Li cell discharge tests

How do you store and manage your electricity?
User avatar
coulomb
Site Admin
Posts: 3835
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Li cell discharge tests

Post by coulomb »

Nevilleh wrote:
weber wrote:

1. A voltage-sag or internal resistance figure is utterly meaningless if you don't know the internal temperature of the cell at the time. The figures can almost halve with a temperature increase of only about 10 degrees Celsius....


Self-contradictory! How can it be "utterly meaningless" in one sentence and "almost halve" in the next?
Ah, you need path-sensitive analysis to parse Weber's sentence. There is a predicate "if you don't know the internal temperature of the cell at the time". It is possible for this predicate to be false, i.e. you know (something about) the internal temperature. So then the predicated statement "A voltage-sag or internal resistance figure is utterly meaningless" doesn't apply in paths where the predicate is false. So it's not always a contradiction.

To summarise the above (sorry, couldn't help injecting some terminology from work    Image ), if you measure the temperature, voltage-sag and internal resistance measurements are no longer meaningless.
...there is NOT a lot of rapid heating in the cell at 3C.

Neville, I suppose it depends on what you mean by a lot. We've found temperature rises of about 30°C after full 3C tests; the temperature only stops rising some 10-20 minutes after the test has ended. (With the fan on, it rises maybe 20°C and starts falling perhaps a little before the 10 minute mark.) A 20°C rise from about 20°C only brings the cell to blood temperature, and a 30°C rise brings it to warm but not hot. But the voltage sag seems to be very dependent on cell temperature, so these moderate temperature rises are important.

See for example this thread: Current Limiting on Curtis and TS cells thread (probably posted to the wrong forum, sorry). It's with a big fan blowing on the cell, but it compares with previous data that was not using the fan.
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.

antiscab
Senior Member
Posts: 2499
Joined: Mon, 26 Nov 2007, 05:39
Real Name: Matthew Lacey
Location: Perth, WA

Li cell discharge tests

Post by antiscab »

Nevilleh wrote:
I think I read somewhere on here that a guy with an operating vehicle found that his BMS gave a low voltage alarm on cold mornings. He also said the alarm point didn't change very quickly as he used the vehicle, implying that the cells were not warming up all that much with quite heavy current drains. He would have been sucking at least 100A to even drive the thing and that didn't seem to create a great deal of heat in the battery.


that was Peter C IIRC.
100A isnt alot for 90AH cells.
Id suggest this is why he isnt seeing any increase in his commute.

i do know of a US bloke who pulls 500A on his 320v 90AH pack to get it to warm up.
sags to 2v initially, but after a bit of warming, internal resistances drops.

on my emax, pulling 250A for a few secs, then 50-60A continuous out of my old 40AH cells warmed them up a bit.

Matt

EDIT: added quote due to "leap frog"
Last edited by antiscab on Fri, 17 Jul 2009, 11:21, edited 1 time in total.
Matt
2017 Renault zoe - 25'000km
2007 vectrix - 156'000km
1998 prius - needs Batt
1999 Prius - needs batt
2000 prius - has 200 x headway 38120 cells

Nevilleh
Senior Member
Posts: 773
Joined: Thu, 15 Jan 2009, 18:09
Real Name: Neville Harlick
Location: Tauranga NZ

Li cell discharge tests

Post by Nevilleh »

All this stuff is about to be made redundant!

http://www.elektor.com/news/lithium-air ... paign=news

User avatar
weber
Site Admin
Posts: 2643
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Li cell discharge tests

Post by weber »

