Weber and Coulomb's MX-5

Post up a thread for your EV. Progress pics, description and assorted alliteration
Post Reply
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 » Thu, 18 Jun 2009, 16:19

Coulomb / Weber, that new lab supply I was telling you about has turned up here (up to 8V, up to 300A, fully programmable) so if you'd like to do some charge testing on these cells then let me know and we'll sort out a time.

User avatar
coulomb
Site Admin
Posts: 3780
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Sun, 21 Jun 2009, 02:39

Tritium_James wrote: Coulomb / Weber, that new lab supply I was telling you about has turned up here (up to 8V, up to 300A, fully programmable) so if you'd like to do some charge testing on these cells...

Well, we were certainly happy to take up TJ's generous offer.

The cell on the large cables:

Image

The front of the power supply:

Image

Spotted at the back of the desk:

Image
Looks like acmotor isn't the only one with a "dyno" set up. TJ also has a larger one... sorry no pictures of that one. I'm told the pancake motor on the right is actually more powerful than the large motor on the left.

We eventually hit the cell reasonably hard:
Image

We had the trusty thermometer on hand, but completely failed to record anything from it:
Image

Well, we figure that at these short bursts, measuring temperature on the side of the cell doesn't make much sense. As you can see from the photo, the cell did get a bit warm (and swelled a little too, but that may have been the morning's discharges); it reads 39.1°C in that photo, and got into the low 40's.

There was a PC connected to the power supply, setting maximum current and voltage. Every command was logged to a simple text file. I massaged the data a little, and ended up with this graph:

Image

Edit: "Corrected" voltage means subtracted 0.00083 times the current, to compensate for the resistance of the cables.

The first two bursts were of the order of 10-15 seconds each; the last burst was about 40 seconds. (Edit: actually the samples seem to be every 100 ms, so the last burst was more like 90 seconds.) The current is scaled such that a height of 3 represents a 3C charge; it happened to work out best that way to scale the data.

As you can see, 3C doesn't seem to cause an alarming voltage rise; occasional bursts like this of hard regen should be fine. Even 4C doesn't cause too much of a rise, depending on how you feel about charging to over 4.0 VPC.

Before this test, the cell was discharged to 2.5 V @ 3C, with about 5 short burts after it reached that level to 2.5 V again. So it would have been about 90% DOD, I would guess. However, we did the tests twice due to a glitch with the data file. The above data is therefore at roughly 80% DOD. In hindsight, we shouldn't have discharged it that low before the tests. At least it shows that there are at least some circumstances where a Thunder Sky cell can accept 4C for at least half a minute.

If the Sky Energy cells arrive before the power supply goes off to its job, TJ has kindly offered to repeat the test on one of those.

Thanks, TJ!   Image

Edit: the vertical lines in the graph are due to commands from the keyboard to change the current limit, etc. They are treated as "invalid data" and plot as zero. I didn't bother to weed them out; it nicely separates out the different tests.
Last edited by coulomb on Sun, 21 Jun 2009, 15:47, 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
weber
Site Admin
Posts: 2626
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sun, 21 Jun 2009, 03:41

The temperature was about 34°C when we first set up, and it may have been about 37° at the start of the recorded run.

We can summarise the above by saying the voltage rose by 0.2 V for every "C" of charge current at this temperature (taking 3.3 V as neutral). That's 5 mR internal resistance on charge, for this 40 Ah Thunder Sky cell. [Edit: Not 0.2 V. See below.]
Last edited by weber on Sun, 21 Jun 2009, 07:09, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
coulomb
Site Admin
Posts: 3780
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Sun, 21 Jun 2009, 03:48

Oh, that reminds me: the power supply was not set up with 4-wire Kelvin leads. The leads (as you can see) are quite fat, but there is still about a milliohm of resistance, so the voltage on the cell would have been a little lower than indicated. We measured 3.65 V on a multimeter when the power supply was at 3.75 V; the current at the time was probably 120 A (3C).

Edit: the values are corrected now. The old values are still there, since I'm not 100% sure about the value we had on the power supply when we saw 3.65 V on the multimeter.
Last edited by coulomb on Sun, 21 Jun 2009, 15:49, 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
weber
Site Admin
Posts: 2626
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sun, 21 Jun 2009, 04:12

Er, coulomb, how about you correct for that in the spreadsheet and update the graph above. It was definitely at 120 A where we saw 0.1 V drop in the cables.

