DIY Electric Car Forums banner

1 - 20 of 63 Posts

·
Registered
Joined
·
971 Posts
Discussion Starter #1 (Edited)
VW Polo 86C (1991) AC industrial conversion

Now that I'm getting close to finishing the conversion that started out with a Renault R4, I shall start a build thread.

First, the specs:
- 152 LFP040AH in series for an idle voltage of ~500V
- 18.5kW Industrial ACIM (Lenze MFEMAXX132-22C1C)
- Homebrew AC controller (using 1200V, 400A IGBTs)
- Homebrew BMS
- Inverter Charging
- LiFePo4 System battery charged by an ebay AC/DC (no DC/DC converter)
- Audi vacuum pump (8E0927317) for braking with Audi pressure sensor (036906051G) and homebrew control electronics (some resistors, an NPN transistor and a relay)
- Heating element from a 230Vac heater with a custom solid state relay
- 800V, 16mm² "solar cable" for the DC connections
- 4x6mm² for the motor connection
- Lots of wood, aluminum and fibreglass

I have kept as many original components as possible. For example the original throttle had a pot already, so I could use that. I added another rear defroster switch for cruise control and a pot that controls max regen torque.
I use the existing temperature and fuel gauge as well as the warning lights. Reverse gear is selected with the transmission.

Drive video: http://youtu.be/70uQymh2TGQ

Nov 2014: The car was approved by TÜV Nord to be road worthy

Okt 2015: The car has now done 9000km with 120km commutes. Had a few issues with the charger and a screw from the transmission came loose. No other issues.


June 2017: replaced EMW charger with "software charger" i.e. using inverter and motor as a boost mode charger










 

·
Registered
Joined
·
682 Posts
Forgot to ask on the other topic, what voltage is your motor? 208V Delta?
I would add some temperature probe to those batteries.
Being that they are all in series it might make driving in winter a bit tricky as the rear ones may be hotter and offer better soc that the front ones.
Regeneration below freezing is a big no as well.

Looking Nice, good work
 

·
Registered
Joined
·
971 Posts
Discussion Starter #4
Forgot to ask on the other topic, what voltage is your motor? 208V Delta?
I would add some temperature probe to those batteries.
Being that they are all in series it might make driving in winter a bit tricky as the rear ones may be hotter and offer better soc that the front ones.
Regeneration below freezing is a big no as well.

Looking Nice, good work
Yes 208V delta but its not connected in delta right now. So [email protected] nominal.
The BMS has a temperature probe. Not for each cell, just for ambient temperature.
Regarding the freezing issue: in my other car I witnessed that the batteries become a bit "soft" as in high resistance when its freezing. So the sag or rise of voltage happens much faster than in warmer temps. Thus, the usual voltage monitoring should be sufficient for protection.
As you've driven along for a while the issue heals itself as heat is produced inside the cells.

That is a lot of cells.
Excellent job. Very inspiring.

Jeff
Thanks :)
 

·
Registered
Joined
·
682 Posts
208V delta but its not connected in delta right now. So [email protected] nominal.
What have you done to change that, rewind? Weakening the field above 90Hz or is it in star?
For most motors 208V refers to 60Hz (190 @ 50Hz), so 415V @ 120Hz for constant torque (double horsepower).

The BMS has a temperature probe. Not for each cell, just for ambient temperature [...]
So the sag or rise of voltage happens much faster than in warmer temps. Thus, the usual voltage monitoring should be sufficient for protection.

As you've driven along for a while the issue heals itself as heat is produced inside the cells.
I was worried about regeneration. No charging should be done below 0C, so the voltage should be regulated to 3.3VPC by the controller when this is happens (Average cell voltage) so that:

a) There is not excessive regeneration (Forcing you to use the brakes)

b) Any excess will be consumed by the loads (DC-DC, Heater) but only a small amount will go into the batteries

c) Your controller will dump the excess into a resistive bank (useful in winter anyway)
 

·
Registered
Joined
·
971 Posts
Discussion Starter #6
What have you done to change that, rewind? Weakening the field above 90Hz or is it in star?
For most motors 208V refers to 60Hz (190 @ 50Hz), so 415V @ 120Hz for constant torque (double horsepower).
Ok, thats a conincidence. It is connected in star.
Lenze has a special series of motors for inverter operation only. So all rated values are at 120 instead of 50 or 60 Hz.


