Renard's BMW

Post up a thread for your EV. Progress pics, description and assorted alliteration
User avatar
weber
Site Admin
Posts: 2790
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Renard's BMW

Post by weber »

Renard wrote:I need a little help with my IGBTs please.

I followed Johny's heater control circuit, except that I used two IGBTs (one for each side) and 600V 40A ones for my lower voltage. I put 0.47uF caps across the IGBTs, 15V zeners from gate to emitter, and am careful to install them with pins shorted. Well, they work for a short while, then go short collector-emitter, and in doing so, take out the opto.
I wonder what I'm missing here.
My fall back position is to simply put in a contactor since I don't need any PWM, but that seems inelegant.

I assume you are referring to this circuit of Johnny's.
viewtopic.php?title=electric-heater&p=2 ... 359#p29868
It looks fine to me. But you say you put 0.47 uF across the IGBTs. You should not have any caps directly across any IGBT terminals. Johnny's circuit doesn't have them. A cap across collector emitter could explain the failure of your IGBTs (due to massive peak current from the cap at turn-on).

Do you have the fast recovery diode between collector and positive supply as in Johnny's circuit? Omitting that would explain your failures (due to turn-off voltage spikes from the inductance of the wires to your heating element).

The other thing you should be aware of is that PTC ceramics initially decrease their resistance with increasing temperature. So the current rises before it starts to fall. You should ensure your IGBTs can take the peak current.

Also, paralleling IGBTs isn't such a great idea. Unlike MOSFETs, they don't tend to share the current. They tend to be "loser takes all".

We put in a contactor. Kilovac EV200. Seems elegant enough to me.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
User avatar
Johny
Senior Member
Posts: 3733
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Renard's BMW

Post by Johny »

Renard wrote:...., except that I used two IGBTs (one for each side) and 600V 40A ones for my lower voltage. I put 0.47uF caps across the IGBTs, 15V zeners from gate to emitter, and am careful to install them with pins shorted.
When you say you placed a 0.47uF accross the IGBTs - whay do you mean by that?
User avatar
Johny
Senior Member
Posts: 3733
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Renard's BMW

Post by Johny »

I do regard the PWM of the ceramic element as a bit of an overkill. If I turn it on at full current it draws 6A for about 1 second then trails off to about 2 Amps. At 1200W it's very good but not excessive. A contactor would have been easier but then again I like the idea of winding the heat down to the lowest level that will do the job.
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Renard's BMW

Post by Renard »

Johny wrote: When you say you placed a 0.47uF accross the IGBTs - whay do you mean by that?

With my limited understanding of electronics, I was thinking that the caps should be across the switches as they are in AC equipment. They were from collector to emitter. However, I was having the same trouble before I put them in. Perhaps also the diodes I put in were not fast recovery types -- if that matters.
The BMW has two separate controls for the left hand and right hand sides, and I had wanted to preserve this duality (no quarrels between driver and passenger over temperature) by having two separate optos and IGBTs -- not paralleling them.

But now I'm leaning towards the contactor option, though I have room for only one, as it would be much safer. The last thing I want to happen is for the heater to switch itself on without the blower fan.
Renard
User avatar
weber
Site Admin
Posts: 2790
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Renard's BMW

Post by weber »

It may be the long wires you have on the collector and emitter. The triangle made by the IGBT CE, the fast recovery diode and the DC bus capacitor should have the smallest perimeter and area possible. Yes, Fast recovery matters. Move your 0.47 uF to be across the series combination of IGBT and FR diode. Any inductance in this triangle will cause voltage spikes across the IGBT.

Edit: Even mechanical switch contacts should never have a pure capacitor across them. They may have a snubber which looks like a capacitor, but actually has resistance in series with its capacitance.
Last edited by weber on Sun, 19 May 2013, 16:09, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
zeva
Senior Member
Posts: 422
Joined: Sat, 15 Dec 2007, 05:09
Real Name: Ian Hooper
Location: Australia
Contact:

Renard's BMW

Post by zeva »

The 0.47uF capacitor across IGBT probably isn't doing much harm nor good, except maybe harm to itself, and possibly a small amount of snubbing. Either way I'd be inclined to remove it, or put it in series with say a 10ohm resistor to make a proper RC snubber.

