[mfor1000]

I'm starting my first EV project...so be gentle.

Here is what I have so far:

1987 Porsche 944
Tesla Large Drive Unit (Thanks @WK057! and HSRmotors.com)
2x Chevy Volt Battery (second on the way)
Tesla 2nd Gen Charger

Fitting the drive unit to this car should be pretty straight-forward. It has a semi-trailing arm rear suspension and had room for a transaxle. Once I remove the transaxle, fuel tank, and exhaust, it looks like there is plenty of room. (see pics) For half shafts, I plan to weld a flange to the inner cups from the Tesla shafts to allow the use of the stock 944 CV's.(similar to @jackbauer's Panzer project)

The plan for the battery is to widen the transmission tunnel on the car and sacrifice the back seat for the first battery. Then build a box for the engine bay for the second.

I will definitely use a BMS system for the batteries. I'm hoping @bigmouse can help me with that. It looks like he has that figured out for the Volt BMS.

I'm still debating on controlling the charger. It looks like my choices are the custom board by @jackbauer and the upcoming controller from @WK057.

I haven't bought a DC converter yet...looking for suggestions there.

Please feel free to tell me when I'm screwing something up. I'm going to try to keep this up to benefit from everyone's advice!

 

[Duncan]

Are you going to disassemble the Volt batteries and use the modules?
Trying to use the complete T shaped units ....

When you take the modules off the bottom plate you can then drill out the spot welds and use the mounting hardware without the heavy bottom plate

 

[mfor1000]

Unfortunately, the suspension lateral tube prevents keeping the T completely intact. I'll have to split it at the T.

The longitudinal tunnel portion of the pack will go into the widened trans tunnel, while the lateral T section will have to reside above that suspension tube in a box created where the rear seats currently are.

The plan is to split it and then use longer hoses and wires to connect the two parts of the pack. I also am going to attempt to retain the fiberglass cover by splitting it as well and then glassing in sections to seal the holes.

I'm currently try to read the voltages from the BMS in CAN...struggling a bit. A friend did see a post on this:

http://www.diyelectriccar.com/forums/showthread.php/attempting-hack-chevy-volt-drivetrain-107946p6.html

However, the voltages from that are not making sense.

 

[Duncan]

Hi
Unless the T dimensions fit with your space (lucky) I would advice using the modules to alter the dimensions to fit better

Taking the battery apart into it's modules is easy! - and they can then be configured the best way for your car

 

[itchyback]

this is going to be a sweet conversion!
looking forward to updates

 

[mfor1000]

I think it will work out ok to break it into just two pieces. Then they well be nice and protected inside the fiberglass housings. The question is whether or not I can widen the tunnel enough to fit over the whole thing and still have enough room for the seats. I believe it will work with the stock seats. I have some racing seats in there right now, which would be too wide.

I'll definitely have to break up the second pack to go in the engine bay.

I finally figured out what I was doing wrong with reading the BMS. It looks like it is big-endian...(backwards). I attached my Arduino code to do this. It is using a CanDue with Collin's libraries.

Now I get decent numbers. Attached is a screen shot from this code. You'll notice that a bank of cells reads zero and the min is messed up because I found one cell that was a bit low. I unplugged that bank and confirmed it with the volt meter. It was 3.85 volts compared to 3.90-3.91 for all the rest of them. The screen shot was taken with that plug out.

I'm charging that cell now by itself with my trusty drone battery charger set on 1S...a little scary...but slow charge at 0.7 A.

 

[electro wrks]

Unfortunately, the suspension lateral tube prevents keeping the T completely intact. I'll have to split it at the T.

The longitudinal tunnel portion of the pack will go into the widened trans tunnel, while the lateral T section will have to reside above that suspension tube in a box created where the rear seats currently are.

The plan is to split it and then use longer hoses and wires to connect the two parts of the pack. I also am going to attempt to retain the fiberglass cover by splitting it as well and then glassing in sections to seal the holes.

I'm currently try to read the voltages from the BMS in CAN...struggling a bit. A friend did see a post on this:

http://www.diyelectriccar.com/forums/showthread.php/attempting-hack-chevy-volt-drivetrain-107946p6.html

However, the voltages from that are not making sense.

Could you show a picture of the interfering lateral tube?

 

[mfor1000]

Basically, the T part of the pack will have to sit above this tube and high enough that the suspension doesn't hit it. That is the torque tube for the torsion bars. The rear seats are right above that.

 

[electro wrks]

All right, I see a version of it in this photo: http://bestautophoto.com/imgs/4306-tuning-porsche-944-25-turbo-03.jpg.html#gal_post_4306_tuning-porsche-944-25-turbo-03.jpg

And this one: https://newhillgarage.files.wordpress.com/2013/06/dscn0117.jpg

Looks like it has torsion springs in the tube? And coil overs? The coil overs I saw must be auxiliary springs. Thank you.

 

[mfor1000]

Mine doesn't have coil overs on the rear. It just has the torsion bars and shocks.

Some racing guys take the torsion bars out and substitute coil overs for adjust-ability I think. I may end up doing that, but for now I'll just try to use the bars.

 

[mfor1000]

Excuse me if a similar post shows up twice...I sent one earlier like this but it didn't show up. I wasn't sure if I messed something up or if it needed approval.