So it's really only 0.17 V rise per "C" of charge or 4.2 mR (at about 20% SoC and 38°C).

If you look at our various discharge tests above you find 0.12 to 0.16 V drop per "C" of discharge or 3 to 4 mR (at about 80% SoC and 25°C to 30°C).

Earlier I suggested that the difference in internal resistance between charge and discharge was so great as to require a diode in the equivalent circuit. It appears I was wrong. On 3C and 6C discharges the internal temperature is probably much higher than the case temperature at the time of the measurement. This could explain the slightly lower internal resistance on discharge.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
acmotor
Senior Member
Posts: 3603
Joined: Thu, 26 Apr 2007, 03:30
Real Name: Tuarn
Location: Perth,Australia

Weber and Coulomb's MX-5

Post by acmotor » Sun, 21 Jun 2009, 05:28

Can you shift the temperature measurement to the actual battery terminal ? Also, go for a thermocoule or at least an LM35. (Give the steak thermometer back to the cook !) Image

I'm glad you clarified the volt drop in the leads. I was beginning to think that the charger had a 'calibrate' mode where you bolted the leads together, it supplied current, and measured the drop itself, then substracted it from the subsequent data. i.e. smarter than the average bear. Image

TJ, tell us more about your dyno. I can't read the numbers on the pic.
Image Is that 1.5kW ?
iMiEV MY12     110,230km in pure Electric and loving it !

User avatar
weber
Site Admin
Posts: 2626
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Sun, 21 Jun 2009, 17:06

acmotor wrote: Can you shift the temperature measurement to the actual battery terminal ? Also, go for a thermocouple or at least an LM35. (Give the steak thermometer back to the cook !) Image

No cook would be interested in that thermometer as it only goes to 150°C, as you can read in the photo. It is sufficiently accurate for our purposes, or at least it agrees well with our non-contact IR thermometers which we use to read the temperature of the terminals.

In fact, in a single-cell setup like this, the terminals do not give any better indication of the internal temperature of the cell than the probe thermometer on the side. Think about it. There is a massive copper heatsink to ambient bolted to both terminals.

In a battery pack it's different. In that case, you would get a good result on the short links between cells. We discussed all this with Tritium_James at the time.

Just testing two cells in series and thermally insulating the link between them would give a much more accurate picture of internal temperature (the average of the two).

Normally, we'd have a styrofoam patch over the probe where it is taped to the cell, but I guess we assumed Tritium's expensive setup would include logging of temperature (and kelvin connections) and so we didn't come well enough prepared.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
coulomb
Site Admin
Posts: 3780
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Mon, 22 Jun 2009, 01:57

weber wrote: Er, coulomb, how about you correct for that in the spreadsheet and update the graph above. It was definitely at 120 A where we saw 0.1 V drop in the cables.
Err, Weber, what a good idea. Done.

The little overshoots are interesting, e.g. at the beginning of the 4C charge. But I think it's just a sampling error:

3.3491
000.98
3.3491
000.98
PC 120 OK
PC 120 OK
3.3491
000.98
3.4386 <<-- new voltage
000.98 <<-- old current
3.4386
120.01
3.4386
120.01

Somehow the machine sampled the new voltage with the old current. Ah, there may be a student that can be blamed for this, so all is well. So just ignore those glitches; the cells don't have overshoots like that (or prescience, knowing that a big current it coming).
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.

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 » Mon, 22 Jun 2009, 02:04

I don't think they are sampled at the same time, which probably explains it. You can only talk to the power supply so fast, and I don't think the 'Request Current' and 'Request Voltage' commands are sent as a pair. I suspect they're sent at equal intervals. So voltage - 50ms delay - current - 50ms delay - voltage - 50ms delay - etc, etc

I didn't want to use the Kelvin connection option because of the need to do the connection 'live' (the power supply has large output capacitance, so needs to be at a matched voltage to the cell when connected). Trying to do up the main bolt lugs onto the battery, at the same time as keeping a 2nd kelvin connection on the lug, would have been difficult. If the connection between main lug and sense connection was lost momentarily during the connection process, then the voltage would shift, and not match = fat scary precharge spark...

I guess what I should have done is crimped the sense lead into the main lug when I made the cables.

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 » Mon, 22 Jun 2009, 02:08

Hmmm, looking at that data more closely, the offset request times on the serial port don't explain it. Weird.

