NEW! Uprising Control Series (Logic and Driver Boards)
 

Hello Everyone,

I have put together a design for a new motor control board based on the Revolt design. I will be calling this board the Uprising to keep confusion to a minimum.

I hope to direct traffic about the Uprising board to this thread as to not clutter Paul's already MASSIVE thread!

Thank You.

-Adam


V. 2.0.0


http://farm8.staticflickr.com/7015/6...733a1c7d_z.jpg
Untitled by AdamBrunette, on Flickr

23 Pin Weatherproof Automotive Connector: One of the biggest things I wanted to include was an easy way to disconnect the controller. I have found an affordable 23 pin weatherproof connector meant for automotive applications.

5 Universal Inputs: Each one of these inputs is capable of accepting a signal from 3.3 to 15v without damage. This means you can feed 5v TTL signals and automotive 12V signals into the controller. Each input is independent of the other so they can be used in any combination. One of these inputs has a special function of accepting a tach input signal.

4 Universal Outputs: Each of these outputs is supplying 12V from the controller, Common to the vehicles 12v system. This means you can easily run indicators that are tied to ground. One of these outputs is wired special for tach output as well.

Ethernet Communication: I have included ethernet communication on the board for easier configuration. Software will provide a web interface that will be accessible from anything device with a web browser (iPhone, Android, ect.)

Built in Pre-Charge: This is done with solid state switching inside the logic board. Just 1 lead to outside the controller, on the live side of the contactor is all thats required.

Bus Voltage Monitoring: The controller measures the voltage inside the controller at the cap. This will be used for pre-charge and for motor and battery voltage limiting.

Universal Driver Output: The output section of the controller took alot of thought. I wanted something flexible, powerful and unlimited. What I came up with I think will fit the bill. The board is capable of driving mosfets directly, indirectly, or driving IGBT driving modules. I think thats everything! :D Directly, the controller can be connected to 10 or so fets through some simple breakout and everything will work good. Indirectly, There is also the ability to driver remote mosfet drivers at high voltage to resist noise and have alot of drivers. And , IGBT's. The driver chip can be changed out to a non-inverting driver to drive VLA-500 modules, and jumpers reset to provide 5V high current output.

Throttle Options

Single 2 Wire Potentiometer
Dual 2 Wire Potentiometer (Inverted or Normal)
Single 3 Wire Potentiometer
Dual 3 Wire Potentiometer (Inverted or Normal)
Single Hall Effect Sensor
Dual Hall Effect Sensor (Inverted or Normal)

Other Features: I decided to keep serial on the controller because I had room and if people want to poll or tell the controller to do things over simple serial they would still be able to. Throttle section is the same, as well as contactor control, temp sensor, current measurement and under-voltage circuitry.

This controller uses a very different AVR though. I have moved to a Atmega 644p. This provides 32 IO, and I used all but 4!

Now, Some of you will be disappointed in my next decision, but believe me, it was not made lightly.

I have designed the board in all surface mount components, except for optos and the DC-DC.

This brings me to my next statement. After prototyping is complete, I will be offering pre built logic boards. These will be professionally assembled on assembly machines. Once a few boards are made and purchased, I hope to always keep a few controllers on hand ready for shipment. They will be tested and loaded with the latest firmware.


Schematic: http://farm7.staticflickr.com/6112/6...71d098cd_z.jpg
Uprising Rev2 Sch by AdamBrunette, on Flickr

Layout: http://farm7.staticflickr.com/6051/6...074cf2bc_z.jpg
Uprising Rev2 by AdamBrunette, on Flickr

Current Revision: 2.0.4

11/23/11
I have finished populating 90% of the prototype of version 2.0.0 . While doing this, I made a video showing the process of hand building an SMD board. Enjoy.
http://www.youtube.com/watch?v=515_1g9G7Mw

Very Nice Layout ! :thumbup:

:D

Thanks for starting this thread. Your sharing of this process is very valuable. Folks can now feel a lot more confidence in building SMD boards.

Good job!

Can't wait for your next posting.

Congrats.

Eric

Nice, I like it already..

What are the specs for this board?

I'm tagging along this thread

Wow I must admit although I am no great fan of smd's I really like this unit, wish I have one.Keep it up Adam.

