AEVA Homepage
Forum Home Forum Home > General > Members Machines
  New Posts New Posts RSS Feed - Weber and Coulomb's MX-5
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

Weber and Coulomb's MX-5

 Post Reply Post Reply Page  <1 3738394041 67>
Author
Message
Nevilleh View Drop Down
Senior Member
Senior Member
Avatar

Joined: 15 January 2009
Location: Tauranga NZ
Status: Offline
Points: 783
Post Options Post Options   Thanks (0) Thanks(0)   Quote Nevilleh Quote  Post ReplyReply Direct Link To This Post Posted: 05 December 2011 at 5:14am
Originally posted by weber weber wrote:



Quote I do hope that snail is alive and well!
You mean Speedy? Yes. In fact he told us we were a refreshing change from most of the humans who photograph him racing cars. All they want to do, he says, is paint an "S" on his shell so they can use the caption, "Look at the S-car go". I don't get it.


Yeah, right! Those reports of a strong smell of garlic thought to be emanating from your workshop are quite irrelevant then?

Back to Top
weber View Drop Down
Admin Group
Admin Group
Avatar

Joined: 23 January 2009
Location: Brisbane
Status: Offline
Points: 2477
Post Options Post Options   Thanks (0) Thanks(0)   Quote weber Quote  Post ReplyReply Direct Link To This Post Posted: 21 December 2011 at 2:13pm
Originally posted by Nevilleh Nevilleh wrote:

Originally posted by weber weber wrote:

"Look at the S-car go". I don't get it.
Yeah, right! Those reports of a strong smell of garlic thought to be emanating from your workshop are quite irrelevant then?
Baseless rumours.

Regarding our BMS: We have at last got rid of the bad packets that were caused by modifying the software to count bad packets.

So now we see no bad packets, even with the unmodified (low-current unfiltered) comms hardware, provided the motor isn't running.

Next job is to reinstall the contactor boxes we pulled out, so we can run the motor again. We pulled them out in order to do those mods to reduce the noise being put on the 12 V bus by the EV200 coil economisers. Before we put them back we want to make them complete. A few things were omitted in the haste to get them going for the EV festival.

One of those things was automatic precharge for the WaveSculptor. This is now working. We decided to keep it simple, as follows.



This provides a 560 ms delay for precharge, which is 7 time constants for the 100 R precharge resistor with the 800 uF capacitance of the WaveSculptor 200. It was necessary to go to a 100 W resistor to meet the resistor manufacturer's peak power and pulse energy requirements. But this resistor will still be suitable if we go to a 750 V system, provided the motor controller does not have any more bus capacitance than the WaveSculptor 200.

My son Hunter, who has just finished first year engineering, made the delay circuit up for us on veroboard (his first). You can see it installed here.


One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
Back to Top
Nevilleh View Drop Down
Senior Member
Senior Member
Avatar

Joined: 15 January 2009
Location: Tauranga NZ
Status: Offline
Points: 783
Post Options Post Options   Thanks (0) Thanks(0)   Quote Nevilleh Quote  Post ReplyReply Direct Link To This Post Posted: 24 December 2011 at 5:27am
BMS lookin' good. Mine gets the odd bad value, saw a few "FF" packets during a 20k drive and I don't seem to be able to link them to any particular throttle setting, except that I don't get any with the motor not running. However, there isn't enough errors to cause concern - you have to be watching the thing every second to see them. Logging framing errors hasn't shown any.
I'll be interested to see how yours gets on when you start driving regularly.
Must have a look on the 'net and see if I can discover how those "economisers" work. I haven't used any in my car, just good old solenoids from ZEVA. I think they draw about 2 or 3 amps each and I have 3 energised normally, so that's 100-odd watts being wasted.

Must be nice having some "child labour" to build things for you
Back to Top
jonescg View Drop Down
Senior Member
Senior Member
Avatar

Joined: 21 January 2010
Location: Albany, WA.
Status: Offline
Points: 2273
Post Options Post Options   Thanks (0) Thanks(0)   Quote jonescg Quote  Post ReplyReply Direct Link To This Post Posted: 24 December 2011 at 10:38am
Originally posted by weber weber wrote:


One of those things was automatic precharge for the WaveSculptor. This is now working. We decided to keep it simple, as follows.



