The Corvair Chronicles

[Volta1r3]

Hello everyone,

I'm still pretty new here (bare with me) and working on my first EV build using a 1965 Corvair Monza as a donor:


Originally, I was going to attempt a conversion similar to that of the Electrovair 3 done by High Voltage Hotrods:

https://www.youtube.com/embed/84xgjU1V0Ws

They accomplished this by mating the motor via adapter plate to the bell housing 86'ing the engine as seen in the following:



This would limit me to stock performance at 110hp, about 80kWh on a good day. The Powerglide transmission for this year is pretty solid from what I've gathered in Corvair forums and can take some tactical abuse. BUT even if I were to modify the governor, trick the vacuum modulator, and pray the hollow input shaft holds, I still would not be able to find out what a motor can handle.

Motor redlining is something that of course is going to vary from motor to motor, but it's something I have been wanting to experience for myself. Back in 2017 I built a 3-phase controller from scratch (I wish I would’ve found this site back then…) that successfully spun a motor for a capstone project, but I still haven't gotten around to refining it: adding features like creep, current/voltage monitoring, etc.. I know my gate driver stage is almost solid (needs minor tuning), and I still need MAJOR development to my logic stage (currently using STM32F0 MCU series, may change platform):






These small boards were designed to be directly mounted on most (check datasheet gate and emitter pin distances) IGBT power modules as seen below:



I have gerber files, if anyone wanted to play around with this stage.. If one were so inclined, they could pop one of these boards on a single IGBT for high-side switching a DC motor (much easier than a six step inverter) or even a few in parallel to control a BLDC motor. I certainly enjoyed this project so far:



[ur]https://drive.google.com/file/d/1g3KCIA ... Qyn0M/view[/url]

From the video you can see the motor spinning, but not quite right. I had my ramp rate algorithm increasing exponentially (instead of linearly) and the motor couldn't keep up. The rotor would go as fast as it could until the voltage slipped away. Hence, the torque bump to the motor when toggling the logic power while running.

Here's some wave forms I had pulled if anyone was interested:

https://drive.google.com/drive/folders/ ... sp=sharing

I hope to get back to this as I have some new circuits and topologies I want to implement; I may even try to go all analog for my logic using 555s and ring counters.

Right, back to the 'vair. With what I've learned regarding the powertrain, I began looking around and around at different motors and others' setups.. DC or AC... AC or DC... while that was going down, I decided to get the 'vair up and running with the ICE so I could get a feel for how it was meant to drive.

I couldn't get her started initially, and looking at fuel off the fuel pump I noticed something foul afoot:



Bad gas... probably older than me! So, I changed the fuel filter, drained the old gas, and inspected the carbs... really, a bread tie holding a throttle rod on?

The carbs were not in the best shape as can be seen:



Thanks to Wolf Carburetors, I was able to get some rebuilt Rochesters for a great price:



With the help of some high-octane fuel, I was able to clear the old crud out and get her running every time (noice!):

https://drive.google.com/file/d/17z-ERb ... jkcbv/view

Not bad, except for this annoying clicking sound in the engine, a friend suggested it may be a sticking lifter, but that's where I draw the line. This is more of an electrical conversion than restoration project after all! After some cruises around the hood, really enjoyable cruises I may add, I caught a break and got my hands on a Model S performance LDU:



Whoops! I dropped my screwdriver, and this happened:



I may as well take a peek:



Cool! But then I noticed some weird bubbles over one of the IGBTs..



It looks like a hot spot forming on this section of the inverter... I read that in the performance models the IGBTS are rated for 120A. From what I could gather on the physical controller and pictures of dissected controllers, the controller has 7 IGBTS per high/low side cluster for a total of 28 transistors per phase leg, 14 per high side total and 14 per low side total. That would mean 1680A? But I read that the max current is 1500A in ludicrous mode? Maybe there is more a bit more power that can be squeezed through?

EDIT: I’ve read that you can get up to 1700A! I wonder how much more it can handle with even more cooling.

While searching for answers I stumbled across a Damien Maguire video that pointed me to Openinverter.org (I’m a little late to the game). Lo and behold, they offer a logic board replacement on the LDU that will help me tap out this motor a lot faster than I could on my own and for a GREAT price. While I was here, I took a peek at their DIY inverter docs and was excited to see how similar my own inverter design was. I guess I really am learning in the right direction.

Next step is to remove all the ICE systems and check the integrity of the frame while my bench test parts come in.

