AEVA Homepage
Forum Home Forum Home > Technical Discussion > Technical and conversion discussion
  New Posts New Posts RSS Feed - Pre Charge Resistor
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

Pre Charge Resistor

 Post Reply Post Reply Page  12>
Author
Message
KDRYAN View Drop Down
Groupie
Groupie
Avatar

Joined: 28 September 2008
Location: Cairns
Status: Offline
Points: 130
Post Options Post Options   Thanks (0) Thanks(0)   Quote KDRYAN Quote  Post ReplyReply Direct Link To This Post Topic: Pre Charge Resistor
    Posted: 25 April 2011 at 10:22am
Question: What would be the minimuim length of time a pre-charge resistor would need to be powered up to protect the controller?

On my vehicle the 120vdc pack is split into 2 x 60vdc units divided by a contactor. I am placing a timed "on" relay on the main contactor. Idea being when the ignition key is turned on the pack splitter contactor will connect the pack supplying 120vdc to the precharge resistor which is accross the main contactor, at the same time triggering the timed "on" relay which will energise the main contactor after x amount of time and hopefully not damage the controller. The precharge resistor is a 1Kohm 10W

Thanks
Back to Top
Electrocycle View Drop Down
Senior Member
Senior Member
Avatar

Joined: 19 October 2008
Location: Sydney
Status: Offline
Points: 1038
Post Options Post Options   Thanks (0) Thanks(0)   Quote Electrocycle Quote  Post ReplyReply Direct Link To This Post Posted: 25 April 2011 at 10:58am
it usually only takes a few seconds.
You could go to a smaller value precharge if you need to (100ohms or so)
The Engine Whisperer - fixer of things
Back to Top
coulomb View Drop Down
Senior Member
Senior Member
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 2799
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 25 April 2011 at 11:02am
Originally posted by KDRYAN KDRYAN wrote:

Question: What would be the minimum length of time a pre-charge resistor would need to be powered up to protect the controller? The precharge resistor is a 1Kohm 10W

I would give it about 3 time constants. The time constant depends on the capacitance of your controller. Suppose you have 10,000 uF on the controller. Then t = RC is the time constant, = 10^3 x 10,000 x 10^-6 = 10 seconds. In this case, you'd need 30 seconds, which is too long.

For 120 V nominal (call it 150 V max), 1K will limit the current to 150 mA, which is very conservative. I'd go for say 5 A, so about 150/5 = 30 ohms.

So then (assuming 10,000 uF, a total wild guess; it could be plenty more) you would have t = 30 . 10^-2 = 0.30 seconds, so 3 time constants is about a second.

Note however that the resistor is now conducting 5 amps briefly, so the instantaneous power will be around 5^2 . 30 = 750 W. I'd use at least a 50 W resistor, or more likely two 50 W resistors (15R) in series. 750 W might be too much; others may be able to suggest the right peak power.

You may find it difficult to look up the capacitance of your controller (and whatever else is connected, like a DC/DC, that also has capacitors). You could use your 1K resistor and a multimeter to estimate it. One time constant will be the time for the voltage across the caps to increase to 63% of the total. So if your pack is 130 V at the time of the test, time how long it takes to get to 130 x 0.63 = 82 V. Then you can either calculate the capacitance, or just scale it (so to get 30x faster precharge, use a resistor of 1/30 the value).

[ Edit: warning: your 1K resistor will charge the capacitors so slowly that it may have time to get hot. It may get very hot; 130^2/10^3 = 17 W, so that 10 W resistor may need a heatsink. Or at least make sure it doesn't burn you or any plastic nearby. ]

Edited by coulomb - 25 April 2011 at 11:12am
Working with Weber on finishing an electric MX-5
Back to Top
KDRYAN View Drop Down
Groupie
Groupie
Avatar

Joined: 28 September 2008
Location: Cairns
Status: Offline
Points: 130
Post Options Post Options   Thanks (0) Thanks(0)   Quote KDRYAN Quote  Post ReplyReply Direct Link To This Post Posted: 25 April 2011 at 11:39am
The controller is a Kelly KDC12603 (120vdc sepex / regen) and also a IOTA DLS-55 DC/DC converter. I cannot see the uF value on there web site.
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: 25 April 2011 at 12:06pm
Like coulomb said, you'll need to know the controller capacitance to work out the precharge value.

