Paul & Sabrina's cheap DIY 144v motor controller

[MPaulHolmes]

Hey squiggles! You little Devil ('s advocate!) hehe. I am really really grateful for your analysis. You are right about the while loops. They have all been eliminated. Everything is done with interrupts now. We need to upload a newer version of the software to the wiki..

I'll pass along your suggestion about the layout possibly diminishing the isolation to the engineer that's doing the testing of the control section in the lab. Thank you for raising that concern!

timer1 is initialized in the function InitPWM();

The NAND gate circuit is responsible for automatic overcurrent shutdown. It's a little tricky, but here it goes!

First, here's the overcurrent protection circuit, taken out of the main circuit:


I'm calling high signals 1 and low signals 0. This is the sequence of events -

I. IF CURRENT OVERFLOW SUDDENLY HAPPENS
1) R_BAR: 1 ---> 0
2) S_BAR: 1 ---> 1
3) Q: ??? ---> 0
4) Q_BAR: ??? ---> 1
5) PWM output to gate driver: ??? ---> 0

Because the TRIP_STATE (which is Q_BAR) is now 1, the microcontroller now knows that the mosfet driver has been disabled. I can re-enable the driver at my leisure.

II. IF YOU CLEAR S_BAR (TRIP_RESET)
1) S_BAR: 1 ---> 0
2) Q_BAR: 1 ---> 0
3) Q: 0 ---> 1
4) R_BAR: 1 // if R_BAR is back to 1, all is well, so let's pretend that.
5) PWM output to gate driver is still 0!

III. IF YOU THEN SET S_BAR (TRIP_RESET)
1) S_BAR: 0 ---> 1
2) R_BAR: 1 // LET'S PRETEND
3) PWM output to gate driver is now 1! Hurray, finally we have power!
///////////////////////////////////////////////////////////////////////////////////
Delay(uSecond) would be way better. I agree.

BY THE WAY, THE OVERCURRENT CIRCUIT HAS BEEN MODIFIED SINCE THE MOSFET DRIVER IS NOW THE INVERTING TYPE. That circuit above was for the original controller (version 1).

[squiggles]

Gate circuit drawn and described like that make sense.
Funny but I would not of thought of including it, would have done it all with the micro. It occurs to me that you have added another level of security as this will work even with a hung processor. Very nice.

Hmm, just had another thought, the over current protection will be a bit instant. Maybe a good thing, can't think of any real downside. Is it self resetting...no. Is there a circumstance where a short time over-current condition is acceptable ... I guess not better to be safe than sorry. All good.

[MPaulHolmes]

Yes, the hardware overcurrent shutdown happens in maybe 3 or 4 uS. In theory, with a well tuned PI loop, the overcurrent shutdown may never have to come on, but it's there as a last line of defense. Also, it lets you tune the pi loop without worrying about destroying the mosfets.

[vinny1989]

Originally Posted by MPaulHolmes

Yes. I used it to run my electric bike hub motor. Well, with 200v components, you don't want to go much over 144v, since when the pack is full, it's quite a bit higher than 144v. You do need a 12v auxiliary battery for the controller though. The control section uses less than 5 watts, so you could even use the last battery (connected to ground) in your series string to power it, and it's such a puny power requirement that I don't think it would get too out of balance, but I don't know.
...

That. Is. Awesome.

For my idea, id need the 12v auxiliary anyway..

[bdrehmer]

Great work by everybody so far, especially Paul! Very inspirational.
Looking at the schematics, it looks like the unused IO pins on the microcontroller don't go anywhere. If that is true, would it be possible to have those unused pins going to a pad? I think that would make it easier for people to extend the controller in the future by adding their own features (the beauty of open source!). One idea I had for an extension was an anti-theft feature. Maybe different features could be enabled by jumpering some of the unused pins high, which the end user could easily do if the pins were available on the controller board.

On the topic of USB devices to replace your hobbled oscilloscope, check out this DAQ:
NI USB-6008 - Products and Services - National Instruments
I could probably get a discount (~10%) on it if you are interested.

[MPaulHolmes]

Hey bdrhemer! Thank you for the encouraging words! I've never seen something like that before. It looks like a very helpful tool. I should definitely get one eventually. I may have to wait for something like that until the fall, though.

Back to ways it can fail! This is fun!

If someone hooks it up such that they basically bypass the motor completely, (like hooking both motor cables to the same post post of the motor) there would be almost no inductance to slow the current rise, which I believe would destroy the mosfets. The hardware current shutdown time of 3 or 4 uSeconds would be too slow I believe, but someone should try it just to see what happens! hahaha!

[patx]

Originally Posted by dcb

Build a few and try all those scenarios and get back to us.

I have gathered most parts now for my conversion... All that is missing is, batteries, controller, charger + desulfater... I would like to build Paul's controller right away, I think there is no point in buying a 1500$ curtis... Are the files on the wiki are the latest ?!?

What about the POWER Board ?!?

Where to get the aluminium plate for heat ?!?

On the BOM, the cristal is not listed... what is the part number ?!?

Paul said somewhere that R16 needed to be more than 1k... what value R16 needs to be !?

Originally Posted by MPaulHolmes

...You are right about the while loops. They have all been eliminated. Everything is done with interrupts now. We need to upload a newer version of the software to the wiki..

Can you please update wiki with all latest file...

Maybe a few more beta tester will build their own if the files are available... Atleast I will

Originally Posted by MPaulHolmes

By the way, somebody out there keeps hogging all the good mosfets from Digikey! They got their supply of them replenished less than a month ago! Maybe Curtis is stockpiling them? hehe. Now they won't ship until June 22! makes me mad!

A few years ago I was into console modchips, I made a few of those ( the latest i made was the xbox 8164 solderless adapter)... I still have a few contacts with chinese pcb boards manufacturer and I sent them your fab files for prices... Chinese manufacturers have their own parts suppliers witch are very very cheap... I will let you know if I have something better than june 22nd...

Thanks again for this awesome work and pls forgive my english

[squiggles]

Being picky again, and I am sure I am stating something that has been said before. However in my opinion this

PORTB &= (128 + 64 + 32 + 16 + 8 + 4 + 1); // this sets PB1 (the pwm output) to 0 at startup. Notice '2' is missing!


Is better written as
PORTB &= 0xFD; // this clears PB1 (the pwm output) at startup.

I appreciate that hexadecimal may be unfamiliar at first, believe me if you are going to be programming regularly it is highly advisable to become familiar with HEX.
Also in programming speak you don't "set" something to "0".

[dcb]

too picky

[squiggles]

I know, I know, but it is all about readability. In two years time when you have all worked on dozens of other programs and you pick this one up to make some changes you will realise the difference