Not simple BLDC controller It RUNS! :)

[Njay]

I do fully agree on a modular design and I've even done it. Breaking in 3 parts (brain, drivers, power switches) is the best choice to end up achieving a high power controller. In that case, design can and should start on the brain. Converters for the drivers and 5V, processor, communications and the interface to the power drivers module.

Arlo, I'm jealous of you! So many guys with the same interests around

[Arlo1]

I do fully agree on a modular design and I've even done it. Breaking in 3 parts (brain, drivers, power switches) is the best choice to end up achieving a high power controller. In that case, design can and should start on the brain. Converters for the drivers and 5V, processor, communications and the interface to the power drivers module.

Arlo, I'm jealous of you! So many guys with the same interests around

Its not that good I never see them. I need more people into this around. I work on this by my self all the time and mostly because not many can understand electricity like me. (I'm not at most of your guys-es levels but around here I do ok)

[Arlo1]

I like rhitee's idea of sensing voltage drop across the low side fet. I think this can be used to feed our trip quite easy and it will help when the fet heats up it will trip faster at lower amps! How long can a fet be held on??

[Njay]

Can be a good idea, but would have to be tested. The temperature and other compensations would have to be done in hw because of the required speed of reaction for this "big" controller.

[jdb]

I would like a faster sensor to protect phase to ground shorts however.

Now we can run the motor with two sensors and do the math; but then you cannot protect that third phase for a phase to ground short. Just something else to think about at a system level as the design matures.

Is this really a problem? A line-to-ground fault shouldn't generate any fault current unless the negative battery terminal is tied to chassis. Is that common? It really isn't a good practice with high voltage systems.

Even if it is commonplace, using low-side fet Rds as a shunt will not detect the fault current in an output line-chassis fault. The path is from B+ through the high-side fet, to the line, to the chassis to B-.

The pictures posted on this board of motor winding faults have been either line to line or line to self, either shorting two phases together or shorting one phase through itself. Both of those faults will be detectable with two-phase current sensing.

[jdb]

I like rhitee's idea of sensing voltage drop across the low side fet. I think this can be used to feed our trip quite easy and it will help when the fet heats up it will trip faster at lower amps! How long can a fet be held on??

It sounds backwards, but if you have a high-speed semiconductor fuse in the battery supply, the best course of action may be to hold the 'fets on and crowbar the fuse! In IGBT systems, the act of turning off the switch while it is conducting high fault current can produce a voltage spike powerful enough to blow the switch. Even if the voltage spike isn't high enough, the I*V switching loss alone can be enough to vaporize the switch, too.

[rhitee05]

Temp compensation would not need dedicated hardware, it could be done in the processor. Temp changes very slowly. All that's required is for the processor to monitor the FET temp (which really should be included anyhow) and then it adjusts the trigger threshold to the comparators if desired. You could only partially compensate for temp, or not compensate at all, and take let the change in Rds be a self-protecting feature. Either way the compensation is a very slow process.

[hardym]

I dont know much about sensorless controllers, it seems a bit magical how they work. However I have build several versions of sensored controller brains using the PIC16F2431 chip. This is a great chip as it is designed for BLDC operations and has everything you need to get going, including cycle by cycle current limit.

If this will help, I'll publish my latest schematic and C software. The software is long and has a lot of debugging stuff in it, but the best part it that it has an output to a serial LCD so you can see what's happening inside your brain. The source uses the Free High-tech C compiler from microchip.

Two files attached: a jpg graphic of the schematic, and a zip file with the source code.

Brain-v2-16F2431.jpg

Brain Schematic

(215.8 KiB) Not downloaded yet

Brain-18F2431.zip

Zip file with software

(42.28 KiB) Downloaded 57 times

[Arlo1]

I just order some IR2113 14 pin dip style, a pickit3 with the 12 code lessons... (might need them ) And a couple dsPIC30F3010-30I/SO Now over the next 2 days I will do some digging and see if I need anything else. Other wise I will start saving for more lipo while I wait and learn code! Man this is exciting we are about to become creationist

[Lebowski]

How about this for a gate-drive (optocoupler + gate drive circuit in 1):

VO3120:

http://www.farnell.com/datasheets/485119.pdf

[Njay]

Arlo hold on, you're burning cash like mad . Decide on the parts and schematics first, then burn only the necessary cash .