User avatar
weber
Site Admin
Posts: 2626
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Mon, 22 Jun 2009, 03:12

Tritium_James wrote:I didn't want to use the Kelvin connection option because of the need to do the connection 'live' (the power supply has large output capacitance, so needs to be at a matched voltage to the cell when connected). Trying to do up the main bolt lugs onto the battery, at the same time as keeping a 2nd kelvin connection on the lug, would have been difficult. If the connection between main lug and sense connection was lost momentarily during the connection process, then the voltage would shift, and not match = fat scary precharge spark...

I guess what I should have done is crimped the sense lead into the main lug when I made the cables.

You can do the kelvin connections after tightening the terminals, by using alligator clips.

But what you really should do is eliminate the need to do live manual connections. If you don't someone is eventually going to get it wrong and blow away half a terminal and maybe lose an eye.

You just need to get your nearest Haymans or Ideal Electrical to order you in a 4-pole 63 A D-curve circuit breaker (e.g. Hager NDN463A) and chop one of your 50 mm^2 leads in half, strip both ends back 70 mm, split the strands into four equal bundles, and run the 4 poles of the circuit breaker in parallel. This should take 300 A for a many minutes without tripping.

You should still aim to have the output voltage the same as the cell voltage before closing the breaker, but at least if you forget, it's not a disaster, and you can do the kelvin connections without stress.
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 » Mon, 22 Jun 2009, 03:49

Oh, I'd never use the test method we used for anything except a one-off test. Anything else would have a precharge controller & contactors.

For what we did, with matched voltages before connection, if we were using a Kelvin connection, then it needed to be connected to the bolt lugs before they went onto the battery, or else the voltage on the lugs wouldn't be right (no feedback via the 4-wire sense to the power supply).


User avatar
acmotor
Senior Member
Posts: 3603
Joined: Thu, 26 Apr 2007, 03:30
Real Name: Tuarn
Location: Perth,Australia

Weber and Coulomb's MX-5

Post by acmotor » Mon, 22 Jun 2009, 07:42

weber wrote: ..........
No cook would be interested in that thermometer as it only goes to 150°C, as you can read in the photo..........


So you don't do much cooking then ? Image

Short of burnt toffee, not much else goes above 100° and meat, e.g. steak is cooked at 71°C at its centre...... I digress. (oven air temperature is another topic)

The response time through the cell case was more my concern.
The terminals are also often a source of heat due to current and represent the path to ambient for much of the cell heat. They would be my first point of measurement. Image
iMiEV MY12     110,230km in pure Electric and loving it !

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 » Mon, 22 Jun 2009, 13:54

We discussed that during the test, but for this particular setup the terminals would not have been an accurate temperature - we were only doing short tests (15 seconds or so) and the cables weighed almost as much as the cell itself, ie there's a lot of copper in them. So any thermal measurements on the terminals would be misleading, a lot of heat can be dumped into those cables before you see much temperature change.

Squiggles
Senior Member
Posts: 742
Joined: Wed, 22 Apr 2009, 03:19
Real Name: Neil
Location: Newcastle NSW

Weber and Coulomb's MX-5

Post by Squiggles » Mon, 22 Jun 2009, 14:38

So, what your saying is that the cells in your battery should be connected with over sized copper busbar that acts as a heat sink. Seems like a simple enough solution.

User avatar
weber
Site Admin
Posts: 2626
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Mon, 22 Jun 2009, 15:12

Tritium_James wrote:For what we did, with matched voltages before connection, if we were using a Kelvin connection, then it needed to be connected to the bolt lugs before they went onto the battery, or else the voltage on the lugs wouldn't be right (no feedback via the 4-wire sense to the power supply).

Oh now I understand. Sorry. I imagined it was programmable as to whether it used the separate sense wires or not. So I guess the sense wire terminals were just jumpered to the output terminals on the back of the unit?
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 » Mon, 22 Jun 2009, 15:26

Squiggles wrote: So, what your saying is that the cells in your battery should be connected with over sized copper busbar that acts as a heat sink. Seems like a simple enough solution.
I know you're only half serious, but this has triggered me to check out something which I've wondered for a while.

Copper is one of the best electrical and heat conductors.
Aluminium is up there, but is way lighter. Which is more suitable for EV use?

