm hub , i could manage an easy 21mph(pedalling of course).
Next. a no load test by lifting the front wheel off the ground, speedo registers 22.3mph.
Now i swap the controllers over and do the no load test.
26.1mph on the highest(173hz) and a legal 15mph on the lowest setting(98hz).

Although i have not tried it out on the road yet as it's already getting dark.
By swapping controllers and increasing the signal frequency I seemed to have gained 4mph.................which is nice, but i have lost speed on the lower setting even though slightly higher frequency (reduced duty cycle has much to do with this).
I will spend tonight installing the controller properly and see how it performs tomorrow and post.

 

[jbond]

Normally, I'd expect the controller to sync with the motor and not the other way round. The sync is going to come from hall effect sensors or from the controller synthesizing the timing from watching the two unpowered circuits. So under no-load, the controller should just follow the motor up to it's no-load maximum rpm. So if the phase timing is the same, you'd expect the two controllers to produce the same no-load speed and hence frequency. If it's not doing that then one or both of the controllers is limiting the frequency and hence limiting the motor rpm. This is what happens I think in low and medium mode. Perhaps it's also happening in the Tongxin controller in order to keep it legal.

If your other controller runs at 26mph no-load, maybe you will be able to get to 23-24mph on the road with enough amps/power and before the motor's torque drops off.

 

[John L]

Hi guys - thanks for sharing this info I don't understand much of it but I get the gist!

Wurly - It'll be interesting to see whether the motor actually runs faster under load. For commuting purposes I won't mind giving up some torque for a little more top speed.

Cheers

John