Weber and Coulomb's MX-5

Post up a thread for your EV. Progress pics, description and assorted alliteration
User avatar
Johny
Senior Member
Posts: 3729
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Weber and Coulomb's MX-5

Post by Johny » Thu, 03 Jun 2010, 22:22

Point taken re alternator...

Tritium_James
Senior Member
Posts: 683
Joined: Wed, 04 Mar 2009, 17:15
Real Name: James Kennedy
Contact:

Weber and Coulomb's MX-5

Post by Tritium_James » Fri, 04 Jun 2010, 04:01

a4x4kiwi wrote:TJs solution does give an over voltage protection over 15v however.
It doesn't actually. The lockout in the pump software (how overkill does that sound!) only prevents the pump from *starting* if it's overvoltage. Once it's started, you can turn the voltage way up and it just keeps running.

User avatar
weber
Site Admin
Posts: 2623
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Fri, 04 Jun 2010, 04:19

How about a 470 uF 6 V capacitor in series with a 22k resistor, across the 4k7 which is the low side of the voltage-sense divider. Size of cap based on observation that it takes about 4 seconds to start. Probably never fit the cap inside the case.

Maybe there is another model pump from Koolance that would be more suitable for future Wavesculptor users, even if it costs a bit more?
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
coulomb
Site Admin
Posts: 3766
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Fri, 04 Jun 2010, 05:37

Now that most of the kinks are out of the software, I've put together a test jig for up to 5 BMU boards, appropriately made from end end of a Sky Energy crate:

Image     Image

I can't help it, I just love those LEDs, even from the back. These boards don't have bypass resistors yet; they're at the other factory. I figure that when testing, we'll want access to what will ultimately be the bottom of the boards, with all the components.

The serial communications wires are on gold plated 0.1" headers; they work nicely with our gold flashed boards. (The gold flashing isn't all that impressive in most lighting; usually it looks like tarnished tin plate. It seems like it must be 3 atoms thick). I have a 9 V battery also with a 2-pin header for testing the end-of-row repeater.

[Edit: the gold flashing looks actually somewhat gold-like in the second photo, possibly due to the reduced light.]

The connector at the bottom is connected to a pair of hand matched 2K resistors. The idea is to use a multimeter on the 2 V range to accurately set the cell voltage; reading 1.800 V across the 1:2 voltage divider means the cell voltage should be close to 3.60 V. I don't know if I'll bother using it much; it's too fiddly setting the voltage on my cheapo Electronics Australia kit power supply. Oops, my age is showing there.

The thick wires were to cope with the bypass current; alas the clip leads that I used with the chunky alligator clips have wire that has very little copper and would be flat out coping with one ampere.

Each new BMU board has to be flashed individually using the JTAG port (4 pins under the yellow LEDs). This downloads the bootstrap loader loader, and when run, copies the bootstrap loader and calibration data to the info flash. This can be done without external power or communications leads. One the boards have the bootstrap loader programmed, I can connect up the power and comms wires, and send the interpreter to the boards in parallel. I can send a simple command to turn on the LEDs to test them, though the error LED is already tested by the calibration process. The error LEDs also light at half brightness with every byte sent over the comms loop.

Edit: the second image seems to have deleted itself; I've just uploaded it again. It happens to be my favourite photo, too!
Last edited by coulomb on Tue, 15 Jun 2010, 16:58, edited 1 time in total.
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.

User avatar
coulomb
Site Admin
Posts: 3766
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Fri, 04 Jun 2010, 05:41

weber wrote: How about a 470 uF 6 V capacitor in series with a 22k resistor, across the 4k7 which is the low side of the voltage-sense divider.

I think we don't want to defeat the speed reduction with greater applied voltage, unless it causes the motor to run slower than it would at 12.0 V. What we'd want is to run the motor at the same speed at 13.X volts (whatever float voltage we use) as it is designed to run at 12.0 V.

If we did need to squeeze in a 470 uF capacitor, a tantalum (or two) might fit.
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.