Your contactors should have a datasheet that describes the effect of closing them without precharging. The relevant numbers for the EV200 type HV vacuum contactors are something like 80% precharge = 50 cycles, 90% precharge = 50000 cycles. You should charge to >90%, which takes 3T, like coulomb said, where T = RC (R in Ohms, C in Farads).

The matter is confused if the DC/DC starts up and begins to draw power while you are precharging - it will probably never successfully get to 90% because of this power draw. This is another good argument for precharging quickly, as the extra current draw by the DC/DC won't affect the precharge rate too much.

The 'correct' way to precharge is to only turn the main contactor on when the voltage difference across the precharge resistor is small. You also need a timeout so if there is a fault, you don't cook the precharge resistors, which aren't rated for full power.

Waiting a fixed length of time only works if there are no faults in the system, or anything like a DC/DC drawing power during precharge.

Coulomb, the slow/fast precharge thing doesn't matter for resistor heating, it's exactly the same number of Joules dissipated in the resistor in either case - the amount stored in the caps when they are full.
Back to Top
KDRYAN View Drop Down
Groupie
Groupie
Avatar

Joined: 28 September 2008
Location: Cairns
Status: Offline
Points: 130
Post Options Post Options   Thanks (0) Thanks(0)   Quote KDRYAN Quote  Post ReplyReply Direct Link To This Post Posted: 25 April 2011 at 12:19pm
Thanks for the above information. I have just noticed that EV Works market a Automatic Precharger unit, that will do exactly what I am trying to achieve. (product code ZEVA-AP1.0)
Many thanks
Back to Top
antiscab View Drop Down
Admin Group
Admin Group

Forum Admin

Joined: 25 November 2007
Location: Perth, WA
Status: Offline
Points: 2132
Post Options Post Options   Thanks (0) Thanks(0)   Quote antiscab Quote  Post ReplyReply Direct Link To This Post Posted: 25 April 2011 at 1:09pm
If looking for a large enoguh resistor is causing problems, try using a mains rated light bulb of around 100W.

The EV Works solution is the best all round though, I agree :)
Matt

2007 vectrix - 96'000km
1997 Prius - Plug in prototype
Back to Top
coulomb View Drop Down
Senior Member
Senior Member
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 2799
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 25 April 2011 at 9:47pm
Originally posted by Tritium_James Tritium_James wrote:

The matter is confused if the DC/DC starts up and begins to draw power while you are precharging - it will probably never successfully get to 90% because of this power draw.

Wow, I never thought of that. If your DC/DC decides that this is a great time to pour 55 A into your battery at 14 V, that's 770W plus losses, at least 800 W. At 130 VDC, that's over 6 A, and if it starts drawing current at say 90 VDC, then that's close to 9 A. Precharging at 130 mA (with a 1K resistor) is not going to work in this case!

My initial thought is that you would want to run the DC/DC from some auxiliary contacts on the main contactor, so that this doesn't happen. But then, the DC/DC will have some significant capacitance, and may wear out the auxiliary contacts. So you need a precharge circuit for the DC/DC as well...

So I guess the best way is to leave the DC/DC across the controller (i.e. it gets precharged at the same time as the controller), but you need to find a way to disable the DC/DC (or make it draw very little current) until the main contactor closes. So if you did have auxiliary contacts on the main contactor, they might be able to be used to disable the DC/DC until precharge is over. If the auxiliary contacts have enough current rating, perhaps they could be between the DC/DC output and the auxiliary (12 V) battery. Or better yet, doing some digital enabling of the DC/DC.

I think we can set our DC/DC current limit (on our Mean Well power supplies) with an external resistor; hopefully it will be possible to set the current limit very low. Dang: I just checked, and the current can only be set to a minimum of 50%, and only with the "B" model (we ordered the "A" model, the only one Mouser stocks). Also, our contactors are all SPST (no auxiliary contacts). Looks like we'll need a low voltage relay that can handle the 30 ADC that the DC/DC power supplies can put out. That assumes that the DC/DC supplies won't draw much current with no output, apart from that needed to charge the input capacitors.