What's your switching frequency? If you're using a 10ohm gate resistor like Johny did, with diode having poor reverse recovery time, my money would be on "shoot through" due to the IGBT turning on before the diode has turned off - effectively a short circuit backwards through the diode and on through IGBT for maybe 100nS, with corresponding huge current spike that kills semiconductors.

I agree with weber's suggestion about keeping the IGBT, flyback diode and ripple capacitor(s) as close as possible to avoid parasitic inductance. Luckily if you're using a 600V IGBT on a 370V system, you do have a lot of headroom for a bit of inductive spikes/ringing.

FWIW in my RX7 I did a similar but somewhat simpler heater controller based on a 600V 50A MOSFET switch, tiny little flyback diode and no ripple caps. It PWMs at a very lazy 10Hz, below hearing frequency but plenty fast enough for a heating element which changes temp slowly. Because the element inductance is modest and it's only switching at 10Hz, the flyback diode can be tiny. It also means you can use a much bigger (ohms) gate resistor which gives softer switching, because switching losses are negligible at 10Hz. Just a thought.

See also: http://zeva.com.au/Projects/RX7/?sectio ... es#Heating
Ian Hooper
--
"Never doubt that the work of a small group of thoughtful, committed citizens can change the world." - Margaret Mead
http://www.zeva.com.au
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Renard's BMW

Post by Renard »

Thank you folks for the help on the IGBTs. But I want to get this car registered before 26th June when the three months grace after the rego expiry is up. So I'm going for the contactor rather than continue experimenting.
Since I don't need PWM, and I don't need to control the heat other than on/off (because the BMW has a climate module where one sets the desired temperature on the dash and it outputs a signal correspondingly to what used to be the water valves) the contactor will be switching only occasionally.
Renard
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Renard's BMW

Post by Renard »

A BIG DAY!

The car is reborn and has escaped from the shed.

I'm reminded of a little verse by Ogden Nash:

Deep in the study of eugenics
There lies that fabled fowl the phoenix,
It lays one egg, not ten or twelve,
And when it's hatched out pops itselve.

I drove up and down the road, up to about 80km/hr. Of course there are a few little faults, but those are for another day.

Image
Renard
User avatar
Johny
Senior Member
Posts: 3733
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Renard's BMW

Post by Johny »

CONGRATULATIONS. Great work.
User avatar
woody
Senior Member
Posts: 1716
Joined: Sat, 21 Jun 2008, 02:03
Real Name: Anthony Wood
Location: Mt Colah

Renard's BMW

Post by woody »

Great work Renard!

Edit: you need to work on your "EV Grin" though...
Last edited by woody on Thu, 30 May 2013, 17:50, edited 1 time in total.
Planned EV: '63 Cortina using AC and LiFePO4 Battery Pack
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Renard's BMW

Post by Renard »

During my test run up and down the road the other day, I was too nervous to observe everything accurately, so I want to renew the test drives asap after I've fixed the faults that emerged.