That gate driver is very interesting Lebowski, but what were they thinking when they made it Vcc >= 15V!

[Arlo1]

Arlo hold on, you're burning cash like mad . Decide on the parts and schematics first, then burn only the necessary cash .

That gate driver is very interesting Lebowski, but what were they thinking when they made it Vcc >= 15V!

I actually had all of this on order before this thread other then the pic programer and the better fet drivers bigmoose pointed out.
As well I have 2 weeks - 1 month before it all shows up so I am not to afraid to order a little extra. I dont have a electronics store in town I would have to burn 20 bux in gas to go to Victoria so ordering online makes sense. I have a rough idea on the schematic all we need to figure out is the trip section for the overcurrents

[Arlo1]

IM bidding on a set of 7 of these 1MBI800U4B-120 http://pdf1.alldatasheet.com/datasheet- ... B-120.html I want to build a controller with epic phase current to I can have epic torque so..... 1200 amps and up to 1200 volts hmmm Might be able to use these for the Megabike!

What would be the harm running in 100-300 volt range with IGBTs like this?

[Arlo1]

So I just found this http://ecomodder.com/forum/showthread.p ... 4-497.html shows it here http://www.melexis.com/Hall-Effect-Sensor-ICs/Triaxis®-Hall-ICs/MLX91205-689.aspx And the data sheet says the Melexis hall current sensor is 8uS responce time so a little slow but you could do the math and predict where the current is in 8uS more time then set your fail safe to trigger there. I will look into it tonight. Looks uber easy to inmplament!

[Njay]

8us seems very sllooowww...

[Arlo1]

Here is a CSA-1V with 6 uS resopnce time. http://pdf1.alldatasheet.com/datasheet- ... SA-1V.html Its a little better 25% faster is a great improvement. You can scale how many amps it can read by chaning how close it is to the wire or busbar.

[Njay]

Latest datasheet says 8us response time:
http://www.melexis.com/Asset/Datasheet-CSA-1V-DownloadLink-5313.aspx

[Arlo1]

Shane Colton. Has a whole bunch of brushless motor control building info on his blog so Ive been studing it. He has some good tricks I will post some of what I saw asap.

http://scolton.blogspot.com/p/motor-con ... #hexbridge

[Electroglide]

So I just found this http://ecomodder.com/forum/showthread.p ... 4-497.html shows it here http://www.melexis.com/Hall-Effect-Sensor-ICs/Triaxis®-Hall-ICs/MLX91205-689.aspx And the data sheet says the Melexis hall current sensor is 8uS responce time so a little slow but you could do the math and predict where the current is in 8uS more time then set your fail safe to trigger there. I will look into it tonight. Looks uber easy to inmplament!

Any of those current sensors that rely on measuring the B field directly from the trace without a flux concentrator are susceptible to cross talk from other high current traces running nearby. Allegro has a product offering that is somewhat similar too the melexis part too, but they control the trace to sensor relationship which limits the maximum current that you can sense. If your currents are going to be higher than 200Amps, I would suggest rolling your own current sensor by selecting a ferrite core and cutting an airgap so that you can fit a hall sensor. See attached document. I have rolled my own in the past up to +/-2000Amps, with decent linearity. You can also purchase off the shelf sensors from LEM if you want to do it without learning about magnetics. As for the comment about 8usec response time, for simple hall base current sensors, the response times are not usually very quick, but in terms of the typical motors that you might be trying to sense the current in, it is fast enough. Only motors with really low winding inductance and resistance values(small electrical time constants) would not work with them.

[Arlo1]

So I just found this http://ecomodder.com/forum/showthread.p ... 4-497.html shows it here http://www.melexis.com/Hall-Effect-Sensor-ICs/Triaxis®-Hall-ICs/MLX91205-689.aspx And the data sheet says the Melexis hall current sensor is 8uS responce time so a little slow but you could do the math and predict where the current is in 8uS more time then set your fail safe to trigger there. I will look into it tonight. Looks uber easy to inmplament!