Nevilleh wrote: It's all interesting stuff, isn't it?
Yes. And its kind of frightening that this sort of data isn't available from the cell manufacturers, and such tests are left to little old us. At best they quote us utter^H^H^H^H^Hfairly meaningless 1 kHz internal resistance measurements.
I am trying to simulate the load an EV would put on the thing by running it at 3C and then "lifting off" monetarily with the cells at "garage" temperature, ie 17 deg C. Now I stand by my internal resistance measurement of 1.67 milliohms and the amount of heat generated in the cell then is 120 x 120 x 1.67 x 10^-3 = 24 watts. Don't know what the Specific Heat of the internal chemicals is, but if it were water and the weight is say 1.5 kg, then 24 watts applied for 23 minutes will raise its temperature by 5.26 deg C. Not a huge heat rise, is it?
I'm not doubting that the internal resistance is as you say it is, at the time that you measure it. I'm just suggesting the internal resistance was a lot higher at the beginning of the 3C discharge, in which case the self-heating of the cell was a lot more rapid at the start, and so by the time you measure the resistance the internal temperature may be a lot higher and hence resistance may be a lot lower than it was at the start when the cell internals were at the ambient 17°C.

We have two different measurements of the internal resistance of a new 40 Ah Thunder Sky, both claiming to be at around 17°C. They differ by almost a factor of 4. We need to get scientific and figure out why this is. Maybe it is because TS quality control is reeeally baaaad, but that seems unlikely. I've suggested a possible reason which should be subjected to further experiment. Do you have some means to measure at least the external cell temperature? If not, you might try repeating the full 3C discharge with say three 5 second "liftoffs". Say at 1 minute, 5 minutes and 10 minutes to see if you get roughly the same value each time.
Solid Li and Ethanol (might not be relevant, but we can always drink it Image )both have specific heats of about half that of water, so if the thing is built out of that sort of stuff, the temp rise would be about double, or say 10 - 11 deg C. Enough to have quite an effect, according to Coulomb.
I understand the lithium is so light that its percentage by weight is so low it can probably be ignored. My best guess so far is that the temp rise would be about 5 times that of an equal mass of water for the same energy input.
But it does drop the internal res, so that can only be good.
Yes that is good.
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

Li cell discharge tests

Post by acmotor »

Nevilleh wrote: ..........But a thermocouple would be pointless as it could only measure the cell temp on the outside and the massive heating is on the inside, and heat sink-ed away by the huge terminals before it can be measured!


Nevilleh, I mean to unscrew the vent cap (at least it can be done on TS)and feed a thermcouple inside the cell to be immediately above (in contact with) the "bag" inside. This would give much improved (but not totally) temperature monitoring. A bit like the winding temperture sensor I put in the little AC motor.
viewtopic.php?p=14860&t=1237#p14860

That type of measurement combined with the TJ standard temperature and EV load profile (Lp.tj.ev) would be good.

edit: quoted RIGHT person Image
weber wrote: See what trouble you've got me into now, ACmotor.


Oz trait. If you like someone, flame them. If you hate them, flame them. Hang on. That's the same thing ? Oh well.
I think you know by now I respect your knowledge. I just hate it ! Such is the price you have to pay for being right all the time Image
Last edited by acmotor on Fri, 17 Jul 2009, 14:32, edited 1 time in total.
iMiEV MY12     110,230km in pure Electric and loving it !

Nevilleh
Senior Member
Posts: 773
Joined: Thu, 15 Jan 2009, 18:09
Real Name: Neville Harlick
Location: Tauranga NZ

Li cell discharge tests

Post by Nevilleh »

weber wrote:
I understand the lithium is so light that its percentage by weight is so low it can probably be ignored. My best guess so far is that the temp rise would be about 5 times that of an equal mass of water for the same energy input.
But it does drop the internal res, so that can only be good.
Yes that is good.


Looking at a table of specific heats for many common substances (Wikipedia) I don't see a lot lower than about 1.5 (graphite) with water at 4.2 (that's Joules/gm/deg). That's .36 of water, so I really doubt the stuff in a LiFePO4 cell is less than that. Which is why I don't think the temp rise is going to be as high as you are speculating. I know it has Li and C and Fe and P and O!
Graphite is 1.534 and there is probably quite a bit of that - carbon fibre - as that is apparently the secret to making these things.
Li is about 1.95, so I reckon the whole thing has to be between 1.5 and 2, so the temp rise is likely to be about 2 to 3 times that of water, or around 10 to 15 degrees!
Speculation - the food of the uninitiated!!