This provides a 560 ms delay for precharge, which is 7 time constants for the 100 R precharge resistor with the 800 uF capacitance of the WaveSculptor 200. It was necessary to go to a 100 W resistor to meet the resistor manufacturer's peak power and pulse energy requirements. But this resistor will still be suitable if we go to a 750 V system, provided the motor controller does not have any more bus capacitance than the WaveSculptor 200.

My son Hunter, who has just finished first year engineering, made the delay circuit up for us on veroboard (his first). You can see it installed here.




The relay on the precharge circuit is a bit of overkill isn't it? Surely the current would only be a few amps; couldn't you get away with a standard relay?
I challenge you to come up with a better invention than the bicycle
Back to Top
Tritium_James View Drop Down
AEVA Member
AEVA Member

AEVA Corporate Member

Joined: 04 March 2009
Status: Offline
Points: 748
Post Options Post Options   Thanks (0) Thanks(0)   Quote Tritium_James Quote  Post ReplyReply Direct Link To This Post Posted: 24 December 2011 at 10:42am
Standard relays aren't rated for 450V DC.

In a fault situation (eg a fault in the motor controller, DC/DC, wiring, etc that presents a shorted bus), it will weld on, which then means you can't turn off the precharge, which means your precharge resistor catches fire (450V,100R =2kW in a 100W rated part) , burning your car to the ground...
Back to Top
weber View Drop Down
Admin Group
Admin Group
Avatar

Joined: 23 January 2009
Location: Brisbane
Status: Offline
Points: 2477
Post Options Post Options   Thanks (0) Thanks(0)   Quote weber Quote  Post ReplyReply Direct Link To This Post Posted: 24 December 2011 at 11:31am
Originally posted by jonescg jonescg wrote:

The relay on the precharge circuit is a bit of overkill isn't it? Surely the current would only be a few amps; couldn't you get away with a standard relay?

You could certainly get away with a lower current relay or contactor, but the precharge contactor must have the same DC voltage rating as the main controller contactor, namely the maximum battery voltage, which in our case is planned to eventually be almost 900 V at full charge.

But surprisingly the EV200 contactor turns out to be the most cost effective solution for our precharge even though it's rated for 63 A at 900 Vdc (10,000 operations) and we only need a 9 A rating for the precharge relay. An EV200 can be had for about $65 on eBay but lower-current 900 Vdc relays cost 3 times as much!

The same goes for the contactors for our heater/demister elements, our chargers and our DC DC converters. We end up using the bulky EV200s for them too, purely based on cost, even though a 10 A 900 Vdc relay would do.

I thought I was onto something once with 10 A 1000 Vdc reed relays, but it turns out they can't do anything like 10 A and 1000 V at the same time. Their typical maximum voltage-current product is 50 VA.

Breaking current at high DC voltages is really hard to do in a small space, as is making current without welding the contacts together.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
Back to Top
coulomb View Drop Down
Admin Group
Admin Group
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 3307
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 24 December 2011 at 11:39am
Oops! Parallel posting with Weber. Not to worry. If you wish, skip to the progress update, now underlined.

Originally posted by jonescg jonescg wrote:

The relay on the precharge circuit is a bit of overkill isn't it?

Yes, it's way overkill on the current capacity. As James has noted, something more than a "standard relay" is needed for the voltage rating.

We've found that all (so far) of the lower current relays we've found to date have been way more expensive than the ~ US$65 EV200 relays from Ebay. So we're using these high-rated relays not only for precharge but also for switching the charger (5.5 A max), heater (about the same), and DC/DC (less than 1 A), even though a 10 A relay with a 450 V rating would do those jobs. Well, actually since we plan to switch to the 750 V configuration, we actually need 900 V DC rating.

If anyone knows of a cheaper source for 10 A 900 V relays, we'd love to know about them (for the next MX-lightning conversion). We might even be able to shrink some of the contactor boxes, assuming that these lower rated relays are physically smaller, and that would also be welcome.