I can imagine that a typical power supply used as a DC/DC would draw quite a spike of current at initial turn on. The 12 V battery has likely not been charged for some time, so its self discharge will mean it is ready for a boost. Also, at least some contactors have just been closed, and they take a spike of current at first switch-on. (In our case, to get voltage to the pre-charge circuit, something like 13 contactors have to close, with an initial surge of some 4 A each. We may have to stagger our contactor closings so that the 12 V supply doesn't collapse if the 12 V battery happens to be weak.)
Working with Weber on finishing an electric MX-5
Back to Top
antiscab View Drop Down
Admin Group
Admin Group

Forum Admin

Joined: 25 November 2007
Location: Perth, WA
Status: Offline
Points: 2132
Post Options Post Options   Thanks (0) Thanks(0)   Quote antiscab Quote  Post ReplyReply Direct Link To This Post Posted: 25 April 2011 at 10:46pm
you could use the "start" position of your ignition to disconnect the 12v battery from the dc-dc.

The EV Works module may not let the main contactor close until the cap voltage is high enough, but the disconnected 12v battery may stop the dc-dc from drawing too much current to prevent that from happening.
Matt

2007 vectrix - 96'000km
1997 Prius - Plug in prototype
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: 26 April 2011 at 8:28am
I don't think the capacitors in the input to the DC/DC will cause too much of a problem for the contactors, they're going to be several orders of magnitude less capacitance compared to a motor controller (well, not ours...)

We get away with having the DC/DC on the controller side of the contactors in the Civic because the DC/DC has a soft-start built in where it takes a few seconds to fire up and then a few more before it's at full power. We've precharged in under a second, so the DC/DC is effectively out of the picture as far as that's concerned.
Back to Top
weber View Drop Down
Senior Member
Senior Member
Avatar

Joined: 23 January 2009
Location: Brisbane
Status: Offline
Points: 1808
Post Options Post Options   Thanks (0) Thanks(0)   Quote weber Quote  Post ReplyReply Direct Link To This Post Posted: 26 April 2011 at 11:52am
I spent several hours yesterday trying to find a cheaper and physically smaller option than the 200 amp EV200 contactors ($65 on eBay), for isolating a bunch of things which are connected to our twin 400 Vdc traction busses but pull (or push) less than 10 amps. The motor-controller precharge resistor is one of those things. The others are the twin PTC-ceramic heater/demisters, the twin 12 V DC-DC converters and the twin 240 Vac battery chargers.

Zeva's precharge controller is great, but it won't handle our voltage.
http://www.evworks.com.au/index.php?product=ZEVA-AP1.0

The best online prices I could find for Kilovac PD5 and PD10 contactors were at http://onlinecomponents.com. The astonishing thing is that these contactors capable of switching only 5 or 10 amps cost two to three times as much as an EV200! I can only guess this is due to lower volume of sales. It seems that the EV200 is the LV DC switching workhorse of the planet. The only thing cheaper than a Tyco/Kilovac EV200 in this range of voltages and currents appears to be a Tyco/Kilovac LEV200, but this does not have the coil economiser and so pulls about 1 amp for the 12 V coil and the only ones in stock seem to have 24 V coils.

To get something cheaper than an EV200 I had to go down to 0.5 A (@ 400 Vdc) when reed relays start to come into the equation. Of course 0.5 A is of no use for chargers, heaters or DC-DCs but it is just barely usable for the precharge resistor (but only because we're using a Tritium Wavesculptor controller).

I looked for the bus capacitance in the Tritium datasheet and didn't find it, so I went and measured ours, at 800 uF. Why not put it in the datasheet, James?

So an 800 R precharge resistor would limit the current to 0.5 A, in which case 3 time constants (95% charge) would be 3 * 800R * 800uF = 1.92 seconds, which is acceptable.