Just what IS in a Li cell?

Nevilleh
Senior Member
Posts: 773
Joined: Thu, 15 Jan 2009, 18:09
Real Name: Neville Harlick
Location: Tauranga NZ

Li cell discharge tests

Post by Nevilleh »

By the way:
Go the All Blacks!

User avatar
weber
Site Admin
Posts: 2643
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Li cell discharge tests

Post by weber »

Sorry Neville. I can be a bit "all rigour and no tact" at times. It's great to have someone else doing these sorts of tests to compare results with, particularly when we only have a sample of one TS cell.
Johny wrote: Hi guys. From my understanding Nevilleh was starting with a cell at near ambient then testing it to ascertain capacity. He used an assumption that internal heating would be minimal when performing short high-current bursts. Useful testing.
My understanding is that Neville was doing (almost) continuous high current to ascertain capacity, with short "lift offs" of zero current to ascertain voltage droop and hence internal resistance.

Perhaps you would clarify, Neville?
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2643
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Li cell discharge tests

Post by weber »

Nevilleh wrote:Looking at a table of specific heats for many common substances (Wikipedia) I don't see a lot lower than about 1.5 (graphite) with water at 4.2 (that's Joules/gm/deg). That's .36 of water, so I really doubt the stuff in a LiFePO4 cell is less than that. Which is why I don't think the temp rise is going to be as high as you are speculating.
But Neville, we're not speculating about the temperature rise at 3C discharge. We've measured it. Twice. Please see the link that Coulomb posted above. Or rather, here it is again to save you the trouble.
viewtopic.php?p=14135&t=1178#p14135
I know it has Li and C and Fe and P and O!
Graphite is 1.534 and there is probably quite a bit of that - carbon fibre - as that is apparently the secret to making these things.
Li is about 1.95, so I reckon the whole thing has to be between 1.5 and 2, so the temp rise is likely to be about 2 to 3 times that of water, or around 10 to 15 degrees!
Speculation - the food of the uninitiated!!

Just what IS in a Li cell?

There's quite a lot of info on that in the Material Safety Data Sheet at the end of the Thunder Sky Instruction Manual I dowloaded some months back. I couldn't find it on their site now, so I tried to upload what I've got, to the forum, but it's too big. You may have better luck finding it at the TS site.
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:

Li cell discharge tests

Post by Tritium_James »

Last edited by Tritium_James on Fri, 17 Jul 2009, 18:22, edited 1 time in total.

User avatar
coulomb
Site Admin
Posts: 3835
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Li cell discharge tests

Post by coulomb »

Info like this:

Image

I can't find now a post of someone who pulled a cylindrical cell apart. He seemed to think it was mostly aluminium and copper, plus a little graphite and negligible amounts of lithium. The above analysis seems to contradict that idea somewhat.

The above may be fractions of the various elements (but then why separate C from graphite, and Li from LiX?), since I'd expect the proportions my volume of aliminium and copper to be about the same.

So how about that: the element with the largest proportion in those expensive cells is plain old iron!

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

Li cell discharge tests

Post by Electrocycle »

I'd say the aluminium and copper plate backing is not counted as active ingredients.
The Engine Whisperer - fixer of things

Nevilleh
Senior Member
Posts: 773
Joined: Thu, 15 Jan 2009, 18:09
Real Name: Neville Harlick
Location: Tauranga NZ

Li cell discharge tests

Post by Nevilleh »

I'll try and cover some of the points raised:
I have 16 TS LFP40AH cells that have been used in my EVT 4000 scooter for quite a while, done 632 kms in fact and they have been charged 14 times in all. I replaced them with the original PbA pack, so I can pick one of 16 at random to "play" with.

I wanted to do a 3C continuous discharge test to see how they would perform and to compare with the Sky Energy cells that are due here about now. I hadn't seen a 3C test on these forums.