A quick progress update, since Weber seems to be busy. We spent last EV day preparing two contactor boxes for re-installation. We want to make sure that these are finished this time, not 3/4 finished as they were for the EV festival. Some of that involved figuring out exactly where some conduits are going to be placed, so we can cut the cables to the right length. Under the bonnet is getting quite full now, and we have to be careful to place anderson connectors out of the way of rain and dust. One promising place is behind the pop-up headlights, but it's a challenge dodging the moving parts for when the lights pop up or down. So that burned a lot of time.

Another thing that took time is the rethinking of the precharge wiring. We had used ordinary 30 A car hook-up wire, but realised that it really needs to be double insulated to handle the eventual 750 V configuration. Some of the precharge wiring passes close to the coil economiser circuitry, which is at chassis potential. It now runs in clear PVC tubing for double insulation.

All this is so that we can get the car running again, so that we can debug the BMS again, so that we can finish the BMS design and order BMS the parts, so that we can finish the battery packs and finally finish the car.

Edited by coulomb - 24 December 2011 at 11:44am
Learning how to repair and re-flash TC/Elcon chargers and PIP-4048 inverter-chargers.
Back to Top
weber View Drop Down
Admin Group
Admin Group
Avatar

Joined: 23 January 2009
Location: Brisbane
Status: Offline
Points: 2477
Post Options Post Options   Thanks (0) Thanks(0)   Quote weber Quote  Post ReplyReply Direct Link To This Post Posted: 24 December 2011 at 12:50pm
Originally posted by Tritium_James Tritium_James wrote:

In a fault situation (eg a fault in the motor controller, DC/DC, wiring, etc that presents a shorted bus), it will weld on, which then means you can't turn off the precharge, which means your precharge resistor catches fire (450V,100R =2kW in a 100W rated part), burning your car to the ground...

In our case we could still turn it off since we have main (or battery) contactors which are separate from our controller contactors. In fact 12 of them in series, to also break the battery up into ELV segments (< 120 Vdc). As shown here for the 900 V case:



The AS3000 definition of a main switch (or contactor) is one which isolates an entire installation (excluding emergency services) from a source of supply. This should be separate from the switches or contactors that isolate the various loads such as the motor controller.

We considered putting a fuse in series with the precharge resistor, e.g. a 500 mA 1000 V multimeter fuse. But we worry about what happens if it eventually fails un-noticed, possibly due to repeated high current pulses. And 500 mA at 750 V is still 375 W into a 100 W part.

We also considered using three 50 W 240 V halogen down-lights in series as our precharge resistor ala Lee Hart, with a clear polycarbonate lid on the box, as these could tolerate the full voltage indefinitely in the event of a fault, but still have low enough initial resistance for a fast enough precharge. But we baulked at the extra space required.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
Back to Top
Nevilleh View Drop Down
Senior Member
Senior Member
Avatar

Joined: 15 January 2009
Location: Tauranga NZ
Status: Offline
Points: 783
Post Options Post Options   Thanks (0) Thanks(0)   Quote Nevilleh Quote  Post ReplyReply Direct Link To This Post Posted: 24 December 2011 at 1:12pm
You could use a 230v ac 10A relay and put a high voltage mosfet across it. The idea is, you turn on the mosfet and then close the relay so the mosfet only carries the current for a few mSecs and the relay closes with almost no volts across it. Turning off is the reverse, turn on the mosfet, open the relay and turn the mosfet off.
No charge for this amazing idea!
Back to Top
weber View Drop Down
Admin Group
Admin Group
Avatar

Joined: 23 January 2009
Location: Brisbane
Status: Offline
Points: 2477
Post Options Post Options   Thanks (0) Thanks(0)   Quote weber Quote  Post ReplyReply Direct Link To This Post Posted: 24 December 2011 at 1:31pm
Originally posted by Nevilleh Nevilleh wrote:

You could use a 230v ac 10A relay and put a high voltage mosfet across it. The idea is, you turn on the mosfet and then close the relay so the mosfet only carries the current for a few mSecs and the relay closes with almost no volts across it. Turning off is the reverse, turn on the mosfet, open the relay and turn the mosfet off.
No charge for this amazing idea!

