Paul & Sabrina's cheap DIY 144v motor controller

[mora]

Originally Posted by adamj12b

On a side note. I am currently working on developing an expansion board that will bring all these new features that have been added in hardware to the users with the original control board. It will be a simple connection to the main board with a 6 pin ribbon cable that will attach to the ISP connector that is in place.

On this board, I will be including the precharge circuity, bus voltage measurement circuitry, and I was thinking a tach pulse input from motor with a tach pulse output to drive an in-vehicle tach with software scaling so that it will display the right RPM. What do you think?

This is great. Go for it. This must be the second best thing after software RPM limiter.

[sawickm]

Hi Adam,

Good stuff !!! Sign me up for one of your PCB sets (controller - driver - expanded I/O). Just add your new proposed power board and you have a very nice system !!!

Soon. lol real soon. I will be able to have them rapidly populated if I get enough orders together. These 2 boards were not worth the hassel of putting all the different value parts into the feeders and then the feeders into the machines. BUT, if im making a panel or 2 at a time, that will be totally worth it.

Don't tell me that’s your own conveyor driven reflow oven in your garage. And I suppose you have a paste stencil and pick-n-place machine too !!!

On a side note. I am currently working on developing an expansion board that will bring all these new features that have been added in hardware to the users with the original control board. It will be a simple connection to the main board with a 6 pin ribbon cable that will attach to the ISP connector that is in place

Why are you connecting your expanded I/O PCB into the ISP port? Is the ATMEGA8 I/O maxed out ??? Is there another compatible ATMEGA that has more I/O ?

Good job,

-Mark

[Grimm]

Wow, that surface mount version is awesome! Would it be a lot cheaper if a bunch of people order one?
Also, any way you could give us a new current/planned feature list now that we're in the hardware adding phase again?
Another thought... is there some way to store the code on some sort of memory card for those of us who don't have the tools to program the at processors? Or maybe just store the log data on a usb/sd card so you don't have to have a computer with you in the car to record all the data like current, temp, and any software faults or errors? Kind of like what the PackTrakr does, only better. That would be awesome!

[dcb]

integrating a vehicle speed sensor signal might have more than just monitoring energy per mile/speed uses.

some possibilities include
1. detecting wheel slip.
2. rev matching for faster clutchless shifting, plug in the ratios and have one small pushbutton finger operated switch on the shifter to select rpm for the next gear (based on vehicle speed) and one to select the previous. controller can send a pulse or a brake to quickly find the next rpm.

[Grimm]

Originally Posted by dcb

integrating a vehicle speed sensor signal might have more than just monitoring energy per mile/speed uses.

some possibilities include
1. detecting wheel slip.
2. rev matching for faster clutchless shifting, plug in the ratios and have one small pushbutton finger operated switch on the shifter to select rpm for the next gear (based on vehicle speed) and one to select the previous. controller can send a pulse or a brake to quickly find the next rpm.

That sounds pretty sweet. Makes me think that this thing should get some code for cruise control as well after the speed sensor is installed. With the speed sensor you could have it act just like any other manual and give you a little light when you should change to a lower gear to get the absolute best range. Also, with fine RPM control and motor amps we could totally make it adapt to icy roads automatically. It wouldn't be too hard to make it determine when the wheels break loose and then have it give just enough power to get moving without making the wheels slip. Try doing that in an ICE!

[Tweety]

Originally Posted by sawickm

Why are you connecting your expanded I/O PCB into the ISP port? Is the ATMEGA8 I/O maxed out ??? Is there another compatible ATMEGA that has more I/O ?

Good job,

-Mark

Probably because the pins not used aren't pulled out to a connector, the ISP port is available in the boards already made... The next version should really bring any and all unused ports to a connector...

[adamj12b]

Originally Posted by sawickm

Hi Adam,

Good stuff !!! Sign me up for one of your PCB sets (controller - driver - expanded I/O). Just add your new proposed power board and you have a very nice system !!!

I will be sure to let you know when I am ready to start shipping kits!!

Originally Posted by sawickm

Don't tell me that’s your own conveyor driven reflow oven in your garage. And I suppose you have a paste stencil and pick-n-place machine too !!!

No, I dont have my own reflow oven. That is at my friends shop. He has 4 of those things!!

Originally Posted by sawickm

Why are you connecting your expanded I/O PCB into the ISP port? Is the ATMEGA8 I/O maxed out ??? Is there another compatible ATMEGA that has more I/O ?

Good job,

-Mark

We have decided to connect the expansion board to the controller through the ISP port for a few reasons.
- First, the original controller did not have signal conditioning resistors on the extra pins. This doesn't allow us to create a new board and allow the old ones to run the same firmware. By adding expansion through an external device allows all controllers to be able to run the same firmware. This method allows users to upgrade their controller with just plug and play.

- Second, the ISP port is actually a high speed serial data connection. After the firmware is loaded onto the micro, it is never used again. The 6 pin connector has the 3 data lines, reset, and 5V power. This means that the only connection needed between the 2 boards will be a 6 pin ribbon cable.

