New Infineon v Old PIC Controller - A 48V Bafang PMGR Test

[Knuckles]

UPDATE! September 5, 2008

Turns out I had the WRONG controller wire combination that had a bizare effect of actually spinning the motor FASTER but with super high no-load current.

See this spreadsheet for an explaination of testing the Bafang and the Infineon ...
View attachment Phase-Hall-Wire-Combinations-Bafang-Infineon.xls
-K

 

[fechter]

The Infineon ran the Bafang motor at a much higher rpm at the same exact voltage!

F'n'A!

-K

Interesting.

Can you measure how much faster? It would seem that the new controller has some degree of timing advance. This may or may not be a good thing. 7 amps seems like a lot for no load. Of course the current under load is generally more important.

It would also be interesting to measure the current from both at exactly the same speed.

If you have an old bicycle speedometer around you could at least get a comparative measurement.

 

[dnmun]

a digital cyclo computer type speedometer on the wheel would be very accurate for comparison.

i have an idea of how to create dynomometer using the Bafang hub motor. a dyno will be useful to determine how efficient the controller is by evaluating the amount of power produced by the motor, when driven by either of these controllers, or any modifications to the new controller, or with other motors like the BMC/puma and C'lyte.

imagine a test rig built such that the tire of the test wheel drives a small DC motor or generator, or belt driven from around the rim to the pulley on the driven DC motor. if you use an alternator, you can use a fixed current to excite the field of the alternator so that the alternator would produce the same output current under identical driven conditions. then you can load up the output of the driven motor with a resistive load like a small electric wire wound heater core. measure the output current and output voltage through the load to determine the power output into the load.

then structure a test where you maintained the 48V and adjusted the current from your power supply using the throttle, and recorded the speed from the cyclo computer and at the same time recorded the current and voltage going into the driven load. do this at different voltages too if possible.

from the two output charts (one for each of these controllers), for a given input power (voltage and current from the power supply) the output power would give you a true comparison of the efficiency of the controller itself.

for modifications of the controller such as adjusting the timing or perhaps even swapping out the power fets this simple dyno would be very useful, and actually could be the format for comparing the different types of controllers, and if you were able to record the temperature of the heatsink on the fets, this may help understand how to tune the controller itself to reduce the power lost in heating the fets from switching.

it would be helpful for wayne to step on my post and clarify what would work best here. maybe keywin would want to do this back at shenzen, let his bossses know this would be a major step for establishing their equipment's performance, and providing feedback on how to continue the improvements in the controller. get him some credibility and visibility too.

thank you knuckles, BOL, dennis

 

[John in CR]

Knuckles,

If that same speed relationship holds up under load with my DD hub, I may have to drop my voltage booster pack, because I'll be dangerously fast then. Yeah right, I'm going higher because the Infineon can take more. gimme speed! Booster batts, booster motor (geared Bafang) for takeoffs and hills, plus I'm already doing some stuff to reduce wind resistance and lose weight, so I should be really flying. I can't wait for the package man next week.

John

 

[Knuckles]

Interesting.

Can you measure how much faster? It would seem that the new controller has some degree of timing advance. This may or may not be a good thing. 7 amps seems like a lot for no load. Of course the current under load is generally more important.

It would also be interesting to measure the current from both at exactly the same speed.

If you have an old bicycle speedometer around you could at least get a comparative measurement.

Very Interesting indeed! I don't have a speed meter yet (CA not yet connected on these - coming soon).
But I was at 1/4 throttle on the Infineon at 48V and hit 2 amps and the no-load rpm seemed identical (by observation) to the old controller at full throttle!

FULL INFINEON THROTTLE ... I mean this thing was spinning FAST! The 7 amps seemed reasonble just to overcome the internal gearing losses at this high rpm.

I was totally surprised by this no-load test because I had never seen this very cool behavior before. (Ha Ha the motor was cool too!)

Even more ironic ... as I had set this all up I heard from Mark at TH in UK. He said his wheelchairs were doing 30MPH on the Infineon! :lol:

Also, Keywin told me the Infineon "auto detects" 60 or 120 degree hall phase advance so I guess this is a "very smart" lil MCU.

Rich ... You have the Infineon already. Just change the power resistors (R01 A,B,C,D) and the R6 resistor and make it a 72V controller real easy.

And Keywin also found a neat trick for installing the sensor-less module. He said it installed super fast on the Infineon ...


Blue "thingy" is variable LVC POT. The "lil" green board is the sensor-less module.

-K

 

[Knuckles]

Knuckles,

If that same speed relationship holds up under load with my DD hub, I may have to drop my voltage booster pack, because I'll be dangerously fast then. Yeah right, I'm going higher because the Infineon can take more. gimme speed! Booster batts, booster motor (geared Bafang) for takeoffs and hills, plus I'm already doing some stuff to reduce wind resistance and lose weight, so I should be really flying. I can't wait for the package man next week.

John

I don't think it can. My DD motors are not drawing more than 30 amps. But I do start off much faster and I do burn rubber if I want.
Mark in UK actually wants a "softer start" mod to tone this down. Me ... I LIKE IT!