The suitable reed relay was a Coto 5503-12-1 which is about $28 from Digikey or OnlineComponents, but neither have it in stock. The only stock I could find was at Element14 where the price was $55.
http://au.element14.com/coto-technology/5503-12-1/reed-relay-spst-no-12vdc-3a-thd/dp/1079784

[ Edit: even though it's specified at 3500 V and 3 A, there is a limit of 200 W of switched power, so these are actually totally unsuitable anyway. ]

This is barely cheaper than an EV200 but it does have the advantages of smaller size, PCB mounting and lower coil current.

I measured the capacitance of our DC-DCs and got about 200 uF. I think we'll have to have separate precharge for those.

Edited by weber - 30 December 2011 at 3:41pm
Putting the finishing touches on MeXy the electric MX-5
Back to Top
gmacd33 View Drop Down
AEVA Member
AEVA Member
Avatar

Joined: 11 January 2010
Location: Brisbane
Status: Offline
Points: 89
Post Options Post Options   Thanks (0) Thanks(0)   Quote gmacd33 Quote  Post ReplyReply Direct Link To This Post Posted: 26 April 2011 at 11:59am
The capacitance is in the user manual, on the top of page 20. Perhaps you were looking at the datasheet?
Greg MacDonald
Australian Electric Infrastructure & Transport
Wireless power transfer to electric vehicles while driving
043-848-0011
greg@olevaustralia.com.au
Back to Top
bga View Drop Down
Senior Member
Senior Member
Avatar

Joined: 01 September 2008
Location: Perth WA
Status: Offline
Points: 496
Post Options Post Options   Thanks (0) Thanks(0)   Quote bga Quote  Post ReplyReply Direct Link To This Post Posted: 26 April 2011 at 12:13pm
My thought is that the main contactor pullin signal (12V across a coil) can also be used as an 'enable' to prevent the motor controller doing anything until the main power supply is properly on.

The same could be used to inhibit the DC-DC converter until it's safe for it to operate.

Some delay to allow the main contactor to close may be required so that it gets a chance to make good contact before anything happens.

I was looking on the Farnell (oops Element14, that is Silicon) website to get a handle on mountings, pricing and availability of suitable resistors.
The wirewound rod types (HL) are really cheap ($6.50 each for a 50 watt unit), but need a metal skewer thingy to mount them.

The 50 watt gold metal extrusion types (HSC50 ot HSC100 etc) are also inexpensive at about $7.50 and have screw holes for mounting to a bulkhead heatsink, but solder terminals, so temperature may be an issue.

It would seem that the price per unit increases approximately according to a squared rule. Thus, 100 watt is about $22 with 200W about $80. I think 50 watt units are looking attractive.

There seems to be stock (USA) of most of the values that are of interest for EV precharge resistors. Probably several in a series/parallel would give enough power handling to make the precharge resistor robust. I would guess that about 200 watts' worth would make the network hard to blow up.

Sticking a button type 75deg.C Normally Closed thermal cutout on one of the resistors and in series with the coil on the precharge relay should make it fool proof - good for testing, when things usually get blown up.
i-Miev - Now with sustainably sourced electrons :)
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: 26 April 2011 at 12:40pm
The correct solution to precharge is to NOT use a fixed time, it should be under control of a micro which monitors the voltage on both sides of the precharge resistor, and cuts in the contactor when they're within 10% of each other. It should also be keeping track of either time, or resistor temperature (or both) to detect the fault situation where the caps aren't charging up. Our old precharge controller (and now BMS master) does this.

bga, depending on the exact family you pick (some are better than others) those aluminium cased wirewound resistors can be up to 12x nominal power rating for brief periods of time, enough for a precharge event.