Any of those current sensors that rely on measuring the B field directly from the trace without a flux concentrator are susceptible to cross talk from other high current traces running nearby. Allegro has a product offering that is somewhat similar too the melexis part too, but they control the trace to sensor relationship which limits the maximum current that you can sense. If your currents are going to be higher than 200Amps, I would suggest rolling your own current sensor by selecting a ferrite core and cutting an airgap so that you can fit a hall sensor. See attached document. I have rolled my own in the past up to +/-2000Amps, with decent linearity. You can also purchase off the shelf sensors from LEM if you want to do it without learning about magnetics. As for the comment about 8usec response time, for simple hall base current sensors, the response times are not usually very quick, but in terms of the typical motors that you might be trying to sense the current in, it is fast enough. Only motors with really low winding inductance and resistance values(small electrical time constants) would not work with them.

The motor I want to push 600 phase amps through has 8uH inductance so I need a very fast sensor!

[Electroglide]

Ok, that is a small inductance...what motor are you using that has that small of an inductance.... Some ideas for sensing could be to use a di/dt sensor that is integrated to produce a value that is proportional to current. What sort of frequency are you planning to switch your bridge at? You may not need as high of a bandwidth on your current sensor if your switching frequency is high enough though that will ultimately limit the performance along with any dsp/ucontroller that you use, unless you are doing it with ASICs designed for BLDC control.

[Electroglide]

Was Just thinking that if this is a seriously high powered BLDC motor, then your best bet is to go with a current source DC link then use a current steering bridge to commutate the motor. That means that your current sensor bandwidth can be dictated by the size of the inductor used in DC link. How fast you can change the DC link current also depends on the size of that inductor, the bus voltage and the switching frequency, but at least you don't need to run the bridge at a high switching speed. It only needs to run a the commutation frequency.

[Arlo1]

Collossus. I am realy wondering about sensing a section of the phase wires for voltage drop. And just figure out the coraltion of voltage drop vs amps!

[Arlo1]

Some of the cool ideas I lernt form Shane Colton include a 17v zener diode which helps stop over voltage from being aplied to the gate and a 1k resistor on the gate to - which helps hold a fet open when not in use handy for when you turn on the controller and if something fails.

[Electroglide]

It took me a little, but I found what at least the collossus is, or what I think it is. If it is the one that HAL is involved with, and by looking at the pictures, it hard to believe that it has only 8uH per phase inductance, especially since they are quoting numbers like 11mohm for phase resistance(not sure if it is a phase to phase measure measurement or phase only). I also saw that it is a delta configuration too. if it works well, it will make a nice compact power solution. As for measuring the drop on your phase wires, it will not work very well because you have both the resistance and the inductance of the wire to deal with. V= Ldi/dit will be the biggest problem because of the frequency dependency. The resistance is another problem because of the positive temperature coefficient. The hotter the wire the higher the resistance. That why the sense resistors have a very low inductance and very low ppm/C. If a kelly can drive one of these motors, then I have to believe that you are dealing with electrical time constants that fall with the bandwidth of traditional hall sensors.

The zeners on the gates is a blessing and a curse. Yes they will help clamp unwanted high voltage spikes which are normally caused by bad layouts, but they also create more ringing on the gates which can lead to higher losses or higher EMI. The pulldown is a good thing if using ASIC drivers and you plan to push your vehicle and it is a direct drive. I got burned on just that a couple of years ago when I had controllers failing when the vehicle was being pushed with the power off. The backemf from the motors caused the body diodes in the bridge mosfets to start conducting back into the capacitor bank which caused the gates to become charged via the miller induced gate charge which eventually caused the bridge to turn on and short circuit the motor. The motor currents at those voltages got very high and blew the mosfets....who would of thought LOL. Anyways, if your miller charge is not extremely large you probably can get away with a 10k resistor on the gates.