User avatar
weber
Site Admin
Posts: 2623
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Fri, 04 Jun 2010, 06:03

coulomb wrote:I think we don't want to defeat the speed reduction with greater applied voltage, unless it causes the motor to run slower than it would at 12.0 V. What we'd want is to run the motor at the same speed at 13.X volts (whatever float voltage we use) as it is designed to run at 12.0 V.
I guess you're right. So I'll put a 27k in parallel with the 4k7 and it should start at voltages up to 14.5 V and it should think that 13.4 V is 12 V.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2623
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sat, 05 Jun 2010, 00:42

We've installed the 27K resistor, and now the pump starts up to somewhere between 14.5 and 14.6 V. Here is the resistor in place:

Image

The resistor needs to be hung off the edge of the PCB to fit. Here is is installed in the case:
Image
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2623
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Tue, 15 Jun 2010, 23:01

[This is Coulomb posting from Weber's Wigwam. Coulomb has relied on his password being saved on his laptop for so long that he can't remember it.]

The world's ugliest and most complex battery boxes are ready for painting:

Image

Note the three separate cages in each box, the cutouts for the chassis rails, the mounting angle iron, some of which are at small anges to two axes, the many bolt holes with welded-captive nuts, and the general fiddlyness of the boxes. Not shown are two gusseted brackets for attaching to the fronts of the lower wishbone arms (there is another battery box attaching to the backs of the lower wishbone arms).

All that work to hold 13 and 16 cells, or 5.7% and 7.0% respectively of the total 228 cells.

[Edit: Added non-Weber disclaimer]
Last edited by weber on Tue, 15 Jun 2010, 17:38, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
woody
Senior Member
Posts: 1715
Joined: Sat, 21 Jun 2008, 02:03
Real Name: Anthony Wood
Location: Mt Colah

Weber and Coulomb's MX-5

Post by woody » Wed, 16 Jun 2010, 00:08

weber wrote: The world's ugliest and most complex battery boxes are ready for painting:

All that work to hold 13 and 16 cells, or 5.7% and 7.0% respectively of the total 228 cells.
Moral of the story, convert a car with plenty of contiguous space for the batteries :-/
Planned EV: '63 Cortina using AC and LiFePO4 Battery Pack

User avatar
coulomb
Site Admin
Posts: 3766
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Wed, 16 Jun 2010, 03:05

woody wrote: Moral of the story, convert a car with plenty of contiguous space for the batteries :-/

You've converted me! (Eeh, poor pun, sorry.   Image )

But seriously, the amount of work needed to pack 200+ cells into a compact car like the MX-5 is something that we vastly underestimated.

The trouble is, Weber believes that now that we've done all this hard work, we have to convert a few more to amortise the effort! Amortisation was always the plan, but I didn't think we'd have so much work to amortise.

It's true that another MX-5 will be a lot less work, but I think it will still be a lot more work than something with more room. A Mercedes A- or B-class perhaps. But that's a fair way into the future.
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.

User avatar
weber
Site Admin
Posts: 2623
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Wed, 16 Jun 2010, 04:15

I note that the previous "Weber" post was actually from Coulomb. I don't think they are ugly at all. Complex to design, yes. There may never in the history of EVs have been so much time expended by so many brilliant minds (Coulomb and Newton) on the design of a box for so few cells.

An enormous amount of trial-and-error was involved as well as a lot of lying underneath the car looking and thinking. Fitting as many cells as possible in the space was one thing, with chassis rails and handbrake cables and hydraulic brake lines and propshaft and PPF all passing through that space. Figuring out how to anchor the boxes to the vehicle to withstand a 20 gee collision was another. But I think they are works of art, and copying them would be a piece of cake compared to designing them.

I won't ask you to make them, Coulomb, if we do another MX-5. Image

These boxes are located in prime battery real-estate on the MX-5. As low as possible and as close to the centre of gravity as possible. Hence the extraordinary effort for a mere 1/8th of the pack.

The Mercedes doesn't sound too bad either. But if your aim is to produce a high performance EV, and you're not jamming cells into every possible location, with the possible exceptions of the ashtray and glovebox, then it wouldn't be as high-performance as it could be, would it? Image

You know the one about the guy who liked to do everything the hard way. He was telling someone that, when they asked whether he had any children. He replied, "Yes. And before you ask: Standing up, in a hammock". Image
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
EV2Go
Senior Member
Posts: 2059
Joined: Wed, 16 Jul 2008, 00:21
Real Name: Paul
Location: Brisbane 1963

Weber and Coulomb's MX-5

Post by EV2Go » Wed, 16 Jun 2010, 20:51

It’s the extra effort that makes a mediocre conversion into a work of art. Maybe it is just the masochist in me, but whenever I have to fight tooth and nail for something to work, it often gives me the greatest satisfaction... and maybe I just like to do things the hard way Image

User avatar
coulomb
Site Admin
Posts: 3766
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Fri, 18 Jun 2010, 16:02

We've been resigned to having to reinstall the alternator in place of a DC/DC, because of our high pack voltage (228 cells at a maximum of 3.65 V under charge (with a little extra for cable and contactor losses) comes to 832 V; even half the pack comes to 416 V. We were just able to get a charger for this (the 312 V nominal Elcon has a maximum voltage of 417 V!).