In the scooter, they were subjected to discharges of about 85A under full acceleration, according to the clamp ammeter and they never had any trouble doing that, even on cold mornings ie outside temp about 5 deg. (It seldom gets colder than that here).

I thought to shut the load off every minute while testing just to see what the no load voltage was, but that was more an afterthought than a planned event, so I didn't record the voltages, just noticed that after 3 minutes (when I started doing it) that the voltage rose to about 3.4 then steadied at 3.2 for the last 10 minutes or so. I didn't measure the case temp, but just put my hand on it occasionally in case anything catastrophic was happening!
After the test was finished, I put my hand on the cell again and it wasn't appreciably warm so I didn't think a lot of heat was generated.

I have now bought a temperature probe for my DMM, so I'll do the test again and apparently the TS cap can be unscrewed quite easily, so I'll put the probe inside the cell and see what happens. I checked the link given by Coulomb about the temp test and it surely was a lot higher than I saw - my hand isn't very well calibrated but I think I could tell if the cell got up to 40 deg!

I'll pick another cell and see if it does the same numbers.

For the test, I charged the cell by itself at 3.65v/5A from my bench power supply.

User avatar
weber
Site Admin
Posts: 2643
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Li cell discharge tests

Post by weber »

acmotor wrote: weber, can you short your load and see if the CB trips ?   Image
Would a fuse blow in this application ?

While testing our collection of 40 Ah cells at Tritium this morning I mentioned your concern to James. (BTW, Coulomb didn't like the idea of deliberately shorting one of our cells). James suggested we use his power supply to put its full 300 amps through our circuit breaker and see what happens. We didn't have the actual 2-pole with us, but I happened to have a single pole of the same kind in the van. A very ordinary Hager NT163C (63 A C-curve 6 kA). So I made up some short 16 mm^2 leads for it and we measured the voltage across it, including the leads and screw-clamp connections, while putting various programmed currents through it. And we timed how long it took to trip.

Here's the setup.

Image

The result: It tripped after about 6 seconds at 300 amps (a 480% overload) and the max voltage across it, before it tripped, was 0.6 V. That's 2 mR.

In our discharge test setup we have the two poles of a Hager NT263C in parallel. So that's 1 mR. Our Thunder Sky 40 Ah cell has an internal resistance of around 4 mR at 17°C. Lets suppose another 1 mR in the other cables and connectors, for a total of 6 mR. Lets use 3 V for the internal voltage source (a near flat cell) and so we can expect about 3 V/6 mR = 500 A if we shorted our load under those conditions. That's a 400% overload so it will take a bit longer to trip. Maybe 10 seconds.

Here's a time current curve.
http://www.accontrols.com/Typical%20Cir ... 0Curve.pdf

What this means is that the cables need to be sized to withstand 500 amps for 10 seconds. For copper with 90°C insulation the required cross-section in mm^2 is given by I/100*sqrt(t) = 5 * sqrt(10) = 16 mm^2. We have two 16 mm^2 cables in parallel so we'd be OK down to even lower cell temperatures (higher cell resistances).

I note that exactly the same considerations apply to the wiring and short-circuit protection used in the vehicle.

Warning: 240 V AC circuit breakers typically only have a 60 V DC rating so they are of no use for breaking complete EV battery packs.

I looked up the data for some size-00 NHP fuses, class gG (general purpose), of the appropriate rating, and it turns out they have similar resistance and similar time-current characteristics in the region of interest. They don't give resistances, but they give power dissipation of the fuse at its rated current, from which you can calculate resistance.
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

Li cell discharge tests

Post by acmotor »

That's good data !! (the actual test that is)

It probably also says that a thermal overload CB is not good for S/C protection since the I^2t is larger than say a semiconductor rated fuse ? Probably fine for your cell tests though.