I just think I am seeing what fechter called "phase advance". Or maybe some kind of "auto phase advance"? Like I have a clue?
On a down hill run? Maybe? Maybe this thing will keep accelerating going down a hill? I just don't know!

But also be warned. I think the 45 amp controller is doing more like 60 amps! Too much solder in the shunts (by accident).
(Rumor Mill says Infineon already melted some French dude's 1500W e ATV motor! Cool! The knucklehead had no in-line fuse.)

Ha Ha. I told Keywin this is great! 8) 60A is better than 45A anyway! But DEFINITELY use a fuse to the bat +. ALWAYS use a fuse.

-K

 

[wrobinson0413]

.

 

[ZapPat]

I finally had my first real infinion test drive yesterday, after installing a makeshift torque arm. First observation: I think Knuckles is right about the infinion being able to pump out over 50A for sure, since my ~45A breaker setup was poping pretty easily if I accelerate hard enough (and it was poping often believe me). Great torque - even using my old golden motor hub (direct drive) it will spin on grass and gravel while accelerating hard!

As for Knuckle's 7amp no load current draw versus 2 amp on his old controller: I don't have the same thing happen to me, but I am not using the same motor either. I measure 2A current draw at full no-load speed, which happens to be 70kph according to my cycle odometer!!! It does seem that it is spinning faster at no-load than with my old controller, but I cannot test this for the moment, since I have to fix the old controller before comparing. BTW, my old controller was not the PIC-based ecrazyman, but a moded golden motors one (moded from 36V-20A to 48V-35A).

As for real-road speed, I got up to 55kph so far on flat ground, so a little ways from the 70kph no-load speed. I got about the same top speed when using my old GM moded controller (actually got up to 62kph on th old one, but I'm sure this new one will be able to do this in the right conditions).

 

[The7]

Also, Keywin told me the Infineon "auto detects" 60 or 120 degree hall phase advance so I guess this is a "very smart" lil MCU.

IMO. Don't think this is good because any high voltage spikes/ noise could confuse the Infineon.

 

[Knuckles]

See my update and spreadsheet at top of this thread ...
http://endless-sphere.com/forums/viewtopic.php?f=2&t=6189&p=92602#p92602

"One test is worth a thousand opinions"
Dang Right it is!

 

[ZapPat]

Also, Keywin told me the Infineon "auto detects" 60 or 120 degree hall phase advance so I guess this is a "very smart" lil MCU.

IMO. Don't think this is good because any high voltage spikes/ noise could confuse the Infineon.

Could you expland on this, The7?

Personnaly I don't think noise would be a problem on the hall inputs, and that's all the micro needs to detect 60o/120o. What's important is that the routine used to detect this should be well made to avoid any confusing behavior were it might shift back and forth between states or something like this. I actually wonder if the lowest-throttle motor-rattle I observe on the infinion may have something to do with this? I use a large direct drive motor with it (golden front hub, 500W...more like 2000W!) has anyone else tried this combination?

Also, I removed the excess solder from the shunt, and now my breakers hold no matter what. I read about 40 amps max controller current now vs over 60 amps before the light mod. It still has really good starting power but it is easier on the battery, so I like this lower limit better.

As for efficiency, I'm reading about 10 amps on the flat at a bit over 35kph (~24mph), I don't think this is too bad. However, at about 55kph I draw about 30-35amps, and I don't know if this is "normal" for a non-aerodynamic rig like mine?

 

[Knuckles]

I use a large direct drive motor with it (golden front hub, 500W...more like 2000W!) has anyone else tried this combination?

Also, I removed the excess solder from the shunt, and now my breakers hold no matter what. I read about 40 amps max controller current now vs over 60 amps before the light mod. It still has really good starting power but it is easier on the battery, so I like this lower limit better.

As for efficiency, I'm reading about 10 amps on the flat at a bit over 35kph (~24mph), I don't think this is too bad. However, at about 55kph I draw about 30-35amps, and I don't know if this is "normal" for a non-aerodynamic rig like mine?

Well My Grubee black motor is almost identical to the GoldenMotor. I also noticed that very slow rpm "chatter" but I really have to go very slow on the DD motor to make it occur (like 1-2 MPH). IMO it is almost a non-issue because it never happens to me during "normal" riding.

The energy required to overcome wind resistance increases rapidly over 15-20 MPH (it rises exponentially). Wind resistance is the greatest "robber" of energy over 20 MPH (on the flat that is). Going slower and pedaling greatly increases your range and is way more efficient. Me? I go as fast as I can!

You removed solder from a shunt! Ha Ha ... that's the first time I EVER heard that in this forum!

 

[Nogwin]

I have now over 100 miles on the Bafang motor, and am extremely pleased at how efficient it runs. So far over a 100 miles averaging 9.5whr/mi, which compared to the X503's 18whr/mi is amazing.

Now I had trouble getting it to run on the Xlyte controller, and using the original motor wiring details from Knuckles, I too had very high loads at full throttle, and the noise was horrible at WOT. Then played with the phases and came up with -

Hall Wires
Control Motor
B B
G Y
Y G

Phase Wiring
Control Motor
B B RUNS GOOD
G Y CORRECT
Y G CONFIGURATION

So the motor at WOT (66V) will spin up quietly, and have a load of under 2A!!