The only power supply application requiring such high voltage input has been the industrial DIN rail power supply that runs off two mains phases, when a neutral isn't available. Even then, most industrial gear runs of 24 or 48 V these days; it's difficult to get the 12 V model in a high enough power version, and being industrial gear, it's expensive unless you luck finding a source of second hand equipment.

However, there is a new(ish) segment of the power supply market: LED lighting. It's finally getting to the situation where outdoor high power LED lighting makes sense. So there has been for a while some cheaper supplies offering constant current output (ideal for driving an array of LEDs without series resistors), but of course there is the usual limitation of 265 VAC maximum, or 370 VDC maximum. What we'd really want is an LED supply that goes to about 300 VAC, or about 420 VDC.

Well, it seems that somehow American lighting finds itself with a source of 277 VAC. This makes no sense to me, except that 277 * √3 = 480 V, so maybe they sometimes have 3-phase transformers wired in star, and they tap one of the 480 V windings for 240 and 120. Or some more complex arrangement like High leg delta that somehow makes 277 VAC available for lighting. There must be a market for it, as Mean Well have produced their HLG-240H series:

Image

The image is of a non-H series (so 370 VDC max), which is slightly cheaper. They are in a weatherproof IP65 (IP67 available) enclosure, with heat conductive potting (so no fiddling the resistors inside to change the voltage, unfortunately). The image is also of the 12 V nominal model; there are also 15 V, 20 V, and 6 other models up to 54 V nominal. The H models offer up to 295 VAC input, or up to 417 VDC (the same as the maximum output of the Elcon charger). They are all around the 240 W mark, although the lower voltage models derate slightly (e.g. 15 V @ 15 A = 225 W nominal).

[Edit: the two black spots at the bottom right hand corner of the case are rubber stops over trimpots to adjust the current and voltage limits. The IP67 version doesn't have these, but you can adjust at least the current limit with an external resistor. I've never seen the IP67 version ("B" in the last letter of the part number) for sale anywhere.]

So a pair of these is ideal for our MX-5: we can use them in parallel across the 365 V nominal (416 V peak) pack, and in series (across the two half packs) when we go to the 730 V pack (832 V peak). The combined current capability will be 30 A, which hopefully will be enough. If we need more, we could add another pair; they're also available in lower power models (but note the MOQ of 24 for these). We've just ordered a pair from Mouser. At AU$195, they're a bit pricey, but a lot cheaper than the industrial alternatives we've been looking at.

One potential problem is that the open circuit voltage has only about a +/- 5% range; the 15 V model is adjustable from 14-16 V, the 12 V model from 11.X to 12.8 V. We'd like 13.5 to 13.8 to float a lead acid auxiliary battery. However, we figure that when the DC/DC will be on, so will a bunch of contactors, and likely some pumps, possibly lighting and heating, the BMS master, the controller logic, and all the other loads that a vehicle has (just not ignition and ECU now). This will likely pull the 14 V open circuit voltage down to well under 13.8 V under the minimum expected load. Even if the voltage exceeds 13.8 for a short while, as long as it doesn't exceed the gassing voltage of 14.4 V, it should do no harm (and lead acid batteries like a "hard charge" now and then, which 14 V might come closer to than say 13.5 V).

[Edit: I note that the 5 A model has a wider voltage adjustment range, 13.5 to 17 VDC.]

So if you have an AC conversion where the pack could exceed 370 VDC with regen, you might consider these or similar models. There is only one left at Mouser since we ordered two this morning, but they might be available elsewhere (I didn't find any last night, at least none that gave the price).

Thanks to Pascal (Warrick) for the idea many weeks ago of trying to find industrial power supplies to use as a DC/DC. It led (no pun intended), rather indirectly as it turns out, to this find.
Last edited by coulomb on Mon, 25 Apr 2011, 20:14, edited 1 time in total.
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.