All the gory details for this stuff are in the user's manual for our (now discontinued) precharge controller. Get it here: http://cesium.office.tritium.com.au/TRI78.001v1_Users_Manual.pdf The interesting part for this discussion is section 8.3, on precharge resistor sizing.
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: 26 April 2011 at 12:44pm
Oh, and your themal cutout will probably just arc over and die if it opens with a few amps at 400V DC across it... don't count on it doing what you expect it to do with DC through it.
Back to Top
weber View Drop Down
Senior Member
Senior Member
Avatar

Joined: 23 January 2009
Location: Brisbane
Status: Offline
Points: 1808
Post Options Post Options   Thanks (0) Thanks(0)   Quote weber Quote  Post ReplyReply Direct Link To This Post Posted: 26 April 2011 at 1:34pm
Yes Greg, "datasheet" is what I wrote. But thanks for pointing out that it is at least in the User Manual.

James, BGA suggested the thermal cutout in series with the coil.

That EV200 operations figure plumetting from 50,000 to 50 seems crazy to me. It can only be because the peak current goes just over the apparently-very-sharp 650 A limit for makes. According to the other EV200 chart, if it was just slightly less than 650 A it would last for at least 10,000 operations. To see this, note that making at 80% charge on a 320 Vdc system means making with 64 Vdc across the contacts. Interpolate 64 V at 600 A on the other set of curves in this EV200 datasheet and see that it's at least 10,000 operations.

So there's nothing magic about whether it's 80% or 90% voltage. It's all about what the peak current is as it makes. And that is very hard to predict as it depends not only on the ESR and ESL of the controller caps but the resistances and inductances of the contactor contacts and the bateries and the cabling and fuses and current-measuring shunts.
Putting the finishing touches on MeXy the electric MX-5
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: 26 April 2011 at 1:40pm
Yeah, it's a bit of a nightmare trying to work out all the interacting factors. But the '50' figure at the worst-case is quite scary, so there's a big incentive to keeping away from whatever causes that. Allowing 4 or even 5 Tau for precharge time is much safer...

weber, it's in the datasheet too! ...now :)
Back to Top
coulomb View Drop Down
Senior Member
Senior Member
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 2799
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 26 April 2011 at 2:14pm
Originally posted by weber weber wrote:

The astonishing thing is that these contactors capable of switching only 5 or 10 amps cost two to three times as much as an EV200! I can only guess this is due to lower volume of sales.

Or maybe that cache of $65 EV200s on Ebay is keeping the price down while it lasts. Maybe EV200s will "go back to" $400+ when the Ebay supply runs out and there is no alternative . I hope not.
Working with Weber on finishing an electric MX-5
Back to Top
bga View Drop Down
Senior Member
Senior Member
Avatar

Joined: 01 September 2008
Location: Perth WA
Status: Offline
Points: 496
Post Options Post Options   Thanks (0) Thanks(0)   Quote bga Quote  Post ReplyReply Direct Link To This Post Posted: 26 April 2011 at 2:51pm
Hi James,

The thermal cutout has to be kept away from the traction DC, I suggested it on the precharge contactor's coil for this reason.

The delta V is the parameter that should control the main contactor, not a clock.

I was thinking of whose minimal projects that would otherwise simply smoke the resistors. Some design and testing is neded to ensure that it will work as intended.

I completely agree that the solution should be more elegant than that proposed above. The controlling parameter is the voltage across the main contactor.

I watched the UWA Lotus blow up the two (quite small) precharge resistors one day. I think it was caused by a foot on the accelerator pedal.
We had to push the car across the campus, fortunately is is light and rolls easily.

For my project, I was going to implement the precharge control and interlock in the motor controller DSP that already monitors the bus voltages and has an interest in the interlock.

I'm currently writing my BMS firmware and hope to get to this soon.
i-Miev - Now with sustainably sourced electrons :)
Back to Top
weber View Drop Down
Senior Member
Senior Member
Avatar

Joined: 23 January 2009
Location: Brisbane
Status: Offline
Points: 1808
Post Options Post Options   Thanks (0) Thanks(0)   Quote weber Quote  Post ReplyReply Direct Link To This Post Posted: 26 April 2011 at 4:44pm
Thanks James! That's what I call responsive.

It seems the best solution for us re DC-DCs is to give them their own contactors and not bother with precharge on those. I just did some tests with a 200 A 50 mV shunt connected to a DSO, and a 200 V 40 Ah lithium pack (the highest voltage we have wired up on the bench at present), and a sacrificial circuit breaker operated manually. The highest impedance device in the circuit is probably the 10 A moving-coil panel meter that we had wired in for various charger and heater experiments.