I figured out my issue with reading the CAN from the Chevy Volt BMS. It turns out to be big-endian, instead of little-endian... That confused me for a while, but I finally figured out the bit shifting and masking to get it right.

Attached is a screenshot of the output. You'll notice one group is zero and my min is messed up. I found one cell that was at 3.85 V instead of 3.90 V like the rest of them. I unplugged that group and was charging that cell while taking this screenshot. The trusty drone battery charger comes in handy...0.7 A really slow charge!

I also included my arduino code. I'm using the CanDue with Collin's libraries.

 

[mfor1000]

Progress! Motor is turning...

Using the controller from HSR Motors, it is turning in my shop! Jason did test it in his shop first, but this is with my little mockup dashboard.

https://www.youtube.com/watch?v=zhv7QqqKDzA

I'm still working on the software, but have been able to interface with both the HSR controller and the BMS using the two can channels on the CanDue. I'll go ahead and post the latest code as well.

The HSR controller can use 12V in from a momentary switch for each drive mode. So I did that and I then use the CanDue to light up the appropriate LED on the switch once the drive unit has accepted the gear request.

 

[chrishazell]

Regarding your half shafts I would not weld cups on there hardened steel and don’t weld well. With the torque the large drive unit can output I think they will shear off. I am currently working with a company that are going to scan the tesla drive shaft splines and will then be able to make up shafts that use the tesla inner joint and will be able to put whatever spline on the other end to fit to any outer cv joint. I will be using them on Kevin split screen conversion and my 350z conversions should have something end of January so will update you once I get them sorted.


Sent from my iPhone using Tapatalk

 

[mfor1000]

Chris,

That does sound interesting. Definitely keep me posted.

 

[mfor1000]

More Can Decoded...

I decoded all but just a couple utility channels on the HSR controller and worked on the Volt BMS some more.

For the Volt BMS I've been using the following for reference:

http://www.diyelectriccar.com/forums/showthread.php/attempting-hack-chevy-volt-drivetrain-107946p6.html

The 0x210 didn't seem to match completely.

The voltage in the first 12 bits worked ok (Big Endian), but the current didn't match up well. I played around and found that bits 3 and 4 combined to a signed 16 bit that when divided by -50 gave similar amperage to what I saw on a clamp meter.

Here is the code I used:

case 0x210: // Pack Voltage, Main Pack Amps
BMS1_packVolts = (((incoming.data.bytes[0] << 8) | incoming.data.bytes[1]) >> 4) / 8.0; // Volts
BMS1_packAmps = ((int16_t)((incoming.data.bytes[3] << 8) | incoming.data.bytes[4])) / -50.0; // Amps

0x264 and 0x304 didn't seem to do anything, but they may need the charger.

Attached is the full code I'm using for both the HSR controller and the Chevy Volt BMS.

 

[Boxster-warp]

Hello
Why you don't use the tesla modules?
Is there not enough place in the 944?
Greetings Boxster-warp

 

[mfor1000]

Tesla modules are very expensive, and you need 16 of them to support this motor since they have such a low 'C' rating. I don't have to have that much range. I also don't think it is a great idea to spend a ton on batteries right now. Electric car production is increasing rapidly and the prices of the salvage battery packs are likely to drop quickly in the future.

My thought is to get the car running with a couple Volt packs and then see what happens. If I can get 5 years out of these packs, then I'll consider upgrading to whatever is the best deal at that time.

 

[Kevin Sharpe]

Tesla modules are very expensive, and you need 16 of them to support this motor since they have such a low 'C' rating.

The Tesla open source controller allows you to use a lower system voltage. iirc Damien is using 11 modules (250V Nominal) in the BMW 8 Series conversion and I'm using 14 modules (320V Nominal) in the VW Bus.

 

[Duncan]

Re - "C" ratings and voltages

I believe an AC system and a DC system are similar in that the motor/controller are - Power In = Power out devices

So you will start off with a high motor current but a LOW battery current and only as speed/revs rise will the battery current rise

This means that you don't need as high a "C" rating as you might think

On my car I have a 1200 amp setting on the controller - worse case (for me) is the drag strip

My battery current will start out down below 100 amps - it will rise as I accelerate to about 1200 amps (100% controller) and then drop as I continue to accelerate and my motors increasing Back EMF reduces the current - I think it's down at about 600 Amps at the end of the 1/8th when I lift off

This means that I have not (yet) blown the standard Chevy Volt fuse! and my batteries don't seem too unhappy

Theoretically I'm pulling about 20C - but that is only for a second or two

 

[rwaudio]

Very nice, that's exactly what I'd like to transplant into my 944, the Siemens/DMOC combo is reliable, but lacks power. You may find useful info in my blog related to either the 944 or volt packs. http://www.electricporsche.ca

As for the volt packs, I won't work so hard to keep the whole T with fibreglass shells, just a huge waste of space. I did two volt packs in a Porsche 914 cramming them in every space I could find, while keeping the liquid heating/cooling.

And just an FYI, the front of the volt module has built in precharge resistors, you could use those instead of the giant ebay one. There are also a couple of contactors and a heater you may choose to use along with some small high voltage relays good for heaters or dc/dc converters, plus some fuses etc.

Good luck!