User avatar
Johny
Senior Member
Posts: 3729
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Weber and Coulomb's MX-5

Post by Johny » Fri, 18 Jun 2010, 16:17

Good find. I guess if 30 Amps isn't enough you could throw another one (or two for the 800 V split pack) at the problem. The cold, rainy nights are the problem.

Then again, for the 800V version of the pack you should be able to tap in at the 1/3rd points of the whole pack and use 240VAC SMPSs as I have seen SMPSs that can do 100-240VAC without switching. Maybe worth pursuing for the next one.

User avatar
coulomb
Site Admin
Posts: 3766
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Fri, 18 Jun 2010, 16:38

Johny wrote: Then again, for the 800V version of the pack you should be able to tap in at the 1/3rd points of the whole pack and use 240VAC SMPSs as I have seen SMPSs that can do 100-240VAC without switching.

Great point! In fact, the non-H versions will work 127 to 373 VDC (with some possible derating at the low end of the voltage range). [Edit: so they would work with 1/3 and 2/3 taps.] That would also allow tapping at the half way point of the 114 cell pack, for 183 V nominal and 205 V peak) (or quarter way point for the 730 V pack). So one H and two non-H models might work out well if the auxiliary drain starts approaching 30 A more than rarely. [Edit: when the pack is in 365 V mode.]

Being able to tap the pack at many different points (e.g. at 1/3, 2/3, and 1/4) becomes problematical. It's non trivial to connect a cable with high voltage fuse part way through a battery box, and despite our 8 different boxes, we'd have to juggle a lot to arrange for all those options at the ends of battery boxes. Of course, the next model won't need the 1/4 tap, but we still might connect the centre tap to chassis, so that's still 3 taps.

The non-H models are not in stock till mid July, so as you say, perhaps for next time.
Last edited by coulomb on Fri, 18 Jun 2010, 16:03, edited 1 time in total.
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.

User avatar
coulomb
Site Admin
Posts: 3766
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Sat, 19 Jun 2010, 01:37

Oops!

Silly post deleted. No need to cut the (charger communications) connector open. See post on the next page.
Last edited by coulomb on Mon, 25 Apr 2011, 20:12, edited 1 time in total.
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.

7circle
Senior Member
Posts: 540
Joined: Thu, 04 Feb 2010, 02:18
Real Name: Ken
Location: Melbourne

Weber and Coulomb's MX-5

Post by 7circle » Sat, 19 Jun 2010, 03:10

coulomb wrote: ...
Being able to tap the pack at many different points (e.g. at 1/3, 2/3, and 1/4) becomes problematical. It's non trivial to connect a cable with high voltage fuse part way through a battery box, and despite our 8 different boxes, we'd have to juggle a lot to arrange for all those options at the ends of battery boxes. Of course, the next model won't need the 1/4 tap, but we still might connect the centre tap to chassis, so that's still 3 taps.


Hi Coulomb, some high voltage challenges.

I was concerned that the multiple supplies in output/parallel may not share the load evenly, and cause imbalance between the taps into the battery pack. One PSU will be at current limit and the other in constant voltage with only small power passing through it.

I was thinking if a voltage string could be arranged only connected to the pos and neg of the whole pack with out taps. For a quad PSU 250VAC(353VDC), perahps 4 x 100k in parallel with 320V Zeners to limit max input voltage to any of the PSU's.

Need some more thought, baut maybe you have been down this path and found it no good.

The other thought to limit imbalance is outputs say 4 x (3.3V to 4V) in series so match current.

User avatar
coulomb
Site Admin
Posts: 3766
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Sat, 19 Jun 2010, 03:58

7circle wrote: I was concerned that the multiple supplies in output/parallel may not share the load evenly, and cause imbalance between the taps into the battery pack.
Yes, this is always an issue. It will be necessary to adjust the voltage and current trimpots to ensure that the supplies share the load equally. It remains to be seen whether once set up they continue to share well, or drift out of balance.

Some suggest small resistors in series with the outputs of each charger, so small voltage imbalances in output voltage are offset by the voltage drop across the resistors. The one carrying the highest current will drop the most across its resistor, causing the other one to have to take up the load as its voltage threshold is crossed. I'm not a great fan of this, due to the wasted power, but would do it if necessary.
I was thinking if a voltage string could be arranged only connected to the pos and neg of the whole pack with out taps. For a quad PSU 250VAC(353VDC), perhaps 4 x 100k in parallel with 320V Zeners to limit max input voltage to any of the PSU's.
The resistors and zeners would have to be very high powered, and would waste a lot of power. Much more than the series resistors mentioned above.