1. The motor fan failed to come on, and when I checked the motor temperature ten minutes after coming home, it was still over 90deg. Perhaps that had caused a power time-out I had experienced.
It was just due to a poor solder joint on the control FET.
2. The power steering pump fuse blew, so I replaced it with a 25A fuse. It blew again, so I went to 30A. This time it lasted longer until I used full lock. I'll try 40A. Since the pump wiring is all 2.5mm^2, and the full lock moments are brief, I'm not concerned that this would cause problems.
3. As described earlier, the heater control is being re-worked.
4. I had installed the coolant flow pump and reservoir early on in the build, and it was not until I had mounted the Wavesculptor that I realised that the reservoir was slightly below the level of the Wavesculptor's coolant block. I made up a revised bracket to raise the pump about 40mm to keep the coolant head above the aforementioned block.
5. The BMS master readout goes bananas when significant current is drawn by the Wavesculptor. Yes, there's interference. It disappears when the Master re-sets. This problem may be more difficult to treat, but it's one of those issues that can wait until I've had the car approved and registered.
6. Another job which has nothing to do with the drive, is that over the last year of fiddling around, I've managed to ruin the 17Ahr gel-cell auxiliary battery, probably by flattening it with interior lights from open doors and suchlike. Fortunately I bought a spare a few months ago; replacement entailed detaching the two ends of the four-cell pack in the boot.
7. I had not yet tried to control the PS pump by the car's speed. This was mainly because my old oscilloscope died a couple of months ago, and I have only just overcome my stinginess to buy a new one. I hooked up the new 'scope to the vehicle speed signal which emerges from the Instrument Cluster (X17 pin2) and goes to several modules around the vehicle: the radio, climate control, cruise control (now removed), and wiper module. For 3-series BMW owners, this is a 0.5 black/white wire. The signal is 13V or so at rest, with short negative-going pulses at speed, supposedly 13 pulses per 10km/hr, though my speedo seems to think the ratio is 1.2 not 1.3.
My initial set-up was to have three pump states: full-on; part-on and off, with two 5k 20-turn trimpots determining the transition set-points. Full-on draws about 18A at wheel centred, and part-on draws 9A. The set-points remain to be adjusted after actual driving experience. I have not yet tested this on the road.

I've been foolish: I put a Nanfeng 50A contactor from EV works in the heater control. I have one as my auxiliary contactor, but I forgot that that one never breaks under load as it operates simultaneously with three Kilovacs which split the pack. Result: my first smoke.

Image

My other foolish omission was failing to test the shunting side of the BMS cell top boards when I checked them after they'd been soldered up. So far three have failed to shunt and have had to be replaced. But in an amazing piece of luck, they are all among the 37 boards in the boot and easily accessed.
This fault appeared when I charged the pack up to the charger's maximum voltage of 401.5V. The BMS master showed a cell sailing on past 3.6V and showed which one it was. I replaced the board and the same process was repeated with another runaway board.
Renard
User avatar
weber
Site Admin
Posts: 2790
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Renard's BMW

Post by weber »

Thanks Renard, for telling it like it is. I can identify with so many of those things from early tests of our stuff too.

I see the Nanfeng contactor pages on the EV Works website are lacking the all-important voltage rating, and when you click on the supposed datasheet link all you get is a bunch of Chinese text.
http://www.evworks.com.au/index.php?product=REL-ZJW50A

[Edit: Coulomb pointed out that it contains the Arabic numerals "404", indicating a bad link (not just a Chinese version of the page you want).]

I found this page that says they are only good for 80 V, but I guess you know that by now.
http://www.corpmarket.com/company/DC-co ... ngnan.html
Last edited by weber on Mon, 03 Jun 2013, 05:07, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Renard's BMW

Post by Renard »

weber wrote: Thanks Renard, for telling it like it is.


Well, I wanted to sweep my screw-ups under the carpet, but then I realised that they could possibly be helpful to others. And present a more realistic picture -- that it's not just you who makes mistakes.
Renard
User avatar
Johny
Senior Member
Posts: 3733
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Renard's BMW

Post by Johny »

Yes, it's tempting to paint a rosier picture but it doesn't help anybody. My thanks too Renard, for being honest about the problems.
I think I started one of my blog entires with "Bloody prototypes".
That's what our cars are now. I get the feeling that you're enjoying the "ride" though.
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Renard's BMW

Post by Renard »

I fixed a few faults after my first run on 20th May, and today took the car out on the road again. This time I was not quite so anxious and was able to make more observations.