Hi Adam please keep the info flowing can't wait to know more , guess most are getting curious / interested as I am.

HAHA Im glad people are interested.

I have hit a little bit of a road block financially. I have to wait till my birthday rolls around in a couple weeks to order the last parts for the logic board. I still needed dc-dc converts, one of which will take some extra time to get because its special order. I also forgot the optos for the inputs so i cant test those yet.

Ive been working on wiring up a test stand for it and adding wires to the connector.

On monday the driver boards will be here. I was graciously donated the money to build them and plan to get that tested right away. I was hoping I could use the uprising board to do the testing of the driver board, but im going to have to use the old revolt board for pwm signal.

In other news, Ive got most of the 1kA controller built! The water plate is done and the cap and modules are mounted. All I got left is bus bars and install some new washers that came yesterday.

Here are some pics of the controller.

http://farm7.staticflickr.com/6117/6...129031e2_z.jpg
Untitled by AdamBrunette, on Flickr

http://farm8.staticflickr.com/7018/6...84055db1_z.jpg
IMG_1285 by AdamBrunette, on Flickr


-Adam

This is awesome!

Nice Adam - lets throw it in a Battle-bot :eek:

Nice work Adam!

Which isolated DC-DC(s) are you short?
I have a spare for a P&S controller for certain. Would that work?

Mark

Hey Mark,

That dc-dc is of the same series, but is the wrong output.

The new board uses 2 dc-dc converters for full 3kv isolation. One of the converters is for 5V output and the other is 15V output for driving the igbt's. I designed the output section to use both 12v and 15v output modules so that you can use the 15v one for powering the vla500's and you can use the 12v output one to power a modfet driver board.

Here is a link to the 2 dc-dc converts:

15V: EC4A03H Cincon DC/DC Converters & Regulators

5V: EC4A01H Cincon DC/DC Converters & Regulators

-Adam

OK, I think that I understand.

The board can be populated at one spot with either a 12v output DC-DC for mosfet drive, or a 15v module for IGBT driver.
The second spot is for a 5v output module in each case.



For testing, I sometimes use old laptop/TV powerpack/charger modules.
They are mostly quite well isolated, but one has to check to be certain.
Just be careful with cell phone chargers as they more commonly lack full isolation.

Sometimes they are big and ugly, but allow testing and further refining to continue until you can get the proper module.

Yes Mark,

I plan to find a 5v supply and I have a 15v industrial supply kicking around. Now that I think of it, I have a 5 v one as well. I will probably use these for the tests.

-Adam

Well this weekend I was able to start testing the 3 module driver board. I assembled the driver section up to the gate resistors. I started adding gate resistors with just a 3R3 for on and off. This yielded slower then I had thought rise and fall times.

http://farm8.staticflickr.com/7015/6...dc256ed6_z.jpg
Untitled by AdamBrunette, on Flickr

http://farm8.staticflickr.com/7173/6...f13140fd_z.jpg
Untitled by AdamBrunette, on Flickr

Heres a scope shot of the single 3R3 gate resistor. (Switching at 16kHz)
http://farm8.staticflickr.com/7172/6...826bfa86e3.jpg
3R3GateOnOff by AdamBrunette, on Flickr


After this, I decided to speed up the turn off as much as I could without going overboard. The recommended gate resistor value is 3.1R so I went a little smaller with 3.0R. This is put in parallel with the 3.3R with a ultra fast diode blocking it from being used for turn on. This yield's about 1.57R for turn off. This dropped turn off times from 1.14uS to 720nS. This is still considerably slower then the 250nS fall time in the spec sheet. The spec sheet is talking about switch side and not gate side though. This could all change later today when I connect an actual battery and load motor.

Here is a scope shot of the rise. (Switching at 8kHz)
http://farm8.staticflickr.com/7173/6...a0a3f88d3c.jpg
3R3On 3ROff Rise by AdamBrunette, on Flickr

Here is a scope shot of the fall. (Switching at 8kHz)
http://farm8.staticflickr.com/7019/6...69d557c3ea.jpg
3R3On 3ROff Fall by AdamBrunette, on Flickr

On a side note, I have been doing some loss simulation calculations and have found out that in order to get the 1000A out of 3 modules, and keep them from overheating, The controller will have to run at 8khz.

