Paul & Sabrina's cheap DIY 144v motor controller

[vwdevotee]

Originally Posted by adamj12b

In order to fit the controller inside the amplifier case, you would have to basically redesign the power section from scratch. It would require a special design to fit circuit board and most likely caps that are larger in diameter and shorter in height. Its defiantly a possibility to do, but time consuming. I was working on going this route once, but stopped because of many factors.

As for tach output, The controller has the ability to output a pulse to run a tach with a little bit more hardware, but you first need to feed a tach signal into the controller to know the speed the motor is spinning. This also requires some hardware.

There currently code being worked on and tested to monitor and limit RPM based on one of the input pins on the ISP port that is not used after initial programming.

As for driving a temp gauge, This would just require some hardware and some code.

-Adam

Dang, I finally reached the end. Almost 4500 posts is definitely a marathon read . About 2000 posts ago, someone posted a powerboard sketch with the heat spreader hoizontal rather than vertical, was that just a one off thought experiment, or did anything come of that? That was kind of the way I was thinking of bringing the height down to fit in the amp case. I'm still thinking about the caps. Perhaps a whole bunch of smaller ones in parallel. That wouldn't be cost optimized, I'm sure, but would bring down then height. Also, about 1000 posts ago, there was talk of using motor voltage and a motor constant to calculate the RPM in the MCU without actually needing any new hardware. Did that fizzle or would that still be an option? In order to use the MCU to drive one of the factory gauges, what would the additional hardware consist of? Unrelated, instead of a precharge relay, would it be possible to use another mosfet? It seems like anything solidstate is going to be just a bit more robust than somethign electromechanical. I'm an ME, so I'm a bit out of my element, but learning a little more EE each day. Actually it's kind of funny because the EE at my work is starting to look at me funny when I show up almost every day with rather specific technical questions.

Jackbauer, do you have schematics for your Bauer-con posted anywhere? Or are they functionally similar to one of the Cougar revs? Just with an IGBT instead of a mosfet array?


Well, now I'm off to rad the rest of the threads in the subforum. None of those are too long, so hopefully I can cruise through them at lunch tomorrow. Or while I'm dodging real work.

[jackbauer]

The Bauer-con is basically an igbt output stage. Some modifications to the control board will help such as using a dc dc with a 15v output and the little two resistor and ultra fast diode circuit posted way back to help with the turn off.

[jyanof]

Originally Posted by vwdevotee

Also, about 1000 posts ago, there was talk of using motor voltage and a motor constant to calculate the RPM in the MCU without actually needing any new hardware. Did that fizzle or would that still be an option? In order to use the MCU to drive one of the factory gauges, what would the additional hardware consist of?

It's been implemented in the form of overspeed protection. It doesn't calculate a specific RPM, but calculates a number; if that number is above the threshold you provide, power to the motor is cut.

The calculation could probably be modified to produce an actual rpm value or a proportional PWM signal. It will be very inaccurate at low speeds and will also vary in accuracy at different output currents. For example, I noticed that the overspeed would trip at lower RPM if the motor was lightly loaded. The difference was small though, maybe 400 RPM between an unloaded motor and a highly loaded motor.

Also, no calculations are performed when you're coasting (since motor voltage is zero) so the tach would also go to zero. Might be a little weird to get used to...

Lastly, this calculation only works with series wound motors, hence the modifications required for Jack and his compound motor (i think it's a compound, anyway).

If you try it though, let us know how it works! I've had a dead tach for 2 years now...

btw, you can feed a traditional tach with the signal produced by a couple of magnets on the motor tailshaft (or shaft adapter, i guess) and a home made sensor of a couple hundred turns of fine magnet wire around a screw. Put the screw/magnet wire close to where the magnets pass and they will produce a voltage as they pass by. Use a 12v zener to clamp the output to 12V and feed it to your tach signal input. That's the jist of it, let me know if you want more details...

[vwdevotee]

Originally Posted by jyanof

It's been implemented in the form of overspeed protection. It doesn't calculate a specific RPM, but calculates a number; if that number is above the threshold you provide, power to the motor is cut.

The calculation could probably be modified to produce an actual rpm value or a proportional PWM signal. It will be very inaccurate at low speeds and will also vary in accuracy at different output currents. For example, I noticed that the overspeed would trip at lower RPM if the motor was lightly loaded. The difference was small though, maybe 400 RPM between an unloaded motor and a highly loaded motor.

Also, no calculations are performed when you're coasting (since motor voltage is zero) so the tach would also go to zero. Might be a little weird to get used to...

Lastly, this calculation only works with series wound motors, hence the modifications required for Jack and his compound motor (i think it's a compound, anyway).