With our two DC-DCs in parallel and no precharge I see a peak of 80 A. With the WaveSculptor200 and no precharge I see a peak of 1080 A. So double both of those for 400 V. Even with 160 A at 400 V for the DC-DCs the EV200 datasheet says we should get 5,000 operations.
Putting the finishing touches on MeXy the electric MX-5
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: 26 April 2011 at 6:55pm
That's some interesting hard data - thanks! I'm not surprised it's over a kA for the Wavesculptor, the caps in it are very low impedance. Don't try the same experiment at a higher voltage though! The inductance of your cabling, batteries, etc, will cause some fairly high voltage overshoot on the caps, and you might break something if you try.

You might be able to put some (several watt) power resistors in series with your DC/DC inputs, to cut the inrush down a bit. I suspect they'll operate just fine with a few volts drop into their input while in operation...
Back to Top
coulomb View Drop Down
Senior Member
Senior Member
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 2799
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 26 April 2011 at 8:02pm
Originally posted by Tritium_James Tritium_James wrote:

You might be able to put some (several watt) power resistors in series with your DC/DC inputs, to cut the inrush down a bit. I suspect they'll operate just fine with a few volts drop into their input while in operation...

That's an interesting idea, TJ. The trouble is, for a fixed resistor, the numbers don't stack up. Let's consider our 225 W dc-dc units drawing 240 W from 400 V, or 0.6 A. To drop "a few" volts, make it 3 V, so that's a 3/.6 = 5 ohm resistor. That will limit the inrush current to 80 A, a little better than the 160 A that we might expect (but maybe that was for two... doesn't matter). But now the instantaneous power when the capacitors are totally discharged is 400^2/5 = 32,000 W! Even with the 1 second overload ratio of 25:1 that would require a 1280 W resistor.

What we really want is a high resistance initially, and a low resistance later, to limit the inrush current. Aha, what about an ICL (Inrush Current Limiter)? The infamous CL-40 found on many PC power supplies starts with 5 ohm resistance and decreases to 0.11 ohm at load current. Well, that's no better, but there are others in the series, such as the CL-90 (datasheet here). It starts at 120 ohms, and reduces to 1.18 ohms, with a full load current of 2 A. This would reduce the inrush current to around 400/120 = 3.3 A (plus or minus 25%; these are not precision devices). At 0.6 A, or 30% of rated load, the resistance might be about 6 ohms (interpolating); that would be a 6 * .6 = 3.6 V drop, or about 1% of the pack voltage. These things are designed to work on 240 V mains, which often go to 265 V, so surely they often see 265 * sqrt(2) = 375 VDC, so 400 V might be safe (I don't see a voltage rating for them). The CL-90 is supposed to be able to handle 1250 uF at 240 VAC.

Oops, I seem to remember suggesting this before, and having it shot down in flames. Perhaps someone will point out the flaw in the above.
Working with Weber on finishing an electric MX-5
Back to Top
coulomb View Drop Down
Senior Member
Senior Member
Avatar

Joined: 22 January 2009
Location: Brisbane
Status: Offline
Points: 2799
Post Options Post Options   Thanks (0) Thanks(0)   Quote coulomb Quote  Post ReplyReply Direct Link To This Post Posted: 26 April 2011 at 9:24pm
Originally posted by coulomb coulomb wrote:

These things are designed to work on 240 V mains, which often go to 265 V, so surely they often see 265 * sqrt(2) = 375 VDC, so 400 V might be safe (I don't see a voltage rating for them).

No voltage rating, but at last I found a recommended energy rating:

http://www.ametherm.com/datasheets/ms2212102.html

This one states 220 J; 200 uF at 400 V is 1/2 . C . V^2 = 1/2 x 200 x 10^-6 x 400^2 = 16 J. It also gives maximum capacitance at up to 680 VAC, suggesting that their voltage limit is quite high. It's about twice the volume of a CL-90 (same 23 mm diameter and 10 mm thick), so I'm guessing that a CL-90 will be fine (these particular Ametherm devices have an MOQ of 200). The 240 V capacitance for this unit is about 3x that of the CL-90, so presumably the CL-90 can handle about a third of the energy, or about 220/3 = 73J, a very comfortable margin.