Here are screen shots of 16khz and 8khz. Look at switching loss's for each and junction temp average.

http://farm8.staticflickr.com/7019/6...7dbd3075_z.jpg
Screen shot 2011-12-11 at 11.14.14 PM by AdamBrunette, on Flickr

http://farm8.staticflickr.com/7024/6...9ce69e95_z.jpg
Screen shot 2011-12-11 at 11.14.01 PM by AdamBrunette, on Flickr

-Adam

Nice work ,I am sure you keep away from the little parties scattered here and there at this time, to keep your work in progress.Keep it up.
Thanks.

Don't forget that your gate waveforms will differ greatly when driving the transistors with no load or dc bus than when driving on load. That pesky miller effect really takes its toll!

Well I spent some time this weekend tuning the driver board. I ordered new carbon composition resistors and matched them in groups of 4 to make specific high tolerance values.

I first started with 3.0R because I messed up with my meter and didnt offset the lead resistance of 0.1R. Heres a shot of what That looked like on the scope.

http://farm8.staticflickr.com/7029/6...88ff883b6f.jpg
NewFile7 by AdamBrunette, on Flickr

You can see the turn off was ringing like crazy. It would actually turn the switch fully on with the ring.

I then went to 3.65R and it started looking alot beter. But im curious why there is still such a large ring... I plan to try 4 parallel 16R resistors for about 4R gate resistance and see if that helps. I want to stay as small as possible because switching loss's are already high enough. I dont want to counter my slower 8khz frequency with long switching times.

http://farm8.staticflickr.com/7025/6...f38bb6511c.jpg
NewFile17 by AdamBrunette, on Flickr

The turn on of the module has always been pretty good.

NewFile8 | Flickr - Photo Sharing!

Here are the shots of turn on and off with the 3.0 gate resistor. They look pretty good.

Turn on:
http://farm8.staticflickr.com/7031/6...89a55efa96.jpg
NewFile1 by AdamBrunette, on Flickr

Turn Off:
http://farm8.staticflickr.com/7031/6...fae9917df8.jpg
NewFile2 by AdamBrunette, on Flickr


Heres a longer time withe the gate off for comparasion

http://farm8.staticflickr.com/7014/6...76999e0d3e.jpg
NewFile20 by AdamBrunette, on Flickr

This is where I left it. About 33V spike on turn off. This is down from about 68V!!

http://farm8.staticflickr.com/7028/6...b884905092.jpg
NewFile22 by AdamBrunette, on Flickr

Heres what damien was seeing on his controller. 5V/div 20uS/div
http://farm8.staticflickr.com/7172/6...20a1c054_z.jpg
ce1 by AdamBrunette, on Flickr

More testing tonight!!

-Adam

I have used ferrite beads with Bots to quench the ring - that help?

Adam , what load are you running on the controller?

Where would the ferrite bead be placed? I have taken great care at making the gate and return traces perfectly overlapped. Ate on top and return on the bott of the board.

Damien, the load is a 12v starter motor and te supply is a 13.8v 10A wall supply that's isolated. The ring doesn't change from about 8% to 100% throttle.

-Adam

That IS what the gate resistor is for, to slow the electrons down and force them to leave the gate in an orderly fashion. You can also aid this problem by putting a ferrite bead on the driver leg of the gate. I use it on mosfets but it is true for all drivers. Experiment ....

ALRIGHT!

Last night I was able to finish tuning the driver board. I did have some good scope shots, but my computer crashed when I tried to open them and they got currupt....:mad:

Yesterday I showed you this shot:
http://farm8.staticflickr.com/7009/6...8b009b2a63.jpg
NewFile17 by AdamBrunette, on Flickr

There is about 35V P-P. Well even though I lost my best capture, I have this one thats very close.

http://farm8.staticflickr.com/7021/6...f5bd7408b7.jpg
NewFile23 by AdamBrunette, on Flickr

The scale is a little different, but you can see its getting better. The one I lost was showing about 22 or 24v P-P. I will get a new shot tonight. It also has very little ringing now. This is were I will leave it.