I was worried about regeneration. No charging should be done below 0C, so the voltage should be regulated to 3.3VPC by the controller (Average cell voltage) so that:

a) There is not excess regeneration (Forcing you to use the brakes)

b) Any excess will be consumed by the loads (DC-DC, Heater) but only a small amount will go into the batteries
Frankly, turning off regen below some temp limit wouldn't be a big deal.
Speaking of charging, the batteries will have to bear with that. I will charge them at -10C. That may wear them out sooner, it doesn't instantly break them.

Taking care of it or not largely depends on the quantities we talk about.
If I can increase the cycle life from 8 to 10 years by implementing cold weather cases I'd consider it worthwhile. If I increase it from 19 to 20 years or 9.5 to 10 years I'd consider it a waste of time and convenience.
 

·
Registered
Joined
·
682 Posts
Ok, thats a conincidence. It is connected in star.
Lenze has a special series of motors for inverter operation only. So all rated values are at 120 instead of 50 or 60 Hz.
Now I got it. But don't they work in constant horsepower above 50/60Hz? They might just be rated for operation up to 120Hz. Some even for 200.

This might be useful:
http://forums.aeva.asn.au/forums/changing-an-induction-motor-voltage_topic1237.html

Frankly, turning off regen below some temp limit wouldn't be a big deal.
You wont turn off, only reduce the theresold voltage.
This way the motor will only brake enough to power your loads, BUT it will still do regeneration needed for traction control specially useful for a car without ABS on winter. You should also increase the delay time so that acceleration/deceleration are done a bit slower to avoid locked wheels

Ideally your controller should enable an external load when it reaches the theresold, so you still have regeneration, but you aren't killing the batteries with 2-3C Charge bursts

Speaking of charging, the batteries will have to bear with that. I will charge them at -10C. That may wear them out sooner, it doesn't instantly break them.
You can charge them but at a very low C rate. Capacity is reduced by 50% (Check out the new leaf).

Taking care of it or not largely depends on the quantities we talk about.
If I can increase the cycle life from 8 to 10 years by implementing cold weather cases I'd consider it worthwhile. If I increase it from 19 to 20 years or 9.5 to 10 years I'd consider it a waste of time and convenience.
Its not always about how long they will last, sometimes its how much range you need. These cells usually fail with increased internal resistace, so altought the capacity may be similar they wont acept a fast charge or discharge so well. Bit like the old lead batteries.

Almost forgot, I tested a few LIFEPo4 and Li-Ions last year on cold weather. The Li-Ions, even at 0.2C had a permanent loss of capacity below 5C. The LiFEPO4 were a bit more forgiven but the capacity drifted and not equally between all the cells. Some most lost 5% or less, but one or two lost almost 20%
 

·
Registered
Joined
·
971 Posts
Discussion Starter #8
Now I got it. But don't they work in constant horsepower above 50/60Hz? They might just be rated for operation up to 120Hz. Some even for 200.
No no, this motor has never been designed to run off the wall. It can only be operated off an inverter. So everything thats usually rated at 50Hz is now rated at 120Hz. Judging by the weight (66kg), this motor would be rated 7.5kW at 50Hz. But at 120Hz its 18.5kW.


Almost forgot, I tested a few LIFEPo4 and Li-Ions last year on cold weather. The Li-Ions, even at 0.2C had a permanent loss of capacity below 5C. The LiFEPO4 were a bit more forgiven but the capacity drifted and not equally between all the cells. Some most lost 5% or less, but one or two lost almost 20%
Ok, that actually sucks. What exactly was your test?
 

·
Registered
Joined
·
682 Posts
No no, this motor has never been designed to run off the wall. It can only be operated off an inverter. So everything thats usually rated at 50Hz is now rated at 120Hz. Judging by the weight (66kg), this motor would be rated 7.5kW at 50Hz. But at 120Hz its 18.5kW.

Mine does 15KW at 100Hz / 380V with 55Kg, so it sounds about right what you are saying.

Ill have a look at those motors, you can probably have some serious power with delta connection if your drive can withstand 200Ams!

Ok, that actually sucks. What exactly was your test?
A few discharge tests with the batteries frozen.
Ive run some equipment with a data logger out in the street when everything was covered in snow and would fast charge them (2C) at those temperatures (-8, -10, -20 from the freezer).

Capacity was mainly constant, internal resistance increased (So less current could be pulled, limiting the C-Rate at high discharges).