Originally Posted by Grimm

Wow, that surface mount version is awesome! Would it be a lot cheaper if a bunch of people order one?

YES!! If I can get a large order together before they go into production, prices will be lower. This will have to do with bulk ordering the expensive parts like micro controllers and opto-isolators.

I hope to keep some boards in stock after the initial run, but they will be slightly more because of smaller qty runs.

Originally Posted by Grimm

Also, any way you could give us a new current/planned feature list now that we're in the hardware adding phase again?

I will list what will be added with the expansion board.

- Bus Voltage Measurement
- Internal Pre-Charge Control
- Tach Input/Output w/ Scaling
- 1 Button Cruise Control w/ Brake Input
- Through Hole "Play Ground" for Custom Expansion

Originally Posted by Grimm

Another thought... is there some way to store the code on some sort of memory card for those of us who don't have the tools to program the at processors? Or maybe just store the log data on a usb/sd card so you don't have to have a computer with you in the car to record all the data like current, temp, and any software faults or errors? Kind of like what the PackTrakr does, only better. That would be awesome!

Yes!! You could use something like this : SparkFun Electronics - Logomatic v2 Serial SD Datalogger

It has inputs for a serial terminals. it takes a micro SD card which the adaptor is basically an SD card! I've never used one, But it says it can take 19200 baud serial data so it should work.

Originally Posted by dcb

integrating a vehicle speed sensor signal might have more than just monitoring energy per mile/speed uses.

some possibilities include
1. detecting wheel slip.

The only problem I have about integrating with the vehicle speed sensor is every make can be different.

Detecting wheel spin could still be done with just a tach sensor on the motor as the VSS will be inside the transmission which is close enough to the motor to work the same.

Originally Posted by dcb

2. rev matching for faster clutchless shifting, plug in the ratios and have one small pushbutton finger operated switch on the shifter to select rpm for the next gear (based on vehicle speed) and one to select the previous. controller can send a pulse or a brake to quickly find the next rpm.

Rev matching could be a cool feature. I will have to think of this more. The only problem comes back to interfacing to the VSS. The other thing that im not a huge fan of is connecting to the VSS is more intense of an install. You will have to locate the wire that is the VSS, and tie onto it and go back into the controller.

Originally Posted by Grimm

That sounds pretty sweet. Makes me think that this thing should get some code for cruise control as well after the speed sensor is installed.

The cruise control system I came up with would keep circuitry, wiring and installation very simple. The method I have came up with would require mounting 1 momentary push button for on/off and adjustment and a wire attached to the brake pedal sensor switch under the dash. The control switch would be be held down for 1 second to activate cruise control. 2 quick push's of the button would increase speed and 1 push would decrease speed. When the brake pedal is pressed, the signal will be sent and turn off the cruise control.

Originally Posted by Grimm

With the speed sensor you could have it act just like any other manual and give you a little light when you should change to a lower gear to get the absolute best range.

To do this you would probably need to have an efficiency variable in the controller that if you are over that, you the light would come on. Problem with that idea would be on the highway when you w/h per mile is high, the light would always be on...

Originally Posted by Grimm

Also, with fine RPM control and motor amps we could totally make it adapt to icy roads automatically. It wouldn't be too hard to make it determine when the wheels break loose and then have it give just enough power to get moving without making the wheels slip. Try doing that in an ICE!

I have thought about this before. It could be a very handy thing. It will take some more looking into and talking to Fran to see if this is possible. I have a feeling it might be harder then you think without removing and changing some parts of the firmware that are in place now.


So one other thing I think I forgot to mention is that this controller expansion will require that the main boards micro controller be upgraded from the atmega8 to the 168. The mega 8 doesn't have enough room to implement ANY thing else, including the serial communications to the new board.

Originally Posted by Tweety

Probably because the pins not used aren't pulled out to a connector, the ISP port is available in the boards already made... The next version should really bring any and all unused ports to a connector...

Yes, All of my bords will have expansion connectors and pull down resistors on all unused pins to allow for safe operation of firmware with features that are not implemented in hardware.


So what do you guys think?

-Adam

[dcb]

Originally Posted by adamj12b

.
The only problem I have about integrating with the vehicle speed sensor is every make can be different.

You are preaching to the choir here But it would make a nice optional feature for those who want that sort of functionality (mileage/speed performance tracking, anti spin, anti-lock,cruise,easy on your synchros shifting) and are willing to sort it.

Originally Posted by adamj12b

.
Detecting wheel spin could still be done with just a tach sensor on the motor as the VSS will be inside the transmission which is close enough to the motor to work the same.

Not quite. If like most FWD cars you have the speed sensor tied to one wheel, then that is how you detect when a wheel is spinning. The motor rpm picks up but either the wheel is spinning too fast for the motor rpm or it is too slow, and the rpm to speedo ratio does not match any known ratios. Regen braking is the same way, but that is another issue, for another controller.

