cheap controller ??

[gwhy!]

do anyone know anything about these controllers ? they seem to be a good price..
http://stores.ebay.co.uk/VirtualVillage ... =241440152

[cohberg]

i'm sorry I dont know anything about those controllers but if you are looking for a cheap controller there is a good 1500W controller that some of us in the forum have been using.
The only reason i'm suggesting it is because its cheaper and because it would take equally as long for both to ship to you. It would be easy to mod it for 48V usage and you would have 100V 50A support if you needed it

[gwhy!]

Seems a very good price for a 72v controller, thanks for the heads up.

[gwhy!]

do anyone know anything about these controllers ? they seem to be a good price..
http://stores.ebay.co.uk/VirtualVillage ... =241440152

Ok I got one of these controllers to play with ( to add my power stage too ) I have just set it up to make sure it all works before I start messing with it. for some peoples information it runs sensorless mode if the correct timing from sensors is not present which I was not expecting but on a big plus note the sensorless startup seems really strong I may have to put it on a bike and see just how good it starts up under really big loads.

[bobale]

So you're saying this 48V 600W controller is working in sensorless mode?

[gwhy!]

So you're saying this 48V 600W controller is working in sensorless mode?

yes, I have just done some load testing on it with a couple of my rc motors ( 170kv motors) . It works very well sensorless ( im impressed ) it do suffer from the cutting out problem at mid throttle ( if using sensors or sensorless ) but this can be sorted.

[gwhy!]

In sensorless mode it only appears to work well upto around 4500rpm then starts mis-firing ( 14pole motor ). When running with sensors this is much higher, I dont know how high yet because i need to stop the cutting out mid throttle but I have measured upto 6500rpm before it cuts out. the cutting out is to do with the age old problem of hardware phase current cutoff the same as with the infineon controllers.

[zombiess]

Anyone know what board is used and if it's programmable?

[gwhy!]

Anyone know what board is used and if it's programmable?

there is a program header on the board, I will do some pin tracing tomorrow from the controller chip to see if it ties in with the infineons

[LHelge]

I also bought one of these a while ago, the 48 V 350 W version. It had 6 x STP75N75 as a power stage and is controlled by a STM8 microcontroller. I'm pretty sure you can't program them in the same way as the Infineon controllers but you can download source code for a BLDC controller from ST's homepage. Maybe the designers of this controller have used it without modifying it to much, then you could write your own firmware.

I replaced the FETs in mine with IRFB3006 instead, beefed up all current traces with 3 x 1,5 mm copper wire and replaced all high power cables with 6 mm^2. Yesterday I tried it for the first time with a Turnigy 80-100 rewinded for 90 rpm/V. 350 W @ 48 V would mean a current limit of ~7 A which wasn't enough to start the motor in sensorless mode. I had to manually place the rotor in balance between two cogging locations first to get it running. After modifying the current shunt to approximately 1/3 of the resistance the motor started up every try. My next step is get the sensors, that I installed on the Turnigy, running.

I plan to run this controller on 12S LiPo, how many amps do you think it could handle with my setup? Could I push it up to 50 A?

[gwhy!]

I also bought one of these a while ago, the 48 V 350 W version. It had 6 x STP75N75 as a power stage and is controlled by a STM8 microcontroller. I'm pretty sure you can't program them in the same way as the Infineon controllers but you can download source code for a BLDC controller from ST's homepage. Maybe the designers of this controller have used it without modifying it to much, then you could write your own firmware.

Thats someting well worth looking into

I replaced the FETs in mine with IRFB3006 instead, beefed up all current traces with 3 x 1,5 mm copper wire and replaced all high power cables with 6 mm^2. Yesterday I tried it for the first time with a Turnigy 80-100 rewinded for 90 rpm/V. 350 W @ 48 V would mean a current limit of ~7 A which wasn't enough to start the motor in sensorless mode. I had to manually place the rotor in balance between two cogging locations first to get it running. After modifying the current shunt to approximately 1/3 of the resistance the motor started up every try. My next step is get the sensors, that I installed on the Turnigy, running.

I will be intrested to see what you come up with

I plan to run this controller on 12S LiPo, how many amps do you think it could handle with my setup? Could I push it up to 50 A?

Just had a quick look at the fets and they have a very reasonable spec, I dont see any reason why you cant run these upto about 70A in this controller. I have the 75v version of these fets in a infineon controller that I push at around 80A without any problems except the controller do get a bit warm.

[LHelge]

A small load test with the controller running in sensorless mode. The throttle is set to the lowest possible and the battery voltage is about 46 V, with the current reaching at most 5.5 A it is around 250 W.


Will write a little bit more about the controller and my modifications when I find the time.

[LHelge]

I've written a little about this controller here:
Cheap controller

I will look into creating my own firmware for the STM8 processor.

[amberwolf]

If you do, could you see about making an optional SIN/COS sensor input (to be used instead of the three halls)?

It would enable the controller to be used with external encoders that arent' exposed to the hot conditions most hall sensors are inside motors, and also to use motors that are already designed with SIN/COS sensors instead of the ebike-typical 3 halls.

Of course, I have an ulterior motive: with such a controller, I could easily use the powerchair motor I have here:
viewtopic.php?f=30&t=32838
without any modifications.

[LHelge]

That would probably not be possible without changing the hardware, depending on how the microcontroller AD pins are muxed with the digital.

I would say that there is a better chance of either replacing the sensors in the motor or making some kind of circuit to transform the sin/cos to normal Hall pattern easliy done with a microcontroller with something like this pseudo-code:

Code: Select all

sin = getADSin();
cos = getADCos();

deg = atan2(sin, cos) * 180 / pi;

PORTA = (deg > 120 && deg < 300);
PORTB = (deg > 180 && deg < 360);
PORTC = (deg > 240 && deg < 60;


I haven't read your whole thread but from the pictures it doesn't look like distributed windings (often called PMSM instead of BLDC) so the gain from a sinusoidal drive would not be that big. Easiest way to determine is measuring the back-emf.

Maybe I'll order another controller to reverse engineer, would be a fun project, playing around with sinusoidal drive as well.

[amberwolf]

I wasn't really thinking of sinusoidal drive, but rather just using the SIN/COS sensors (whcih output their namesakes' waveforms for the positioning data) instead of typical 3-hall sensor setups.


Since the encoders on each motor would be differnet, it would probably require programmability for the relationship between the SIN/COS outputs and the alignment with the phases/stator teeth.

If there is programmability, then the hall sensor inputs could also be programmed to work as analog rather than digital in, if that's what's being used. If not, well, then another pair of analog inputs (or externally muxing the SIN/COS signals to use with a single analog input on the MCU) would be necessary. I dont' know if there are any unused pins on the MCU, much less if any that do exist are also analog inputs.

The STM8 is a neat little chip;
http://www.st.com/internet/mcu/class/1738.jsp
I have one here in a USB stick, though I can't seem to get the hang of programming well enough to do anything with it beyond the downloadable demos and such. The STM32 would be really nice for a controller chip on ebike controllers, but I'm not aware of one that uses it.

[gwhy!]

Ok This controller works very well in sensorless mode but there is a problem when using it with sensors, the 2 systems conflict with each other and this is what is causing the cutting out.. there must be away of turning the sensorless mode off completely so some more detective work needs to be done

[amberwolf]

Got some pics of the PCB, both sides, clear enough and large enough to see the markings? Might be able to decipher what they mean, if you can also make a chart of what markings go to what known wired-in functions, so we can eliminate those when guessing about the un-wired (or unknown but wired) pads.