The Li-Ion did perform in overall worse that LiFePo4.
I don't know how they will react with short peaks from regeneration or if they are at different temperatures. My batch was 12Cells.

One Thing: Its an old myth from lead that these batteries warm up when driving.
The average current is only enough to warm up the middle cells during a drive IF they are inside a box.

My 10Ah pack would get 10C above ambient, inside a box, in the boot with the DC-DC and the inverter heating the surrounding air. With brief 5-8C Discharges during aceleration. I imagine that yours, at 40Ah would get even less.

My Test pack would get 5C higher than ambient with a constant 2C discharge. (25 minutes)
 

·
Registered
Joined
·
971 Posts
Discussion Starter #10
Quicks heads up.

I've put all the HV wires into orange tubes as thats a requirement for getting the car registered. Apart from that it looks cool and adds some protection of course. Will take some pictures sometime soon.
I installed the BMS on the front pack. I left the laptop with a small php script in the car and could watch the charging process in my warm appartment. What an improvement :)
Unfortunatly I also found, that the IR communication almost breaks down as soon as the inverter is running. Didn't happen with the trunk pack so it must be caused by the proximity to the inverter. On the plus side, if the communication catches some values, they look plausible.

I also made a decision on the DC/DC issue. There are devices available for 500V+ for around 300€ (http://www.feas.de/netzteil-snt9412-p-588.html). But since I'm a cheapshit I will use my ebay AC/DC converter (14V, 20A) to charge the 12V battery while the main pack is charged. Should I run into trouble I can always upgrade to something better. Don't comment on that, I'll try it anyway :D

Apart from that I tackled the legal issues. Or at least tried to. Here is the best case: http://zeropolomobil.blogspot.de/
He ended up at an inspector that spent 3h looking at the car, trying it out etc. and in the end approved. No EMC/EMI certificate was needed so the whole procedure cost him 300€.

My own research was frustrating as hell. "Experts" told me about "be prepared to spend 10000€ on expert opinions", "the laws have changed significantly 2 years ago" bla bla. I don't know why they do it, maybe they have orders to keep DIY off the road.

Anyway, the zeropolo story gave me hope again that its not so difficult after all.
 

·
Registered
Joined
·
682 Posts
Quicks heads up.

I've put all the HV wires into orange tubes as thats a requirement for getting the car registered. Apart from that it looks cool and adds some protection of course. Will take some pictures sometime soon.
I installed the BMS on the front pack. I left the laptop with a small php script in the car and could watch the charging process in my warm appartment. What an improvement :)
Unfortunatly I also found, that the IR communication almost breaks down as soon as the inverter is running. Didn't happen with the trunk pack so it must be caused by the proximity to the inverter. On the plus side, if the communication catches some values, they look plausible.

I also made a decision on the DC/DC issue. There are devices available for 500V+ for around 300€ (http://www.feas.de/netzteil-snt9412-p-588.html). But since I'm a cheapshit I will use my ebay AC/DC converter (14V, 20A) to charge the 12V battery while the main pack is charged. Should I run into trouble I can always upgrade to something better. Don't comment on that, I'll try it anyway :D

Apart from that I tackled the legal issues. Or at least tried to. Here is the best case: http://zeropolomobil.blogspot.de/
He ended up at an inspector that spent 3h looking at the car, trying it out etc. and in the end approved. No EMC/EMI certificate was needed so the whole procedure cost him 300€.

My own research was frustrating as hell. "Experts" told me about "be prepared to spend 10000€ on expert opinions", "the laws have changed significantly 2 years ago" bla bla. I don't know why they do it, maybe they have orders to keep DIY off the road.

Anyway, the zeropolo story gave me hope again that its not so difficult after all.
Good to know you are up and going again,
Thought you were all set some time ago? Where did you get the orange tube?

I had the same issues. Everything worked nicely on bench, put the motor on the car and the encoder signal gets erratic values. This happens because the extra torque/magnetization under heavy load makes the wires act as antennas and its a real pain in the ass. To solve I designed a circuit with a few transistors to make the signals low impedance (Using current mode) and the host converts this back into a high impedance voltage source (1's and zeros).