I ended up adding a diode to the turn on resistor and re-installed the turn off one. I am now running a 3.3R wire wound turn on, while not the ideal type, the inductance of the wire really helped without adding any other parts. The turn off is 2 carbon composition resistors 16R in value. In this pair, they create a resistor that is 8.45R. The carbon composition resistors tolerance sucks.... You cannot find 2 that are the same. While when I measure the 3.3R wire wound one, the 3 I picked are identical.

I was also able to measure some delays. From the driver to the gates everything is in the range of delay of 12nS max. And from one gate to the next, they are all in sync better then my scope can measure. Somewhere around < 3nS. Im pretty happy with this.

Here are the shots of rise from driver module to a gate.
http://farm8.staticflickr.com/7157/6...cffc4e61e2.jpg
NewFile25 by AdamBrunette, on Flickr

http://farm8.staticflickr.com/7029/6...a63a2ffcf1.jpg
NewFile26 by AdamBrunette, on Flickr

And delays between modules.

Rise:
http://farm8.staticflickr.com/7014/6...9a6ab7eca8.jpg
NewFile27 by AdamBrunette, on Flickr

Fall:
http://farm8.staticflickr.com/7151/6...63be003917.jpg
NewFile28 by AdamBrunette, on Flickr

I will be doing high power and voltage tests starting tonight.

-Adam

Adam;

Any new updates that you can share?

30 days since your last post.

Curiosity is getting the best of me. giggle-giggle

R J

Hey RJ,

Ive been busy.

I have been testing and modifying the logic board for the controller, as well as working on my corvette and working on the 1000A controller. Its all coming together. I plan to try and test the controller tonight with the Uprising logic board on a hydraulic pump at 24V. This setup is capable of 110A for 10 minutes.

As for my car, The frame is finally done and now, I just have to strip the body and do the body work. My goal is the middle of april to be ready for the EV-ent down in Macungie PA on May 5th.

-Adam

Update?
 

Any update on your progress? Great job!!

Well its been too long since an update, and im sorry.

I never stop working on things and its hard to find time to update my threads.

Well a good amount of work has happened since the last update.

I have updated the driver board and I believe it is ready for production now. Mostly small changes that make it easier to mount and work on the IGBT modules. I have notched the edge of the board that is along the IGBT's so the mounting bolts can be removed now. I have also added diodes to the turn off section of the gate lead. This way, you can have 2 different resistors, one for turn on and one for turn off. I also added holes for 3 parallel gate resistors on the on side and 2 parallel on the off side. My plan is to use carbon composition resistors because they hold their resistance very well, but need to be manually paired to create the value needed. Im using 2 - 16R ones for about 8.3R for turn off and will be trying 3 parallel 10's for 3.33R for turn on. On the previous board, which is running happily in Isaac's snowmachine, is using a 3.1R wire wound resistor. The extra inductance helped a bit with waveform, but I plan to revisit this.

Here are some shots of the New driver board.

http://farm8.staticflickr.com/7202/6...29054866_z.jpg
Untitled by AdamBrunette, on Flickr

http://farm8.staticflickr.com/7053/6...c29e1e8d_z.jpg
Untitled by AdamBrunette, on Flickr

http://farm8.staticflickr.com/7209/6...e5a201e1_z.jpg
Untitled by AdamBrunette, on Flickr

http://farm8.staticflickr.com/7061/6...d9d4a318_z.jpg
Untitled by AdamBrunette, on Flickr

http://farm8.staticflickr.com/7178/6...7f518ae6_z.jpg
Untitled by AdamBrunette, on Flickr

http://farm8.staticflickr.com/7066/6...9503ec8e_z.jpg
Untitled by AdamBrunette, on Flickr


As for the Uprising control board:

Eric and I have been working on the software for the controller. Lots of testing and bringing all the pieces together. We hit a roadblock when the original ATmega644p didn't have enough data memory (4k) and couldn't fit the new web interface information. We were trying to fill it 253% full!! We then upgraded the chip to a Atmega1284P which has 4x as much data space. It was a simple upgrade, just had to use some hot air to remove the old chip and some flux to mount the new chip. That took care of the hardware. Then we recompiled the code for the new processor and tried it out. It didnt work....:( The status LED is now blinking much faster then it was with the original processor, and I cannot figure out why. Im still working at it though. My plan is to go back to the simple test programs that test each hardware section of the board and see if everything is working properly.