If you try it though, let us know how it works! I've had a dead tach for 2 years now...

btw, you can feed a traditional tach with the signal produced by a couple of magnets on the motor tailshaft (or shaft adapter, i guess) and a home made sensor of a couple hundred turns of fine magnet wire around a screw. Put the screw/magnet wire close to where the magnets pass and they will produce a voltage as they pass by. Use a 12v zener to clamp the output to 12V and feed it to your tach signal input. That's the jist of it, let me know if you want more details...

I had another thought while at work today, what about hijacking the rev counter for use as am ammeter? Instead of reading 0-7500 1/min, maybe set the scale to 0-750A. I suspect the hardware off the MCU wouldn't be to tough; maybe an isolated op amp to boost the pulse signal to something the rev counter will be able to read.

Originally Posted by jackbauer

The Bauer-con is basically an igbt output stage. Some modifications to the control board will help such as using a dc dc with a 15v output and the little two resistor and ultra fast diode circuit posted way back to help with the turn off.

OK Jack. Do you have any pictures of your system? I'm curious about how you arranged your capacitor bank. After I read the rest of the ReVolt sub-Forum, I'm going to watch all of your YouTube videos.

[vwdevotee]

I was looking at the Rev D drawings, and it looks like all of the pins have been taken up by current features. With absolutely no real basis for it, I was looking at DACs and it looks like this one <http://www.rohm.com/products/databook/general/pdf/bh2220fvm-e.pdf> would be a good fit. It looks like it could run the three gauges I want to use (Tach as a big ammeter, Fuel as SOC, and temp as, umm, temp) with some op-amps. Am I reading the schematic wrong, or has it changed so that some pins would be available? Or would it just be simpler for something like this to switch over to an XMega since those can be had with up to 16 ADCs and 4 DACs onboard? Would most if not all the code port over from the Atmega168 to an Xmega MCU? Thanks in advance.

[DJBecker]

Originally Posted by vwdevotee

I was looking at the Rev D drawings, and it looks like all of the pins have been taken up by current features. With absolutely no real basis for it, I was looking at DACs and it looks like this one <http://www.rohm.com/products/databook/general/pdf/bh2220fvm-e.pdf> would be a good fit. It looks like it could run the three gauges I want to use (Tach as a big ammeter, Fuel as SOC, and temp as, umm, temp) with some op-amps. Am I reading the schematic wrong, or has it changed so that some pins would be available? Or would it just be simpler for something like this to switch over to an XMega since those can be had with up to 16 ADCs and 4 DACs onboard? Would most if not all the code port over from the Atmega168 to an Xmega MCU? Thanks in advance.

I believe that the best way to expand is to hook up a MCP2515 CAN bus controller on the SPI port. It only takes one free pin as a chip select, and gives the most flexibility.

Once you have a CAN network, many different options are available without changing the motor controller. It's easy to add features such as custom gauges, and even cruise control. We previously were using a ScanGauge for reporting, and have now upgraded to an Android phone for reporting. The latter is via a OBD2 bluetooth module on the CAN bus and the "Torque" app.

Our long term plan is to put a controller in the dash to drive the instrument cluster gauges based on the values read from the motor controller, but that's a lower priority since it we already have the info on the Android display, complete with data logging.

[MPaulHolmes]

I FINALLY ordered the current sensor PCB today. I thought I could make them on the mill, but the 7.5mT sensor has few capacitor requirements than the 25mT version, so I had to send the dang thing in to have it made! 0.125" thick. They just shipped the control boards, power boards, and driver boards today finally. Jerks. So pretty soon I'll be able to ship the stuff to everyone that got something. If anyone wants one of those "1200 amp" current sensor boards and a bill of materials, let me know and Sabrina can send a paypal invoice. By the way, it can also be a 2000 amp current sensor just as easily (just one more piece of pcb underneath the main one). I think those will be $5. That will help me make back the stupid $50 or whatever it was for the stupid "tooling fee".

[mora]

If 1000 amp version doesn't use less ham (or was it HASS from LEM), then I want one.

mmm... ham

[MPaulHolmes]

hahahaha. I love ham too! Ya, the 1000amp version doesn't use the LEM HASS, since it only goes up to 900 amp. That board will arrive on "today + 12days + 2 days shipping time". or so... I would have ordered it right away, but I was broke! haha.

[Greg Fordyce]

Originally Posted by MPaulHolmes

That will help me make back the stupid $50 or whatever it was for the stupid "tooling fee".

Hey Paul, for smaller boards you might want to check out PCB prototyping service

10 boards up to 5cm by 5cm for $12, and no tooling fee. They also do boards up to 10cm x 10cm.

Greg