"Amazing" isn't quite the word I'd use.
If I used a high voltage MOSFET (or more likely an IGBT), why would I need a relay? An IGBT would work (Tritium used to make a precharge controller like that), but it would require additional circuitry to provide floating isolated gate drive, and regulations generally don't allow solid state devices to be used as isolators, so it would need to be backed up by a contactor in series.

Edited by weber - 06 January 2012 at 8:57pm
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
Back to Top
Nevilleh View Drop Down
Senior Member
Senior Member
Avatar

Joined: 15 January 2009
Location: Tauranga NZ
Status: Offline
Points: 783
Post Options Post Options   Thanks (0) Thanks(0)   Quote Nevilleh Quote  Post ReplyReply Direct Link To This Post Posted: 24 December 2011 at 3:34pm
Because the thing couldn't carry the current without a decent size heatsink. I thought you were an engineer? Never mind.
Back to Top
weber View Drop Down
Admin Group
Admin Group
Avatar

Joined: 23 January 2009
Location: Brisbane
Status: Offline
Points: 2477
Post Options Post Options   Thanks (0) Thanks(0)   Quote weber Quote  Post ReplyReply Direct Link To This Post Posted: 24 December 2011 at 4:55pm
Originally posted by Nevilleh Nevilleh wrote:

Because the thing couldn't carry the current without a decent size heatsink.

Sorry Neville. I thought you were joking. You probably aren't aware that the Tritium WaveSculptor200 has only 800 uF which is about 1/10th of the bus capacitance of a typical motor controller. So the MOSFET would only have to carry the current for a few tens of milliseconds anyway. The energy dissipated in the resistor (plus switching device on-resistance) is 0.5*CV^2 = 0.5 * 800u * 750^2 = 225 J. With a MOSFET this would be shared between the 100 R resistor and the MOSFET in proportion to their resistances. With an IGBT it would be approximately a constant 2 volts, with the initial current of 7.5 A dropping off with a 80 ms time constant. So only 15 W peak. Not likely to need a heatsink.

BTW, I meant to say, re assistance from family and friends: I do admire the way you just get on and get things done _without_ needing any such assistance. I have trouble staying motivated even when I have it!

Merry Christmas my friend.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
Back to Top
coulomb View Drop Down
Admin Group
Admin Group
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 3307
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 31 December 2011 at 12:00pm
With the Christmas to New Year break, we managed three EV days this week. Amazingly, it took us nearly all that time to get the three contactor boxes that are required to operate the vehicle finished (as far as possible) and installed. We're so confident about two of them that they actually have their rubber gaskets installed .

Part of the delay was figuring out where to route some of the conduits. I've already mentioned the one behind the headlights, and we've had to move another one that would was fouling the headlights already. We've decided that now that we can see where all the cables have to go, we really need to redesign one of the contactor boxes completely for next time (if there is one).

We decided to test the pre-charge circuit, to make sure that the voltage across the motor controller contactor was low enough when it connected after the pre-charge delay. This was a little tricky; we had to convince ourselves that the pack was properly floating, so we could connect one end of the pack (the positive end, as it turned out) to mains earth at the Digital Storage Oscilloscope probe. All worked well, and we saw the classic exponential rise (from negative voltage towards zero) as expected, then the very low and very steady zero volts when the contactor connected. There were two slight surprises. The voltage, even after about 5-7 time constants, didn't rise completely to zero; there was about 10 V across the contactor that would just not go away.

We assumed that this was probably due to a small load from the motor controller of about 10/100 = 0.1 A (our pre-charge resistor is 100 ohms), or about 24 Watts. Perhaps this drives some DC/DC converters or something, but I thought those should run off the 12 V from the CAN bus. Later, we noticed how slowly the capacitors on the motor controller bled down (the manual says 10 minutes). It seemed to bleed down a lot faster with the ~ 10 megohms of load from the multimeter! So now that I write this, I think it can't be a load on the motor controller high voltage input, or certainly not one as high as 100 mA. So that's a bit of a mystery.