Well I think that's it for now. If anybody has ideas of what happened with the control board, I would love to hear them. Maybe I accidentally cooked it with too much hot air? I doubt it though...

-Adam

P.S. Here is the new picture of the control board!

http://farm8.staticflickr.com/7015/6...733a1c7d_z.jpg
Untitled by AdamBrunette, on Flickr

Adam , your worse than AC propulsion:) You tease us with these wonderful products. I want one! ....... actually 2:)

Hi Adam,

Thanks for all your hard work !!!

I don't have an answer for you about that microchip, debugging electronics can drive you crazy! You might have a solder bridge under the chip, you could try to do a quick toaster oven reflow of the whole PCB assembly. Flux around the chip and heatup the PCB, that would melt any small solder bridges. Just a long shot, might work?

A breakout board for the surface mounted micro would have make your life easier too ???

-Mark :(

The worst debugging problem I've had after replacing LQFP chips is some pins not making contact. You can almost always spot misalignment and solder bridges. But if you don't have exactly the same bump of old solder on the pads, some pins on the replacement chip can be held off the surface just enough to not join the solder pool. The pins might make contact only when cold, or only when you touch the chip.

I once spent a long day trying to figure what design or component problem caused my crystal to not oscillate unless I touched the chip. I was assuming that it was the added capacitive load from my fingers, when it was actually the pressure causing the pins to make contact with the board.

Well, Just a little update.

I have come back to this a few times getting frustrated each time...

I have found the problem, but according to datasheets, there is no problem...

The issue, I believe, has something to do with the watchdog shutting down the controller. Ive retested all hardware and everything is working perfectly.

If anybody feels up to it, the old processor was an Atmega644A and the new processor is the Atmega1284P. If anybody can find any differences, I would LOVE to hear about them.

-Adam

Adam:
On the Atmega1284P does the watchdog stop the processor or does an event of lost oscillation stop it? I asked as DJBecker triggered a memory of a problem I had with a physical layout of the driving Xtal - I had to ground strap the metal can of it to stop an internal shutdown glitch due to stray spikes - I know - I know a stab in the dark :) ....

Well It looks like the processor is still doing something because I still get a 16.0001Mhz clock on the CKOUT pin of the chip. If I enable the watchdog to always be on, the LED blinks rapidly and I get a clock out of about 2Mhz. This makes me think that when the watchdog is always on, the chip just cycles and cycles.

The way it is right now, the status LED is solid on.

Im going digging through the code again and going to try and set up some serial prints at points of start up hopefully to find where its failing.

On a side note, I commented out all the code in the enable and disable watchdog functions with no difference in operation..... This lead me to wanting to add the serial prints.

-Adam

Hi Adam,
I know it's a pain debugging such an issue and that you are really busy chasing 1001 other matters, this I ask as I am sure you have many of us on edge ,also how is the vette doing???

I still haven't gotten the code to run on the new processor. Im sending it to a friend that is doing the code and hopefully with the controller in hand, he will be able to get it working...

The vette is coming along, but its slow. It all comes down to money. Currently I have enough to either, 1. finish the car's hardware and body work, or buy 1/3 of the batteries needed. The batteries are enough to make it move at 120V 80Ah, which would be about 20ish mile range, but then I will have to find money to finish the actual car...lol decisions decisions

-Adam

I guess we're on the same boat the money saved always end up spent elsewhere.
Wish you luck and keep up the good work.

Adam,

What's the latest on your controller? Any progress?

Thanks,

Jim

Hello,

I have sent the board to a friend of mine to work on the software. I still have not gotten it to run the cougar firmware with the new 1284p processor, but its a SW issue and not hardware.

When Eric makes some progress, Im sure he will post about it here.

-Adam

This is awesome! I am planning to build a Open Revolt controller for my hybrid hot rod. I have an 11" GE SepEx motor to drive the front wheels and the idea of using the AC setup to control the arm and field from the same controller is more than intriguing. :thumbup:

Will there be, or can there be, two control boards to enable individual and/or manual control of the arm and field separately?