Density: Copper 9kg/L Ali 2.6kg/L
Electrical Conductivity: Copper 58 Siemens Ali 38 Siemens
Thermal Conductivity: Copper 380 Ali 220
Price: Copper US$2.24/lb Ali US$0.74/lb

So:

If you are concerned about volume, Copper is the clear winner. 50% better electrical, 70% better thermal.
If you are concerned about weight, Ali is the winner
If you are concerned about cost, Ali is the winner.

There are a few other things to think about though.

cheers,
Woody
Planned EV: '63 Cortina using AC and LiFePO4 Battery Pack

User avatar
coulomb
Site Admin
Posts: 3780
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by coulomb » Mon, 22 Jun 2009, 15:28

Tritium_James wrote: Hmmm, looking at that data more closely, the offset request times on the serial port don't explain it. Weird.

It sort of does if the current is sampled first, then the voltage. But then why is the voltage sent first? Maybe they are cached.
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 » Mon, 22 Jun 2009, 15:40

If Kelvin Connection is the same as the remote sense system we used to use, then we got around some of the problems by having 47 Ohm resistors in the power supply from remote sense to output in case the remote sense wires were disconnected at the load.

User avatar
Thalass
Senior Member
Posts: 741
Joined: Sun, 12 Aug 2007, 07:29
Real Name: Ben Rypstra
Location: Perth, WA, AU

Weber and Coulomb's MX-5

Post by Thalass » Mon, 22 Jun 2009, 16:36

Aluminium is too brittle, I think. Many aircraft have huge aluminium feeder cables from the generators, but then the vibration is different in a turbine aircraft to any road vehicle. I could be wrong.
I'll drive an electric vehicle one day.

User avatar
Electrocycle
Senior Member
Posts: 985
Joined: Sun, 19 Oct 2008, 20:23
Real Name: Andrew
Location: Sydney
MSN: dumhed@dumhed.com
Contact:

Weber and Coulomb's MX-5

Post by Electrocycle » Mon, 22 Jun 2009, 17:24

I used aluminium for the busbars in the bike.
It was a lot cheaper, easier to get, and did the job very well.
Any connections subject to vibration should be adequately damped to prevent fatigue over time.
The Engine Whisperer - fixer of things

User avatar
weber
Site Admin
Posts: 2626
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber and Coulomb's MX-5

Post by weber » Mon, 22 Jun 2009, 19:11

Johny wrote: If Kelvin Connection is the same as the remote sense system we used to use, then we got around some of the problems by having 47 Ohm resistors in the power supply from remote sense to output in case the remote sense wires were disconnected at the load.

Yes. That's all that's needed. Replace the jumpers at the back of the power supply with resistors. Only I think they should be much higher resistance than 47 ohm. I was thinking 1k0. Of course it depends on the internal resistance of the sense input.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
Simon
Senior Member
Posts: 431
Joined: Sun, 19 Aug 2007, 19:38
Real Name: Simon
Location: Perth WA
Contact:

Weber and Coulomb's MX-5

Post by Simon » Mon, 22 Jun 2009, 20:50

Regarding aluminium busbars I found this - The Great Bus-Bar Challenge
I wonder how copper bus bars dipped in solder would compare?

User avatar
Electrocycle
Senior Member
Posts: 985
Joined: Sun, 19 Oct 2008, 20:23
Real Name: Andrew
Location: Sydney
MSN: dumhed@dumhed.com
Contact:

Weber and Coulomb's MX-5

Post by Electrocycle » Mon, 22 Jun 2009, 21:22

I think that shows that the connection between the terminal and the bus bar is probably the most important part (assuming you have a semi decent busbar)

I wonder how much difference there would have been using bars of half the thickness.
The Engine Whisperer - fixer of things

User avatar
acmotor
Senior Member
Posts: 3603
Joined: Thu, 26 Apr 2007, 03:30
Real Name: Tuarn
Location: Perth,Australia

Weber and Coulomb's MX-5

Post by acmotor » Tue, 23 Jun 2009, 06:04

This is part of what I was fishing at regarding the BMS board copper and the inter battery leads.
I was under the impression ? that weber was saying that the heat (generated by) coming out of the terminals and conducted away was a big factor.
In this case TJ, if the copper wires were in any way non representative of a typical battery pack arrangement yet have a big part to play, then should this representativelessness be addressed post hast ? Image

On that point, is there an optimum inter-cell wire size based on thermal properties ?
iMiEV MY12     110,230km in pure Electric and loving it !

Post Reply