Note: If I misinformed anything in my ramblings, please call me out as I am novice at best and still striving to better understand this stuff!

[SRFirefox]

That's a good lookin' 60 year old car! I am so excited for this project; I've loved the Corvair since seeing a V8 mid-engine conversion at the local car show. Do you have any suspension upgrades planned for after the drive swap? Also, what batteries are you thinking about?

[Volta1r3]

Thanks Fire! The frame seems to be straight by the way the doors, trunk, and "frunk" close, but I won't know until I remove the ICE systems and get further into it. The driver rear is lower than the rest of the car so that shock must be bad, but I'm going to upgrade the set anyway, possibly doubling up with new coils as well. I haven't done too much homework on upgrades because it's hard to tell what I'll need until I've got the final weight of the battery pack squared away. Let it drag, am I right? jk

I've been on the hunt to find a couple 2016-17 Nissan leaf packs, but I haven't had any luck.

[P.S.Mangelsdorf]

The second generation Corvair was on the list of cars I was looking for before picking up my '40 Chevy. Can't wait to see what you do with it!

[Volta1r3]

Thanks Mangelsdorf, I'm looking forward to the arrival of my logic board so I can get to playing!

How are you approaching your instrument cluster?

[P.S.Mangelsdorf]

Thanks Mangelsdorf, I'm looking forward to the arrival of my logic board so I can get to playing!

How are you approaching your instrument cluster?

All new stuff, the car is gutted and going to be race focused, so I'm not worried about updating the old gauges

Current plan is GPS speedo, SimpBMS screens for SOC, and some idiot lights from BMS and inverter.

In the future I'll add two CAN driven temp gauges (battery and motor) and possibly a CAN driven tach used to show amps instead.

[Volta1r3]

Gotcha, I wanted to replace the old ones with some custom gauges, but the shop I was talking to doesn't want to play ball. I may just find a display, run it with raspberry pi (or something) and hide it where the cluster lives to keep as "original" looking as possible. Maybe try to show RPM, current, voltage, MPH, batt%, temp (controller, motor, batt pack).

Your ride is going to be a monster! Can you imagine running 'shine in that thing?!

What's your best time on a 1/8th (if I may ask)?

[P.S.Mangelsdorf]

What's your best time on a 1/8th (if I may ask)?

Well the car hasn't run yet, so not a clue.

Based on Jon Volk's BMW and a few others out there, low 11's in the 1/4 is clearly attainable. Goal is mid to low 10's, but it'll have a cage for 9's.

[P.S.Mangelsdorf]

Gotcha, I wanted to replace the old ones with some custom gauges, but the shop I was talking to doesn't want to play ball. I may just find a display, run it with raspberry pi (or something) and hide it where the cluster lives to keep as "original" looking as possible. Maybe try to show RPM, current, voltage, MPH, batt%, temp (controller, motor, batt pack).

I've been looking at some of the CAN options from Speedhut and Autometer, not sure which way I'm going yet.

[Volta1r3]

I've been looking at some of the CAN options from Speedhut and Autometer, not sure which way I'm going yet.

Thanks for the leads. I saw this guy print out his own gauge faces, I may take another DIY project on and just custom make my own using what I've got.

CAN is new territory, but I see the advantage of eliminating so many wires.

Have you dealt with CAN in before?

[P.S.Mangelsdorf]

Have you dealt with CAN before?

No, that is one of the items I am still nervous about working with. I need to get my car up and running because I do best learning as I go.

[Volta1r3]

I do best learning as I go.

I'm with you on that.

[muehlpower]

Gotcha, I wanted to replace the old ones with some custom gauges, but the shop I was talking to doesn't want to play ball. I may just find a display, run it with raspberry pi (or something) and hide it where the cluster lives to keep as "original" looking as possible. Maybe try to show RPM, current, voltage, MPH, batt%, temp (controller, motor, batt pack).

Your ride is going to be a monster! Can you imagine running 'shine in that thing?!

What's your best time on a 1/8th (if I may ask)?

Would that be something to keep the original speedometer?
https://www.speedhut.com/Speedbox.html

[Volta1r3]

Would that be something to keep the original speedometer?
https://www.speedhut.com/Speedbox.html

That's pretty cool, but expensive. I keep leaning back to mounting a small tablet, and displaying through the existing holes... Either way I've got some time to think about it before I get to that stage.