Actually, I notice that the minimum current for the CL-90 is listed as 0.5 A. With the DC-DC at full output, drawing 0.6 A, this is OK, but hopefully they won't be running flat out all the time. The CL-140 has the lowest minimum current (0.2 A), but unfortunately it's smaller size means its 240 VAC capacitance limit is only 150 uF. The one with the lowest minimum current and well above 200 uF @ 240 VAC is the CL-180, with 0.4 A minimum current, 400 uF @ 240 VAC, and 16 ohms 25C resistance. So the current would be limited to about 400/16 = 25 A.

I wonder why they have a minimum current rating? Is it just that the resistance gets embarrassingly high, or it will get too hot, or what? The CL-40 commonly found in PC power supplies has a minimum current of 1.5 A; surely a PC power supply could be drawing less than 240 * 1.5 = 360 W?

It is possible that the power supplies we're using as DC-DCs already have ICLs in them. They're more likely one of the lower resistance types, so it's a matter of whether we want to reduce the inrush even more than they already have.

In email, Weber noted:
Originally posted by Weber in email Weber in email wrote:

I did find the inrush current somewhat inconsistent. It's possible I only got the 80 A result when I did a make just long enough after a break for the voltage to have decayed back near zero, but not long enough for the ICL to cool down.


Unfortunately, we can't just open one up and look inside, since they are potted for their IP65 rating.

Edit: minimum current -> minimum current rating

Edited by coulomb - 26 April 2011 at 9:26pm
Working with Weber on finishing an electric MX-5
Back to Top
weber View Drop Down
Senior Member
Senior Member
Avatar

Joined: 23 January 2009
Location: Brisbane
Status: Offline
Points: 1808
Post Options Post Options   Thanks (0) Thanks(0)   Quote weber Quote  Post ReplyReply Direct Link To This Post Posted: 26 April 2011 at 10:17pm
James, thanks for the precharge controller docco. I figure there has to be some high voltage switching device(s) inside your precharge controller box to switch the precharge resistor between charging and discharging. Do you mind telling us what you used?

I find it misleading to use an earth/ground symbol for one side of a floating battery, even if it is clearly a different symbol from that used for the chassis and even if they are both labelled accordingly.

But irrespective of the symbol, don't you think that both sides of a floating battery should be disconnected by contactors as close as practical to the positive and negative terminals, although only one of them needs to have precharge across it, and although any branch circuits for loads need only be isolated in one of the conductors?

This is required for floating batteries in standalone power systems (AS4509), floating PV arrays (AS5033) and separated (floating) AC power systems (AS3000). I believe the thinking is that because a single fault to earth (or in our case chassis) can go undetected in a floating system, either side may become lethal with respect to earth/chassis, and if only one side is switched it may be the wrong side, leaving the other side lethal, but with a false sense of security.

The same problem (of one-sided isolation, not misleading earth symbols) occurs in Figure 2 on page 16 of NCOP14 Jan 2011.
http://www.infrastructure.gov.au/roads/vehicle_regulation/bulletin/pdf/NCOP14_Guidelines_Electric_Drive_01Jan2011.pdf
Putting the finishing touches on MeXy the electric MX-5
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: 27 April 2011 at 6:01am
Weber, there was a couple of 1200V 10A IGBTs in there in a half bridge configuration, to switch the resistor either high or low.

There's nothing to stop you putting the 2nd contactor in the negative side of the pack (and it's how we have it drawn in our documentation for the BMS, not up on the website yet), the precharge still works the same way.

Someone changed my original drawing for the NCOP, which had contactors in both +ve and -ve.
Back to Top
 Post Reply Post Reply Page  12>
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 11.01
Copyright ©2001-2014 Web Wiz Ltd.

This page was generated in 0.203 seconds.