I would say though that If I had used your extrapolations that you would have gone blow torch. Image

Now, have you built up enough courage to test the CB in you battery test circuit ? Image
iMiEV MY12     110,230km in pure Electric and loving it !

User avatar
weber
Site Admin
Posts: 2643
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Li cell discharge tests

Post by weber »

acmotor wrote:It probably also says that a thermal overload CB is not good for S/C protection since the I^2t is larger than say a semiconductor rated fuse ?
Yeah. Semi-rated fuse probably much better, but CB convenient as it does switching too. Hence I suggested fuse when contactor already used for switching. Can you find us some time-current curves for semiconductor-rated fuses?
I would say though that If I had used your extrapolations that you would have gone blow torch. Image
Yeah. But only an LPG torch, not oxy. Image Actually, if you had overengineered the cable by a factor of 2 I wouldn't have bothered.
Now, have you built up enough courage to test the CB in you battery test circuit ? Image

The issue isn't one of courage, but whether we've done enough other tests to the poor TS cell, such that we are willing to go outside the manufacturer's limits like that and risk significant permanent loss of capacity. Being pulled down to less than a volt can't be good, even if it is only for 10 seconds.

What would it prove that would be of use in a EV, where you shouldn't be using AC circuit breakers anyway?
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

Li cell discharge tests

Post by acmotor »

? but if you have used the CB to protect something in your test circuit, are you now saying it is not protection against potential damage ? Agreed, 15C is a lot on a TS.

What the test may prove is whether or not the CB served any purpose (other than on/off) and justified your suggestion that Nevilleh was living dangerously without a fuse in his circuit. Maybe you needed one ?

Then again, it might just trip. Job done. It was just a question.
Remains untested. Image
iMiEV MY12     110,230km in pure Electric and loving it !

User avatar
weber
Site Admin
Posts: 2643
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Li cell discharge tests

Post by weber »

acmotor wrote: ? but if you have used the CB to protect something in your test circuit, are you now saying it is not protection against potential damage ?
It's protection from melting insulation and nasty gasses venting from cell. Of course it's not protection against loss of cell capacity.
What the test may prove is whether or not the CB served any purpose (other than on/off) and justified your suggestion that Nevilleh was living dangerously without a fuse in his circuit. Maybe you needed one ?
Those are two different things.

If you want to prove we didn't need a fuse, we just have to short out our load with the CB in place.

If you want to prove Neville didn't need a fuse (or any short circuit protection), then we have to short out the load WITH NO PROTECTION IN PLACE.

I'll do the first one after you do the second one (leaving it shorted for at least 4 minutes). You have some 40 Ah TS cells don't you? You'd better video it. Image
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2643
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Li cell discharge tests

Post by weber »

me wrote:We have two different measurements of the internal resistance of a new 40 Ah Thunder Sky, both claiming to be at around 17°C. They differ by almost a factor of 4.

I was wrong. Our figure was "only" 2.6 times greater than Neville's.

Tritium_James, Coulomb and I did some measurements of voltage lift (opposite of sag?) at high rates of charge today. Although that's technically off-topic, I thought I'd give a rough summary here. Coulomb will give more detailed info later, probably in another thread.

We tested four nominally 40 Ah LiFePO4 cells. One Thundersky, two Sky Energy SE cells and one specially-made high-rate cell from China HiPower. The two SE cells chosen were the best and the second worst (one of two supposed "runts" in the litter of 212), in terms of the 1 kHz internal resistance data supplied with the cells by Sky Energy.

The cells all started at 17°C and some minutes after the tests they measured between 20°C and 22°C. The higher internal resistance ones achieving the higher temperatures as you would expect. These temperatures were measured by our usual method of a probe in the middle of the broad face, covered by polystyrene.

All cells were 40% to 60% charged and were subjected to successive charge rates of 1C, 2C, 3C and 4C, and where possible 5C and 6C. i.e. multiples of 40 A up to 160 A and up to 240 A where possible. Each charge rate was applied for about 5 seconds with approximately a 5 second rest between bursts.