[Volta1r3]

I got some toys in the mail today, and I thought would be worth testing to better understand the operability (also so I know the components are in working condition) :

Contactors

I picked up two sets of GV200 contactors (GV200-004-D, couldn't explicitly locate this part code but it's still the 200 series, so I've got a ballpark to play in):

https://www.gigavac.com/sites/default/f ... /gv200.pdf

They are much more compact than I expected, and class I div II rated, cool. These contactors include built in fly-back diodes (nice not having to add) and are rated up to 800VDC @ 500+A continuous. The spec sheet suggest current carry beyond 2000A range for a short period (85DEGC) at the loss of cycle life (looks like I will have to do further testing later on ).

The coils were checked to make sure it wasn't broken with a simple continuity test, all passed, yay. The resistance across each coil and contact were measured, each reading about 5.3ohm across the coil and open across the contact, double yay. A power supply was then connected across the coil, and a meter, set to resistance, was connected across the contactor. A voltage was slowly walked up while an eye was kept on the current until a contactor clunk could be heard indicating a "close state", values were then recorded. After reaching close state, the voltage was walked back down until the contact release could be heard, values were recorded. Each time, the contactor would close a small resistance could be seen:



This was done for each contactor yielding the following data:



With all the contacts proved and the power supply still setup, I might as well test the accelerator pedal. I found a tesla pedal, but tbh I'm a little disappointed in the material used, I feel like I'm going to break it in excitement while flooring it.. I'll most likely end up finding a more durable replacement to accommodate the lead in my foot.

Accelerator pedal

In order to test the accelerator pedal assembly, Vref1 and Vref2 were twisted together and connected to a 5V power supply. The same was done with each TPS ground. A scope probe was connected to read each TPS output signal. The values were recorded at rest state and full throttle position for both TPS. This was done three times for each state on both sensors to provide a voltage delta average. This data can be seen in the following table:



From the voltage delta average, it can be seen that TPS signal 2 is half that of TPS signal 1, a typical offset for a dual TPS pedal. So, awesome, the pedal works .

Since the removal of the ICE in the 'vair is on hold, I'll be setting up my test bench to start spinning the motor next.

[Volta1r3]

I decided to go with a model 3 pyro disconnect to get precision overcurrent protection. I received it in the mail, but during the initial inspection I noticed an orange light showing through the casing:



I decided to investigate.. I started by checking the resistance across the terminals and saw that there was a small resistance. So, that would mean it is good.. but wait! Upon further investigation, I noticed that there were two small fuses in parallel across the positive and negative terminals. I removed the housing/fuses, and disconnected the pyro safety switch (PSS-1) from the circuitry:



I proceeded by taking another resistance measurement across the terminals to find that it showed an open load:



Well, that's not good.. I started snooping around the housing and found a black mark under where the PSS sat.. hmm, looks kind of like a burn mark:



Well now I'm intrigued.. I broke the blue cartridge out before realizing that you can just pop it out with a small flat head screw driver , and saw more indications of burn marks:



Might as well go deeper.. after removing the terminals, I found that they were not connected due to one being bent at a (close to) 90DEG angle:



That would explain the open reading when checking resistance, now I'm invested! I noticed a small cylinder with a tab toward the middle:



*Chin scratch* that looks like it comes out, and guess what? It did! Look at this little guy hiding in here:





The squib appears to have been activated (sweet!). It looks like upon discharge the charge pushes (more like blasts) out the cylinder, bending the terminal tab towards the notch seen in the blue discharge chamber (if that's what you call it?). Good thing I was able to find a couple of replacements for under $20 online from...Walmart?! lol

[muehlpower]

i have noticed that the pyrofuse is triggered when the HV control unit is unplugged. Mine was working when I got my battery pack. After removing it, it was broken. I suppose you have to unplug the fuse first!

[Volta1r3]

i have noticed that the pyrofuse is triggered when the HV control unit is unplugged. Mine was working when I got my battery pack. After removing it, it was broken. I suppose you have to unplug the fuse first!

I'll have to try it out, thanks for the heads up! What pack are you using?

[Volta1r3]

It's been awhile since my last post, and I'm finally settling into the new place (please excuse the disarray, not all the way settled in) . Since then, I've added some fun new parts to my build:

(2) parallel Gen 2 Chevy Volt packs I plan to run in parallel
(2) OEM 3.3kW Chargers
OEM DC/DC
C5 Z06 Corvette front and rear sub frames
(2) Orion 2 BMS w/ (1) wifi module (if they ever get built...)
Thunderstruck EVCC

I've been playing around with this design to get the two packs from battling too much... I don't know if I'm happy with loss from the blocking diodes, or resistors:



The IRS, motor, etc. from the 'vair dropped and I've got a better idea of how I'm going to mate the rear 'vette sub frame:




Being a unibody design, I think I'm going to stiffen up the body by reinforcing the rocker rails underneath and maybe add a small 4 point roll cage to avoid turning my grocery getter into a taco truck (not the good kind).