Suppose you are dropping a 730 V pack to 340 V for a typical power supply, outputting 240 W (each), and 95% efficient. It will be drawing 240/.95/340 = 0.74 A; call it 0.8 A for some current into the zener. It's dropping 730-340 = 390 V, so that's 390 * .8 = 312 W. So the efficiency is 100% * 240 / (730*.8) = 41%.

Oh, but the zener has to draw a lot more current; if the minimum power drain is 40 W, then 200 W has to be burned up in the zener. That means the resistor has to be lower in value, and that will make it way worse. The resistor and zener diode are a crude linear voltage regulator; you're much better letting the switching power supply do the regulating.
The other thought to limit imbalance is outputs say 4 x (3.3V to 4V) in series so match current.
Yes, that's a good idea, for fixed voltage power supplies, though again they might need their voltage adjusted to make sure they share equally (but I agree it would be far less critical than fixed voltage supplies in parallel, where a small difference in output voltage will lead to a large difference in current sharing). These particular LED supplies are designed for constant current, so it would be tricky getting them to share the voltage evenly. If they don't share the voltage evenly, then they end up drawing different power and hence different pack current, and they are still unbalanced.

The other thing is that these LED supplies only go down to about 11.X V.
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.

7circle
Senior Member
Posts: 540
Joined: Thu, 04 Feb 2010, 02:18
Real Name: Ken
Location: Melbourne

Weber and Coulomb's MX-5

Post by 7circle » Sat, 19 Jun 2010, 07:05

coulomb wrote: ..
I was thinking if a voltage string could be arranged only connected to the pos and neg of the whole pack with out taps. For a quad PSU 250VAC(353VDC), perhaps 4 x 100k in parallel with 320V Zeners to limit max input voltage to any of the PSU's.
The resistors and zeners would have to be very high powered, and would waste a lot of power. Much more than the series resistors mentioned above.

Suppose you are dropping a 730 V pack to 340 V for a typical power supply, outputting 240 W (each), and 95% efficient. It will be drawing 240/.95/340 = 0.74 A; call it 0.8 A for some current into the zener. It's dropping 730-340 = 390 V, so that's 390 * .8 = 312 W. So the efficiency is 100% * 240 / (730*.8) = 41%.

Oh, but the zener has to draw a lot more current; if the minimum power drain is 40 W, then 200 W has to be burned up in the zener. That means the resistor has to be lower in value, and that will make it way worse. The resistor and zener diode are a crude linear voltage regulator; you're much better letting the switching power supply do the regulating.
The other thought to limit imbalance is outputs say 4 x (3.3V to 4V) in series so match current.
Yes, that's a good idea, for fixed voltage power supplies, though again they might need their voltage adjusted to make sure they share equally (but I agree it would be far less critical than fixed voltage supplies in parallel, where a small difference in output voltage will lead to a large difference in current sharing). These particular LED supplies are designed for constant current, so it would be tricky getting them to share the voltage evenly. If they don't share the voltage evenly, then they end up drawing different power and hence different pack current, and they are still unbalanced.

The other thing is that these LED supplies only go down to about 11.X V.


I thought you could get a cheaper PSU than these special LED Current drivers. I could imagine they maybe charging more per watt as LED lights aren't cheep.

Here is PIC of my idea.

Image
(N.B. NOT LIKELY TO WORK)

If one PSU input current drops the Zener protects it from overvoltage on input. Worst case ?? Image
If only one of the four PSU's was pushing 150W
needing 500mA at 300VDC
BUT the voltage on the input would collapse as the other zeners would be at 300V with lots of current Image
even with 220V Zeners allowing Vmax of 880VDC
so 660VDC is on three and one has say 800-660 = 140V
150W/140V = 1.07A so othe Zeners are about 660W together! Image
mmm with out proper current sharing it doesn't look good.

The PSU's may shut down when the input voltage descreases to a level say 100VDC

Just trying to think of way to not have to tap into pack.
Your dual PSU with center tap in battery bank isn't too complicated.

User avatar
weber
Site Admin
Posts: 2623
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sat, 19 Jun 2010, 15:14

Johny, Coluomb and 7Circle,

I'm afraid you're wasting your time. There is no way you can safely use SMPSs with input ratings of less than half the max pack voltage, and therefore no point in having more than 2 SMPSs with inputs in series and outputs in parallel.