The DC-DC you can use a center tapped pack with 2 HP server power supplies. They have a current share pin, so you don't run the risk of unbalancing the pack. They can work in current mode if the impedance is high enough (lead accessory battery) and two are good for 100Amps. 20Amps will get you nowhere when you decide to turn the fan to defrost the windows, headlights, wipers, rear defroster... I have 50Amps and on cold rainy days I run 80% load with all the crap on and a charged battery.

Get used to that with DIY... is always a battle. I think some people are just jealous, then is my friend said that a friend told him story ;)

Keep up the good work
 

·
Registered
Joined
·
971 Posts
Discussion Starter #12
Good to know you are up and going again,
Thought you were all set some time ago? Where did you get the orange tube?

I had the same issues. Everything worked nicely on bench, put the motor on the car and the encoder signal gets erratic values. This happens because the extra torque/magnetization under heavy load makes the wires act as antennas and its a real pain in the ass. To solve I designed a circuit with a few transistors to make the signals low impedance (Using current mode) and the host converts this back into a high impedance voltage source (1's and zeros).

The DC-DC you can use a center tapped pack with 2 HP server power supplies. They have a current share pin, so you don't run the risk of unbalancing the pack. They can work in current mode if the impedance is high enough (lead accessory battery) and two are good for 100Amps. 20Amps will get you nowhere when you decide to turn the fan to defrost the windows, headlights, wipers, rear defroster... I have 50Amps and on cold rainy days I run 80% load with all the crap on and a charged battery.

Get used to that with DIY... is always a battle. I think some people are just jealous, then is my friend said that a friend told him story ;)

Keep up the good work
I got the orange tube from ebay. A week later I found it in the local hardware store, too.
About DC/DC like said: I'll try if the battery alone supplies all the loads for a typical drive. I will switch the lead battery for 4 LFP40AHA that are left over. 40Ah = 40A for one hour :) The most I've seen were 30A (Radio, wipers, lights, rear defroster). If it doesn't work I'll get an actual DC/DC converter built for the job.

Ok, apart from that here is my idea of the heater and my idea of a solid state relay:



What you see is an isolated DC/DC converter, an 800V npn-transistor and a hair dryer core. The transistors base is switched by the thermal contact inside. If it switches off, the transistor switches off and the led switches on. I've only tried this with some 140V, will get it in the car later.

Besides that I've installed a plug in the fuel filler region as well as an RCD and an interlock switch behind the rear panels. Car plugged in -> inverter won't start. Inverter running and car is plugged in -> inverter is shut down.

Now it started snowing and I'm done for today.
 

·
Registered
Joined
·
971 Posts
Discussion Starter #13
Ok, I have done a few things in the last couple of months and meanwhile the car has done 1500km with no real (see below) problems.

I have installed the heater core from a cheap heating ventilator in the spot where the original water heater resided:


It is switched by a 1200V IGBT that is operated using an isolated DC/DC converter.

More things are:
1. Installed the EMW charger. Its set to charge at a humble 4A - at 500V though
2. Installed a fuse box. 6A for the heater (it draws about 4.5A), 6A for the charger, 4x25A in parallel for the drive system.
3. Hooked up the temperature gauge. It only ranges from 60-120°C where the motor can operate up to 170°C. Oh well... At 125°C an LED starts flashing (original logic from the VW gauge system)
4. Broke the fuel gauge when testing it...
5. Equipped the inverter with the latest revision hardware.

There are two more things to do before attempting another visit to the TÜV people:
1. Get the fuel gauge to display something meaningful
2. Indicate that the car is ready to go. I want to add an LED to a reserved spot in the original VW gauge

Problems that occured in the 1500km test run:
1. On one occasion the inverter shut down while driving for no apparent reason. I.e. one of the shutdown signal fired for no reason. The inverter could be restarted by giving it the "start" signal. Never happened again. But I have a hunch it might happen again...
2. After driving pretty much full throttle for 15km the inverter shut down because the motor apparently overheated. Back then I was driving the motor with more slip than necessary so I hope this is not going to happen again.
 

·
Registered
Joined
·
971 Posts
Discussion Starter #14
I fixed the fuel gauge and installed a "ready to drive" LED. It is in the blind spot that is only used by Diesel cars for the pre-glow (or whatever) light. The ECE-100 requires such a light because there is no engine hum to tell you the car is ready to go. I might find its fairly annoying at night and place another resistor on it. It is hooked up to the DC switch signal:



FB means "fahrbereit" - drive ready.