The other surprise was that the voltage across the motor controller contactor appeared to have spikes of voltage on it, some of which were fifty volts in magnitude. I remember heading an ethereal voice saying "You have noise in your future!" and thinking "what chance does the digital BMS have in an environment like this?".

So we finally got to run the latest BMS software while running the motor up on stands. The result is: as far as we can tell, there is no problem. Arrgh!   

The best we can tell is that the way we were monitoring for faults way back when was actually injecting noise into the system. [ Edit: and that still happens, if we monitor that way. ] But there are other ways we can monitor, and while there are some corrupted characters to the monitoring system (e.g. the "V" in the voltage string will corrupt to an "R" or "F" about one in a hundred times), the communications between the BMS master (in the Driver Controls computer) and the actual cell-top BMUs (Battery Management Units) seems to be unaffected by noise. We can tell this by the lack of retrying, and the counts of checksum errors.

But can we really be sure the problem is fixed / never was there in the first place? Damn it, no we can't be sure... not unless we run the car as close to "real life" as we can get it!    So we were reluctantly forced to take the car down off its stands, and run it under its own power. Well, nothing we could do seemed to provoke any errors.

We want to give it one last try next EV day, to make sure we've provoked the BMS as much as we possibly can. (Damn! Another day when we have to move it around under its own power ). Maybe we'll temporarily attach one of the DC/DC units to keep the auxiliary battery charged, and maybe we'll even hook up the charger as well (after all, it's also a source of PWM noise; presumably less strong than the motor controller, but we should not make unnecessary assumptions). Also, the charger process (which happens under the control of the Driver Controls software) uses BMS communications, so this will be another test of BMS comms.

Edited by coulomb - 31 December 2011 at 12:02pm
Learning how to repair and re-flash TC/Elcon chargers and PIP-4048 inverter-chargers.
Back to Top
FreedomWon View Drop Down
Noobie
Noobie
Avatar

Joined: 03 January 2012
Location: Johannesburg
Status: Offline
Points: 3
Post Options Post Options   Thanks (0) Thanks(0)   Quote FreedomWon Quote  Post ReplyReply Direct Link To This Post Posted: 06 January 2012 at 1:44am
Dear Coulomb

I noticed you mentioned the GuanTuo BMS in one of your threads.

I have recently completed converting a Jeep Grand Cherokee to electric drive. The system includes Thundersky batteries, Elcon chargers and a BMS from Harbin GuanTuo Power Equipment Co (GTBMS005A).

I completed a few succesful test drives with everything functioning correctly but have now come up with a problem with my BMS. So far the manufacturer has been very slow with solutions so I am hoping you might be able to assist.

After previously functioning properly, displaying all cell voltages and controlling the Elcon chargers during the charging routine the central controller unit suddenly began alarming one day and all information on the seperate touch screen display dissapeared (the batteries are no longer displayed and all the settings are zeroed). I cannot charge the batteries because the Elcon chargers are not receiving the CAN signal to enable. The COMM LED on the display is no longer flashing as it was before. The red LED's on the voltage and current collection modules are flashing normally however.

GuanTuo suggested that I load their PC software to check if I can communicate with the central controller with my Laptop but their software will not load properly on my PC's (Windows XP or Windows 7). During the installation process I receive a Runtime Error report (three times) which I "ignore" but then the installed programme fails to open. So far I have been waiting for five weeks for a solution to this software problem from GuanTuo. They say the same installation file works fine on their PC's so it must be a problem with my particular software setup. I loaded all the .NET Framework versiond and it did not help.

Help with sorting out the software issue would be great so at least I can try to narrow down the problem by linking up my PC. Do you have any idea what I should try next?

Regards

FreedomWon
Freedom Won, One EV at a time.
Back to Top
coulomb View Drop Down
Admin Group
Admin Group
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 3307
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 06 January 2012 at 6:32am
Originally posted by FreedomWon FreedomWon wrote:

GuanTuo BMS...