1. A 40A fuse for the PS pump doesn't blow.
2. I adjusted the cut-out speed for the pump upwards as I found cornering at 60km/hr rather heavy with no PS.
3. The comms issue remains, but that can wait.
4. The ABS light remains on all the time, but for all I know it may still work. I will experiment on a nearby gravel road soon.
5. Preliminary observation suggests a power consumption of about 150Whr/km. About what I had expected though not as good as I'd hoped.
6. I'm still working out where to be in the gears. I found that regen. is quite weak in top gear, and I have to shift down to get good braking.
7. The motor gets quite hot fairly quickly. The fan is noisy, though one notices it only outside the car. Next time I'll measure the temperature from the thermistor. On the other hand, the Wavesculptor stays as cool as a cucumber -- well, not over about 35 deg.

More driving soon!
Renard
User avatar
Johny
Senior Member
Posts: 3733
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Renard's BMW

Post by Johny »

Renard wrote: 5. Preliminary observation suggests a power consumption of about 150Whr/km. About what I had expected though not as good as I'd hoped.
My early figures were not great either. Longer trips help just as they do with ICE although I think you may have an issue - see point 7.
6. I'm still working out where to be in the gears. I found that regen. is quite weak in top gear, and I have to shift down to get good braking.
Can you increase regen in the controller?
7. The motor gets quite hot fairly quickly. The fan is noisy, though one notices it only outside the car.
This is a concern. Mine is direct drive and the only time I have had it over 40 above ambient is in stop start traffic for about 15 minutes. It made it to 70 degrees on a 30 degree day - 132 frame alu motor. (Measurement is thermistor in windings) My fan is a 72 Watt 172x150mm.
How sure are you that the Wavesculptor is set up correctly for this motor? (Edit later - I see you got them as a pair - strange getting so hot though).
I guess the question has to start with how hard were you driving?
Last edited by Johny on Fri, 31 May 2013, 19:11, edited 1 time in total.
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Renard's BMW

Post by Renard »

Johny wrote: Can you increase regen in the controller?


It would require re-programming the Driver Controls unit.
But perhaps the present situation is desirable as it allows for better "fuel" economy. With top gear at 1:1 and the diff at 2.93:1 the car is at 39km/hr for 1000rpm -- quite a slow motor rotation. And I am very used to changing down to brake. In more urban environments I would be in a lower gear anyway.
Renard
User avatar
BigMouse
Senior Member
Posts: 600
Joined: Thu, 28 Oct 2010, 02:39
Real Name: Vincent Tannahill
Location: Silicon Valley
Contact:

Renard's BMW

Post by BigMouse »

Renard wrote:
Johny wrote: Can you increase regen in the controller?


It would require re-programming the Driver Controls unit.
But perhaps the present situation is desirable as it allows for better "fuel" economy. With top gear at 1:1 and the diff at 2.93:1 the car is at 39km/hr for 1000rpm -- quite a slow motor rotation. And I am very used to changing down to brake. In more urban environments I would be in a lower gear anyway.
Do you have an externally driven (electric) fan on the motor? If you're relying on the integral fan at 1000rpm, that could explain your motor heating issue. Also, I would expect that the motor would be more efficient closer to its rated RPM. Higher RPM means higher phase voltage and lower phase currents, so less copper losses. Lower current also makes for a cooler motor.
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Renard's BMW

Post by Renard »

BigMouse wrote:
Do you have an externally driven (electric) fan on the motor?

Also, I would expect that the motor would be more efficient closer to its rated RPM. Higher RPM means higher phase voltage and lower phase currents, so less copper losses. Lower current also makes for a cooler motor.


Yes, Papst 12V fan rated at 100 litres/s.
But thanks for the remark about rated rpm, which is 1800rpm. (1800rpm corresponds to 70km/hr in top gear.)

But it's early days yet, and I may be jumping at shadows.
Renard
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Renard's BMW

Post by Renard »

I took the car for a 23km Sunday afternoon drive. Novelty value.