Originally Posted by adamj12b

Rev matching could be a cool feature. I will have to think of this more. The only problem comes back to interfacing to the VSS. The other thing that im not a huge fan of is connecting to the VSS is more intense of an install. You will have to locate the wire that is the VSS, and tie onto it and go back into the controller.

yup, we wrestle with vss a lot with the mpguino. Instant and average Distance/speed is such a useful metric in things transportation though that it is worth some wrestling:
MPGuino - EcoModder and a bit of wire cost a lot less than a gps module and is far more precise on instantaneous values (good enough to shift by I recon).

Originally Posted by adamj12b

The cruise control system I came up with would keep circuitry, wiring and installation very simple.

It looks like plenty of adc and interrupt pins are still open, what I will probably wind up doing is using a resistor network on an adc if spare pins are tight, to put several buttons on one pin. Personally I would rather look up a button function on a range of values than do "signal processing" on it.

And again the speed signal could tell you if they put the clutch in or slipped into neutral if you know the ratios (could also train the ratios with some interface).

I'm just thinking out loud here, nice to have for the most part.

[Grimm]

Originally Posted by adamj12b

I will be sure to let you know when I am ready to start shipping kits!!

Please do!

Originally Posted by adamj12b

YES!! If I can get a large order together before they go into production, prices will be lower. This will have to do with bulk ordering the expensive parts like micro controllers and opto-isolators.

I hope to keep some boards in stock after the initial run, but they will be slightly more because of smaller qty runs.

Count me in for now, when will the initial run start? (Hoping to get the control board done in the next 2 months...)

Originally Posted by adamj12b

I will list what will be added with the expansion board.

- Bus Voltage Measurement
- Internal Pre-Charge Control
- Tach Input/Output w/ Scaling
- 1 Button Cruise Control w/ Brake Input
- Through Hole "Play Ground" for Custom Expansion

Sounds good! Is contactor control still being worked on? Or is it just half turn of the key to precharge the controller, full turn to close the contactor?

Originally Posted by adamj12b

Yes!! You could use something like this : SparkFun Electronics - Logomatic v2 Serial SD Datalogger

It has inputs for a serial terminals. it takes a micro SD card which the adaptor is basically an SD card! I've never used one, But it says it can take 19200 baud serial data so it should work.

Thank you! That should work well. Do all the current testers have laptops? They would probably love this.

Originally Posted by adamj12b

The cruise control system I came up with would keep circuitry, wiring and installation very simple. The method I have came up with would require mounting 1 momentary push button for on/off and adjustment and a wire attached to the brake pedal sensor switch under the dash. The control switch would be be held down for 1 second to activate cruise control. 2 quick push's of the button would increase speed and 1 push would decrease speed. When the brake pedal is pressed, the signal will be sent and turn off the cruise control.

Maybe a button like a screenkey where you could put the speed/RPM in the center? It's easily controlled by atmels. And when it's not engaged it's green (ENGAGE! ), when you engage it turns red? I dunno how expensive the cheap ones are though... Might not be worth it, just a thought for brainstorming. Alternatively, you could use these to switch from eco-driving to drag racing, maybe with the same single key.
Screenkeys - getting started, evaluation/prototyping kits, Controlling and Interfacing with programmable LCD keys

Originally Posted by adamj12b

To do this you would probably need to have an efficiency variable in the controller that if you are over that, you the light would come on. Problem with that idea would be on the highway when you w/h per mile is high, the light would always be on...

I'll have to ask my dad (who's worked on ECM code for a million years) but I think the way that manuals do it is directly proportional to RPM and time spent there, if you maintain around the same rpm in a gear (so it doesn't trip while you accelerate to the next one) when shifting to another would be possible without dipping below say 1500 in the RPMs then it reminds you with the light that shifting would be a good idea.

Originally Posted by adamj12b

I have thought about this before. It could be a very handy thing. It will take some more looking into and talking to Fran to see if this is possible. I have a feeling it might be harder then you think without removing and changing some parts of the firmware that are in place now.

You're the coder, feel free to play with it or not if you have better things to code (like cruise control )

Originally Posted by adamj12b

So one other thing I think I forgot to mention is that this controller expansion will require that the main boards micro controller be upgraded from the atmega8 to the 168. The mega 8 doesn't have enough room to implement ANY thing else, including the serial communications to the new board.

Dang, we filled it up fast. Maybe someone will come along and trim down the code a bit to at least fit in the serial communications?

Originally Posted by adamj12b

So what do you guys think?

-Adam

Sounds great! Color me extremely interested.

[dcb]

re: cruise, probably the simplest is "on or off", with brake/throttle override. you select the speed you want to go with your foot and throw the switch. resume/accel/coast are all nice but they provide much diminished returns compared to an rpm lock switch.

re: mega 8, just spend $3 on an atmega 168 (or maybe 328) please? The cost difference is peanuts especially compared to the confusion and extra maintenance having cpu options causes.