Here is the final layout of the engine compartment. All battery monitoring modules are installed and are working nicely. The auxillary battery is amde up of 4 40AH LiFePos and are still being charged by the AC/DC converter while the regular pack is charged. In summer its full after 30 minutes or so.

The fuse box contains 4x25A fuses in parallel for the inverter, 6A for the charger and 6A for the heater.



Orange tubing underneath the car.
 

·
Registered
Joined
·
971 Posts
Discussion Starter #15
Last week I installed a current sensor board and a beagle bone acting as central unit to my homebrew BMS.

The current sensor board also contains a relay and a PWM output. I use the latter to drive the fuel gauge, The relay turns off the charger when done.

Today was rather frustrating. I installed new software on the voltage sensing board (26 of them). As a side effect some boards lost their calibration data and I needed to unmount and recalibrate them.

Then I figured that the boards restarted every now and then sometimes loosing their address and/or calibration data. I ended up ripping them all out as to finish software flashing on my desk.

That is work waiting for me that week.

I then decided to read a book on an almost fatal mountaineering tour to convince myself that there's worse than spending 3 days bent over a battery pack
 

·
Registered
Joined
·
682 Posts
Last week I installed a current sensor board and a beagle bone acting as central unit to my homebrew BMS.

The current sensor board also contains a relay and a PWM output. I use the latter to drive the fuel gauge, The relay turns off the charger when done.
Just a warning: newer gauges wont accept a PWM signal for the fuel gauge. They have their internal "pulse check" logic and go into a fault mode if the signals don't match.

A digital POT is what I am using on mine. Damm technology ;)
 

·
Registered
Joined
·
971 Posts
Discussion Starter #17
Just a warning: newer gauges wont accept a PWM signal for the fuel gauge. They have their internal "pulse check" logic and go into a fault mode if the signals don't match.

A digital POT is what I am using on mine. Damm technology ;)
That is stupid indeed. So the digital potentiometer, is that an IC? What do the gauges expect? Current? Voltage?
 

·
Registered
Joined
·
682 Posts
The cluster sends a 5V pulse and measures the voltage drop to GND across a resistive element (the fuel sender). The higher the voltage the lower the fuel level. Anything out of a certain range and the needle goes to zero

If the voltage is different from zero between pulses it goes into fault.
 

·
Registered
Joined
·
971 Posts
Discussion Starter #19
The cluster sends a 5V pulse and measures the voltage drop to GND across a resistive element (the fuel sender). The higher the voltage the lower the fuel level. Anything out of a certain range and the needle goes to zero
Thats horrible :eek:

I hooked up the BMS central computer yesterday to be booted by the DC switch signal. Must have made a mistake and blew the transistor. That means: open up the controller and swap the transistor. That in turn means: remove the front battery pack because it sits on top of the controller.

Anyway, I'm glad I had to. When I opened up the controller it was soaking wet inside. That explains why the bus voltage measurement was sometimes off. Actually I'm surprised the controller still worked as it did.

But worse to come: The screws of 2 phase cables had almost fallen out of the IGBT modules. Only 3 more turns and they were out. I'm totally puzzled by the fact that the car still drove normally.

I will clean everything up today, renew the sealing and give the screws more torque.
 

·
Registered
Joined
·
971 Posts
Discussion Starter #20
Finally! I passed the EMC Test!



We tested as defined in the ECE R-10. One run from the left side, one from the front while the motor idled at 6000rpm, one idle run with the car turned off.

The poor car had to sit a the rev limiter for almost 2h. The motor heated up to 120°C.

The EMC expert of the company I work for did all the tests. That is really really nice of him. But he seemed very excited about the project and glad to help. He'll still get the mandatory case of beer.

The diagrams a a bit hard to read. Basically you do the runs and then compare them to the idle run. All frequencies where the idle run surpasses the limit are masked out.

All remaining frequencies that surpass the limits are scheduled for another run at different antenna heights. Sometimes the peak is gone because it was caused by a bypassing airplane or other artifacts like that.

So, by these standards the limits were never surpassed.

Here are some things that I did towards EMC compliance:
- Use shielded AC cables
- Use snubbers and good film caps on the DC bus
- Put a 1nF cap from DC- to GND and DC+ to GND
- Put a ferrite around the DC cables

All DC cables that run through or underneath the car are UNSHIELDED :) I'm glad to still pass the test.
 
1 - 20 of 63 Posts
Top