Both battery packs appear to be good at first glance, voltage is a little high for storage, but I plan on performing a capacity test soon anyway:


I'm thinking of a new name for Joule... "the Voltair S"

[celeron55]

What on earth is happening in that electrical diagram?! You pretty much just have to wire the batteries in parallel and hope (or measure) they match well enough - what you've drawn definitely does not make any sense and definitely won't do whatever you thought it might.

[Volta1r3]

Those are contactors, not caps, and the ( ) are the contact coils. This is how I'm used to seeing contactors depicted in controls diagrams.

Is this what you're talking about?

I'm going to attempt to fully charge both packs to match at the mV level before connecting for the first time.

[m.art.y]

I'm going to attempt to fully charge both packs to match at the mV level before connecting for the first time.

I also have two packs but I decided to make it one pack by rewiring cells into 2P configuration. Even ORION themselves advise against parallel strings - packs wired in parallel. If it's possible to rewire your cells you would save yourself a lot of trouble and would only need one BMS instead of 2.

[celeron55]

Even if they're contactors which makes a lot more sense it's risky to gate regen current using a contactor, at least when using openinverter which is too slow to react to overvoltage to save itself. Also diodes passing hundreds of amps generate a *lot* of heat that you have to somehow get rid of, presumably from the inside of a plastic enclosure which will be challenging.

Also, it's enough to have the packs with each cell something like 50mV from each other. The current caused by such difference is similar to slow-ish charging which is definitely safe, unless you have something very wrong with a cell.

[Volta1r3]

I'm going to attempt to fully charge both packs to match at the mV level before connecting for the first time.

I also have two packs but I decided to make it one pack by rewiring cells into 2P configuration. Even ORION themselves advise against parallel strings - packs wired in parallel. If it's possible to rewire your cells you would save yourself a lot of trouble and would only need one BMS instead of 2.

I thought these packs are from the chevy factory as 2p modules? So that would make 96s4p with two packs, right?

I've considered reconfiguring the packs into a 96s4p and saving some cash on a BMS, but you don't think it would be easier to just lose some power (and cycle life) than reconfiguring these packs at the cell level? After reading the Orion documentation on parallel strings:

https://www.orionbms.com/manuals/pdf/pa ... trings.pdf

I'm leaning toward sacrificing some power, and keep making adjustments to prevent a cascading failure. But with that said, I don't know what's going to happen when I request a huge power demand. One pack is probably going to be favored over the other, but I don't know how these packs will respond until I begin testing them under load.. I tried to get packs around the same mileage, but the owners more than likely drove differently and I couldn't account for that. I still want to see if I can mitigate the issues that come with having strings in parallel on a traction pack.

I'm still making adjustments and continuing to dig on the subject, but I still haven't ruled out the 4p setup..

How to you plan to reconfigure your packs to be in parallel at the cell level?

Even if they're contactors which makes a lot more sense it's risky to gate regen current using a contactor, at least when using openinverter which is too slow to react to overvoltage to save itself. Also diodes passing hundreds of amps generate a *lot* of heat that you have to somehow get rid of, presumably from the inside of a plastic enclosure which will be challenging.

Also, it's enough to have the packs with each cell something like 50mV from each other. The current caused by such difference is similar to slow-ish charging which is definitely safe, unless you have something very wrong with a cell.

I've been kind of worried about when the motor goes into regen... Maybe I'll set up a dummy load in parallel to warm up some coffee when I'm not putting the power back in the battery?

If I did take this two string approach, I was thinking of using something like DO-9 diodes, and setting up a nice heat sink for them to cool off with donor car's original engine compartment blower (Corvairs are so cool, pun intended). Or possibly putting the heat sink on it's own cooling loop as a cold plate.

[m.art.y]

How to you plan to reconfigure your packs to be in parallel at the cell level?

Yes, if one pack is already 2P you'd have to go for 4P. I'm not familiar with volt battery and how are cells connected to each other but on my battery cells are connected with screws and busbars so I will be rearranging the busbars to make each module 2P and only half the voltage they originally had.