I've started a new topic on the GuanTuo BMS here: GuanTuo BMS stopped working.
Learning how to repair and re-flash TC/Elcon chargers and PIP-4048 inverter-chargers.
Back to Top
coulomb View Drop Down
Admin Group
Admin Group
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 3307
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 06 January 2012 at 6:59pm
I'm grateful to our evil twins, Maxwell and Faraday, who visited today and took our MX-lightning for a drive. They said we have to fix that shudder, which might happen when the current limit kicks in; they nearly got stuck on a hill in front of a bus. Well, better them than us   

They must have floored it a few times, because when they came back we found two checksum errors logged in the BMUs. All 18 BMUs that we can easily access (a nineteenth is hard to connect to) logged the same error. So the error seems to have occurred between the BMS master (the Driver Controls Unit) and the first BMU.

So we have a plan of attack: we modify the first BMU with the lower-value pullup resistor after the opto receiver, and the 27 nF capacitor to absorb glitches. We have three BMUs modified with this and another change, and it's some encouragement that they didn't seem to record any extra errors. Fortunately, the first BMU is on its own, so we don't have to remove more than one BMU to make this mod. Then we get the evil twins back to see if they can cause a comms error with the modified BMU circuit.

We've done a few changes which should give those evil twins less to whinge about. The vacuum pump circuitry is now tidied up, and we've started preparing for the motor controller to be raised up, so the contactor box for the charger and DC/DC can be fitted. Fitting a DC/DC will possibly give the auxiliary battery enough oomph to drive the power steering pump.



As you can see, the controller follows the slope of the battery box quite closely now, so it fits will under the bonnet. We had to cut, bend and weld a bracket to make it fit. We'll need some extra welding soon to hold the motor controller in place.

We also connected some wires that the ECU usually takes care of, to enable the air conditioning clutch to pull in. Alas, the pressure switch doesn't detect enough refrigerant gas in the system. This came as a surprise to Weber, but I didn't think it had any chance after nearly three years of inactivity, and pushing and prodding the air conditioner pipes around to make things fit.

All in all, a pretty good day, even if the evil twins ended up having the most fun   

Edit: added image and following paragraph.

Edited by coulomb - 06 January 2012 at 8:10pm
Learning how to repair and re-flash TC/Elcon chargers and PIP-4048 inverter-chargers.
Back to Top
coulomb View Drop Down
Admin Group
Admin Group
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 3307
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 06 January 2012 at 7:54pm
Some more photos have come through from the film laboratory:



Here, the busiest contactor box designated A2 (though they are all pretty packed) is about 80% complete, and we had 95% of the cables poked in to convince ourselves that we will one day be able to pull off the miracle of finishing it completely.

Hopefully, the instructions on the lid are self explanatory; we eventually want to convert from two 375 V packs in parallel to one 750 V pack (made up of two 375 V packs in series, obviously).



These are boxes A4 and A5, ready to be covered and slid into place towards the right of the picture. Well, "slid into place" doesn't quite convey the agony of manipulating the conjoined twins into a rather tight space, with cables and Anderson connectors sprouting out every which way. But we did it, and hope we never have to open those boxes again.

[ Edit: that's an 80 amp fuse in the left hand box, and we plan to draw up to 240 A through that fuse for up to 10 seconds. Brainiacs Coulomb and Weber put the fuse (one of two) into one of the most difficult boxes to access. It's enough to make you stop believing in Intelligent Design ]

[ Edit2: had the wrong fuse maximum current and time. ]

Edited by coulomb - 06 January 2012 at 9:54pm
Learning how to repair and re-flash TC/Elcon chargers and PIP-4048 inverter-chargers.
Back to Top
Jeff Owen View Drop Down
Groupie
Groupie
Avatar

Joined: 13 November 2008
Location: Brisbane
Status: Offline
Points: 144
Post Options Post Options   Thanks (0) Thanks(0)   Quote Jeff Owen Quote  Post ReplyReply Direct Link To This Post Posted: 06 January 2012 at 7:55pm
Originally posted by coulomb coulomb wrote:

I'm grateful to our evil twins, Maxwell and Faraday, who visited today and took our MX-lightning for a drive. They said we have to fix that shudder, which might happen when the current limit kicks in; they nearly got stuck on a hill in front of a bus. Well, better them than us   



What, you have a bus service on the fire-break test track?
Back to Top
coulomb View Drop Down
Admin Group
Admin Group
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 3307
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 06 January 2012 at 8:13pm
Originally posted by Jeff Owen Jeff Owen wrote:

What, you have a bus service on the fire-break test track?