So I knew there was more too this as indicated by his spreadsheet, which ironically had the same data as mine (except for one combo of hall wires)

Here are the speeds on the road for a given voltage (Xlyte 36-72v analogue, 35A)

48V - - 22.5mph flat no pedal
60V - - 27.5mph
66V - - 30.0mph

So amp load at 22.5mph (WOT) is around 175-250watts, and that gets me 24 miles each way to work (220whr used). Actually better than double the efficiency of the X503 at similar speeds. I'm just blown away.

Nog

 

[Knuckles]

WOW!

That reported efficiency blows me away too!

Twice the efficiency at half the weight!

AWESOME!

 

[docnjoj]

While my new Infineon is super on the Bafang, I tried Nogwins wire combos on my Crystalyte 20 amp 36-72 volt old analog and it didnt work at all! I dont understand why the Crystalyte was so poor!??? I need to know incase its a bad controller and I get another motor. BTW I did try all spreadsheet combos, and the best I got was the 12 amp no load. For cmparison, the Infineon runs 1 amp no load at 42 volts!
otherDoc

 

[John in CR]

It sounds like I should have ordered 2 motors, and sacrifice one to find out the real voltage limit. I need my speed, so is a gear change possible for a speed Bafang up front and climber motor in the rear? Maybe I'll have to turn a Bafang into a bottom bracket motor with variable gearing like I saw one guy do with a hub motor.

John

 

[dnmun]

i think bafang makes another motor called the 'hillclimber'. it may be brushed, but i think it has more than the 5:1 of all the other bafang motors. what is the rational for having 2 motors and 2 controllers? do you have the same battery? two throttles too i guess.

 

[John in CR]

i think bafang makes another motor called the 'hillclimber'. it may be brushed, but i think it has more than the 5:1 of all the other bafang motors. what is the rational for having 2 motors and 2 controllers? do you have the same battery? two throttles too i guess.

more power up hills and for acceleration. 2wd traction off road. all while still maintaining high efficiency on the flats. yes 2 throttles and controllers. hopefully one pack. Also, full backup for one failure and cooler running due to alternating use. btw I want speed gearing, I should have hills covered already.

John

 

[Knuckles]

Keywin told me Bafang discontinued all their DD motors so they only make geared motors now.

But I have run the Bafang up to 72V but that is really pushing it IMHO.

 

[Knuckles]

While my new Infineon is super on the Bafang, I tried Nogwins wire combos on my Crystalyte 20 amp 36-72 volt old analog and it didnt work at all! I dont understand why the Crystalyte was so poor!??? I need to know incase its a bad controller and I get another motor. BTW I did try all spreadsheet combos, and the best I got was the 12 amp no load. For cmparison, the Infineon runs 1 amp no load at 42 volts!
otherDoc

Well ... What does this tell you? That old x-lyte is only good for DD motors and not good for geared.
Bada Bing!

 

[The7]

While my new Infineon is super on the Bafang, I tried Nogwins wire combos on my Crystalyte 20 amp 36-72 volt old analog and it didnt work at all! I dont understand why the Crystalyte was so poor!??? I need to know incase its a bad controller and I get another motor. BTW I did try all spreadsheet combos, and the best I got was the 12 amp no load. For cmparison, the Infineon runs 1 amp no load at 42 volts!
otherDoc

Well ... What does this tell you? That old x-lyte is only good for DD motors and not good for geared.
Bada Bing!

Don't think this comment is valid.

 

[dnmun]

knuckles, i think the motor that bafang calls the hillclimber is still a geared motor, with a gear reduction greater than 5:1 of the regular brushless motor, obviously for more torque and lower speed, you may be able to find it in their catalog. do you and keywin have access to the other finishes on the hub motor? or just the black and silver type.

 

[docnjoj]

The gearing on the geared Bafang (whichever I have) is really good for hills with enuff volts. I think a minimum of 36 LilfePo4 but 48 would sensational. Now that I have the controller issue sorted, I have a 48 volt Nicad and a 36 volt Lifepo4 to try out. Will try both this week! I think it would be more economical for Bafang to just change the # of coil turns rather than try to adjust the sun/ planet/ring gear values???? That gets tricky.
otherDoc

 

[dnmun]

i think the hillclimber might be an older model, since it is a brushed motor. and it was continued in production to meet more utilitarian purposes than the brushless motors. i thought that motor might help john in CR to get the torque he wanted in the second motor. but he would have to put the geared motor on the rear to get the power to the ground or just have the front wheel spin on the uphill. there really is only one bafang brushless motor it appears, just different finishes on the cases. all are 250W at 24V and 350W at 36V, i think someone has already counted the gears to get the 5:1 reduction.

 

[Microbatman]

In the fist few post we see a blue square on the back of the board.

You said it is a LVC pot.

Translation you can adjust the LVC (low voltage cuttoff) using this device

Is this included as a standard item from the factory on the new inferion Controller or do you put it on after?
What is the adjustment range?