I pushed the speed up to 100km/hr. Very smooth. (Well, it's a BMW after all.)
I found that things go better when I move through the gears more like the ICE original. Not that every gear is necessary, but it provides smoother running and better acceleration. The gear changing requires getting used to, as the motor revs behave differently from the ICE.
On returning home I measured the motor temperature: 85-90 deg.
About the comms interference: it sometimes occurs when I just put the car into gear -- electrical gear that is -- and nothing more, which at least tells me that it's nothing to do with the DC bus.
In 4th or 5th gear, the car runs away down a hill, so at some later date I think I will re-compile Dave Keenan's regen code with REGEN_MAX set to a value greater than 0.25. Because I can always reduce the regen just by flicking down the slider pot. But such refinements can wait.
Renard
User avatar
BigMouse
Senior Member
Posts: 600
Joined: Thu, 28 Oct 2010, 02:39
Real Name: Vincent Tannahill
Location: Silicon Valley
Contact:

Renard's BMW

Post by BigMouse »

Renard wrote: On returning home I measured the motor temperature: 85-90 deg.


That is excessively hot. Did the Tritium come programmed for that motor, or did you run the motor ID sequence? If the latter, maybe run it again?
Renard wrote:About the comms interference: it sometimes occurs when I just put the car into gear -- electrical gear that is -- and nothing more, which at least tells me that it's nothing to do with the DC bus.
I imagine that the gate drive power supply, likely individual isolated SMPS, are enabled when you put the car "into gear" as you put it. The switching noise from the gate drivers can carry through DC bus, even when there is no actual switching taking place. Also, depending on how the Tritium works, "in gear" may mean 0v is being applied to the motor. 0v on a PWM inverter is 50% duty cycle, so there may be switching taking place even with the car just being enabled but not moving.
Last edited by BigMouse on Sun, 02 Jun 2013, 17:14, edited 1 time in total.
Tritium_James
Senior Member
Posts: 683
Joined: Wed, 04 Mar 2009, 17:15
Real Name: James Kennedy
Contact:

Renard's BMW

Post by Tritium_James »

The gate drive supplies in the WS200 begin switching as soon as 12V is applied to the controller. There won't be any noise from them making out back outside of the controller either on the 12V on on the DC bus connections.

The high power switches begins switching only when a non-zero command is sent to the controller, ie when you step on the pedal.

Is the 85-90°C temperature the winding temp (as reported by the WS200 from the thermistor on the windings?) or the case temperature?
User avatar
Johny
Senior Member
Posts: 3733
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Renard's BMW

Post by Johny »

Tritium_James wrote:The high power switches begins switching only when a non-zero command is sent to the controller, ie when you step on the pedal.
Does that mean you only magnetise the rotor moments before requesting torque James?
User avatar
Johny
Senior Member
Posts: 3733
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Renard's BMW

Post by Johny »

Way back in the first post in this thread, Renard mentions the thermistor:
Renard wrote:There are three modifications to the motor: replacing the fan with a 12V electric fan mounted on the NDE shroud; machining down the NDE shaft to fit an encoder; gluing a 100k thermistor onto the windings.
So it's looks to be winding temperature.
Renard
Groupie
Posts: 219
Joined: Sun, 29 Aug 2010, 18:55
Real Name: Robert Fox
Location: Cobargo NSW

Renard's BMW

Post by Renard »

Tritium_James wrote: Is the 85-90°C temperature the winding temp (as reported by the WS200 from the thermistor on the windings?) or the case temperature?


The windings.

I glued in another thermistor in order to control the motor fan, as well as the one that reports to the WS. It was that winding temperature that I was measuring, as I haven't yet set up a system to use the gauge drives from the DCU.

Bigmouse has suggested checking the WS config file. I had followed all the parameter capturing procedures apparently satisfactorily according to the User's Manual. But I have just noticed that -- I've got a bit confused here -- I may have left the vehicle mass at 30kg. Or perhaps not. I will check the WS later today by hooking up the Bridge. I don't know if that would have affected the situation.
Renard
Post Reply