No, the BMS comms errors seem to need more provocation than our test track can provide. The evil twins took the MX-lightning on a drive around the block. Well, a few blocks, it seems. Evil!   
Learning how to repair and re-flash TC/Elcon chargers and PIP-4048 inverter-chargers.
Back to Top
weber View Drop Down
Admin Group
Admin Group
Avatar

Joined: 23 January 2009
Location: Brisbane
Status: Offline
Points: 2477
Post Options Post Options   Thanks (0) Thanks(0)   Quote weber Quote  Post ReplyReply Direct Link To This Post Posted: 06 January 2012 at 9:27pm
Today we also found out what happens when the encoder drops off the end of your motor shaft.

There's a thump and the WaveSculptor trips off due to "Hardware Overcurrent". Fortunately this happened while revving it in neutral in the carport.

So we decided to just epoxy the darn thing in using super-low-strength Araldite (5 minute ultra clear) and be done with it.

More info on our encoder mounting scheme in this earlier post.
http://forums.aeva.asn.au/forum_posts.asp?TID=980&PID=34430#34430

One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
Back to Top
Renard View Drop Down
Groupie
Groupie
Avatar

Joined: 29 August 2010
Location: Cobargo NSW
Status: Offline
Points: 257
Post Options Post Options   Thanks (0) Thanks(0)   Quote Renard Quote  Post ReplyReply Direct Link To This Post Posted: 07 January 2012 at 6:54pm
I generally admire your work, and I realise that you have very little space, but guys, that pre-charge contactor box is less than beautiful. I can't help feeling that such a degree of crowding could be risky.

On another matter, have you fully dropped the idea, discussed in "Product Idea Digital Fuse pp 5-6" of a soft earth? That is, a pack mid-point connection through a resistor to chassis, monitored by, I suggest, a sensitive optoisolator such as a 6N139 or HCPL2731 with high-gain output? It still seems a good idea to me.

Keep up the good work!

Renard
Renard
Back to Top
coulomb View Drop Down
Admin Group
Admin Group
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 3307
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 08 January 2012 at 6:24am
Originally posted by Renard Renard wrote:

I generally admire your work

Thanks.

Quote but guys, that pre-charge contactor box is less than beautiful. I can't help feeling that such a degree of crowding could be risky.

It doesn't look quite so bad with all (so far) of the cables attached.

Quote On another matter, have you fully dropped the idea, discussed in "Product Idea Digital Fuse pp 5-6" of a soft earth? That is, a pack mid-point connection through a resistor to chassis, monitored by, I suggest, a sensitive optoisolator such as a 6N139 or HCPL2731 with high-gain output?

Our priority at the moment is testing the BMS so its production can get started. We do intend to perform some sort of insulation monitoring. At this stage, a soft earth (as in a permanent medium resistor from centre pack to chassis) is not favoured. For one thing, half way through our paralleled 375 V packs works out to the middle of two battery boxes, and the middle will change as we add boxes to our total of 228 cells and 8 boxes (currently 73 cells and 3 boxes).

If we can't find a reasonably priced commercial insulation monitoring solution, we'll probably alternately connect each end through a high value (in the hundreds of kilo-ohm range) resistor, and monitor the current through something like a high CTR (Current Transfer Ratio, like its "gain") opto-isolator, as you have suggested. Something that bothers me about optos is that their CTR is so variable, and also that they vary over time. Another challenge is detecting small probing currents, while coping with much higher currents that could result from a short circuit to chassis at the far end. Finally, there is the issue of ignoring capacitance to ground without desensitising the monitoring. That might be done by testing alternate ends at a low frequency, say 10 Hz, and testing only in the middle of the probing, or alternatively averaging the currents from probing in both directions.