You are forgetting input-to-output and input-to-chassis isolation voltage ratings. The requirement for these is not reduced by connecting more inputs in series (whether on battery taps or not).

I note that the supplies we've ordered only have their isolation resistance of 100 Mohm given at 500 Vdc.
http://www.meanwell.com/search/HLG-240H ... H-spec.pdf

Yes they have a withstand voltage of 1.88 kVac but "withstand" is usually for one minute, or less.

Even with two of them, to keep under the 500 Vdc from input to output we'd need to chassis-earth the centre of the battery, and the new NCOP 14 (foolishly) disallows that. How did you let that happen, acmotor? Image
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2623
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sat, 19 Jun 2010, 15:33

In our case, the current sharing resistors on the outputs could simply be that we wire each SMPS under the bonnet to the 12 V battery in the boot, with a separate cable of the minimum cross-sectional area required for current-carrying-capacity.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

Tritium_James
Senior Member
Posts: 683
Joined: Wed, 04 Mar 2009, 17:15
Real Name: James Kennedy
Contact:

Weber and Coulomb's MX-5

Post by Tritium_James » Sat, 19 Jun 2010, 16:47

Can't you leave off the current sharing resistors? You've got current-limited supplies (LED drivers) so I think all you need is a diode on the output of each supply (ie, diode-OR the outputs together). The highest output DC/DC will just go into it's constant current region before the others and be putting out full power. The voltage will then drop slightly and the next one will start taking up the slack. This might not work properly with supplies that foldback current limit, but you don't have that.

7circle
Senior Member
Posts: 540
Joined: Thu, 04 Feb 2010, 02:18
Real Name: Ken
Location: Melbourne

Weber and Coulomb's MX-5

Post by 7circle » Sat, 19 Jun 2010, 18:11

Did you check out the 300W High Voltage Rail Series from AMTEX
Input: 525 - 975VDC    Single: 12 ~ 130V options
http://www.amtex.com.au/power_pdf/power_6a.htm
eg http://www.amtex.com.au/power_pdf/HVI300R.pdf

I noticed on the Meanwell site HRT or DRT series is close with 780VDC but they only have 24V adustment wouldn't get you down to 14V
The new Meanwell WDR-120 has a 12V model from 780VDC .. so close.

EDIT : Some more for reference
Saw your comments below about environmental rating of enclosure.
Just found PULS being German could get expensive.
http://www.pulspower.com/pdf/ct5e121.pdf
Puls supplied by either (BUT 12V not listed - not popular)
http://www.snaptec.com.au/acdc/puls_pdfs/sl10e300.pdf
http://www.control-logic.com.au/shop/category/1109

IEC 62103 requires 820Vdc which makes me feel more comfortable that product will be available. Image
Last edited by 7circle on Sat, 19 Jun 2010, 09:10, edited 1 time in total.

User avatar
coulomb
Site Admin
Posts: 3766
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Sat, 19 Jun 2010, 18:13

Tritium_James wrote: Can't you leave off the current sharing resistors?

We'll just have to wait and see. Hopefully the supplies won't be in current limit all the time, so then they are in voltage limit, and if the two voltage limits aren't very close, one will take all the load, and the other sees "nothing to be done here; voltage is >= my limit".

[Edit: PS I don't see what good the diodes will do, as long as neither power supply goes into overvoltage protection (and that's at well over 16 V).]
Last edited by coulomb on Sat, 19 Jun 2010, 08:16, edited 1 time in total.
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.

User avatar
coulomb
Site Admin
Posts: 3766
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Sat, 19 Jun 2010, 18:27

7circle wrote: Did you check out the 300W High Voltage Rail Series from AMTEX
Input: 525 - 975VDC    Single: 12 ~ 130V options

No, thanks for the pointer. They look OK except that the data seems to say 24 V minimum (but the web page 12 V minimum). They seem to be designed for railway use, so I doubt that it would be economical or available in small quantities.

The 500 W unit would be good (for 730 V pack mode): no sharing issues, just one box. Not IP65 though.
Nissan Leaf 2012 with new battery May 2019.
5650 W solar, 2xPIP-4048MS inverters, 16 kWh battery.
1.4 kW solar with 1.2 kW Latronics inverter and FIT.
160 W solar, 2.5 kWh 24 V battery for lights.
Patching PIP-4048/5048 inverter-chargers.

Post Reply