At one point in the digital fuse thread I did state in large letters "No high voltage floating pack without detection!". We're very conscious of being probably the first conversion to exceed about 600 V, so we'd hate to set a bad example.

Edit: added explanation of CTR.

Edited by coulomb - 08 January 2012 at 7:54am
Learning how to repair and re-flash TC/Elcon chargers and PIP-4048 inverter-chargers.
Back to Top
Nevilleh View Drop Down
Senior Member
Senior Member
Avatar

Joined: 15 January 2009
Location: Tauranga NZ
Status: Offline
Points: 783
Post Options Post Options   Thanks (0) Thanks(0)   Quote Nevilleh Quote  Post ReplyReply Direct Link To This Post Posted: 08 January 2012 at 7:26am
You might look at the HCNR201 for a linear optocoupler.

Edited by Nevilleh - 08 January 2012 at 7:28am
Back to Top
weber View Drop Down
Admin Group
Admin Group
Avatar

Joined: 23 January 2009
Location: Brisbane
Status: Offline
Points: 2477
Post Options Post Options   Thanks (0) Thanks(0)   Quote weber Quote  Post ReplyReply Direct Link To This Post Posted: 08 January 2012 at 11:25am
Thanks for calling us on that Renard. In fact Coulomb and I had already agreed that it needed to be completely redesigned from scratch. But just so you know it's not quite as bad as it looks in that on-the-bench photo, here's what box A2 currently looks like in the vehicle.



The shunt in the lower right has a u-shaped clear PVC cover over it and its attached cables, which separates them from the 12 V wiring (in clear PVC tube) that passes over them. The bolt through the solid tinned-copper link goes all the way down to the baseplate, so the link can't swing, and it is insulated with clear PVC tube.

The minimum cable bending radius is the killer in designing these things. You find you have to put a device such as a contactor as far as possible from the two entry ports for the cables that connect to it. But this inevitably leads to more cable crossings than you would otherwise have. So you really have to make use of the third dimension when laying these things out. Intuition developed in designing printed circuit boards is not only useless but positively harmful in designing contactor boxes.

Crossing cables must not press against each other or any other conducting parts with any noticeable force, so that if the insulation were to melt, the conductors would still not contact each other.

For example, the smaller of the two rear contactor boxes shown above (box a4) contains one contactor and 3 studs for joining cable lugs. It had already been completely redesigned from scratch before the photograph. Not even _one_ of the holes in the baseplate from the first pass was reused. The 3 studs are now set at 3 different heights, with the contactor terminals at a fourth height.

We try to maintain at least 3 layers of insulation between conductors at different potentials. The design of the EV200 contactors is pretty dumb in this regard. They have single insulated ELV coil wires right near the bare HV terminals, and the attached lugs typically only have a thin layer of heatshrink. I think there were some photos posted in another forum showing EV200s that looked like they might have arced via the coil wires.

[Edit: Re-worded second-last paragraph. Added others.]

Edited by weber - 08 January 2012 at 12:21pm
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).
Back to Top
Renard View Drop Down
Groupie
Groupie
Avatar

Joined: 29 August 2010
Location: Cobargo NSW
Status: Offline
Points: 257
Post Options Post Options   Thanks (0) Thanks(0)   Quote Renard Quote  Post ReplyReply Direct Link To This Post Posted: 08 January 2012 at 2:41pm
Weber and Coulomb, I would not have had the temerity to comment except that I know this is not just any old conversion but is a quality job.
Certainly it looks better connected up, and the clear tubing looks good, but I just hope the multimeter probes don't slip when you're out on the roadside diagnosing a problem...

I also wondered how to mate the split conduit with the boxes, and here is my less-than-perfect solution. I don't know if you can buy these anymore as they were designed for the old-style grey flexible conduit. The thread is 16mm (smallest size) and the ferrule on the end has been filed down.



Edited by Renard - 08 January 2012 at 5:44pm
Renard
Back to Top
 Post Reply Post Reply Page  <1 3738394041 67>
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 11.03
Copyright ©2001-2015 Web Wiz Ltd.

This page was generated in 0.328 seconds.