All cells showed a decrease in voltage lift (i.e. decrease in internal resistance) with successive bursts. This is presumably due to self heating. Rates of charge this high are apparently not endothermic. However it might be that internal resistance is simply nonlinear with current, and actually gets better with higher current irrespective of temperature (although this effect may itself be due to very localised heating).

The voltage lift per "C" of charge was:

Thunder Sky                     250 mV down to 190 mV

Sky Energy SE "best"            160 mV down to 110 mV

Sky Energy SE "runt"            140 mV down to 100 mV

China HiPower (high-rate type) 100 mV down to 70 mV

Divide by "C" = 40 A to obtain internal resistance.

Note that the supposed runt SE cell was actually better than the supposed best SE cell. So the supplied 1 kHz internal resistance figures ("best" 0.53 mR, "runt" 0.79 mR at 20°C +- 5°C) don't appear to be very useful. They are also less than 1/4 of the measured values.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

Nevilleh
Senior Member
Posts: 773
Joined: Thu, 15 Jan 2009, 18:09
Real Name: Neville Harlick
Location: Tauranga NZ

Li cell discharge tests

Post by Nevilleh »

I didn't use a fuse in my test setup as I could not see the need to do so. All I had to do to switch the load off was turn off the 12v to the breaker. And I certainly wasn't going to leave it unattended! From what I read above, my reaction time was going to be a whole lot faster than your CB.

Your charging figures (above) certainly show how completely useless the SE "internal resistance" figures are.
Last edited by Nevilleh on Sun, 19 Jul 2009, 03:40, edited 1 time in total.

Nevilleh
Senior Member
Posts: 773
Joined: Thu, 15 Jan 2009, 18:09
Real Name: Neville Harlick
Location: Tauranga NZ

Li cell discharge tests

Post by Nevilleh »

coulomb wrote: Info like this:

So how about that: the element with the largest proportion in those expensive cells is plain old iron!


A bit of reading tells me that the cathode is LiFePO4, ie Lithium iron phosphate, the anode is usually carbon (deposited on an Al conductor) and the electrolyte is LiClO4. The anode is sometime Li metal (depends on the manufacturer, I suppose). Li migrates to the anode and one reason why these cells are fairly stable and safe is that Li ions can be removed from the LiFePO4 without altering its structure (physically).
Couldn't find any info on the specific heat capacities of those compounds yet.
The table of elements above simply reflects the atomic weights in the make-up of the compounds.
The conductivity (internal res) is very dependent on how easily the Li ions can travel through the electrolyte. Apparently they move through "tunnels" created by the carbon particles in the electrolyte and one researcher has found a way to vastly increase the size and number of these tunnels, so we can expect maybe a 10-fold drop in cell resistance when this filters through to the manufacturers.

User avatar
weber
Site Admin
Posts: 2643
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Li cell discharge tests

Post by weber »

Nevilleh wrote: I didn't use a fuse in my test setup as I could not see the need to do so. All I had to do to switch the load off was turn off the 12v to the breaker. And I certainly wasn't going to leave it unattended! From what I read above, my reaction time was going to be a whole lot faster than your CB.
That's fair enough. So long as you never did get distracted and leave it unattended, or so long as you don't have kids around. Image And even if you're there, you might not realise there is short in progress for a few seconds.

But as I said, we're still not protected against shorts at the cell terminals, except by insulating tape (when we remember). And no point in spending a lot of money on a test setup that will only be used for a short time.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

Nevilleh
Senior Member
Posts: 773
Joined: Thu, 15 Jan 2009, 18:09
Real Name: Neville Harlick
Location: Tauranga NZ

Li cell discharge tests

Post by Nevilleh »

I unscrewed the cap from my test cell to discover a spring sitting on top of a metal disk, presumably a pressure vent valve, so I gave up on the idea of putting my thermometer inside and taped it to the outside of the cell instead.
Image

So I ran the test again and recorded a bit more data this time.
The columns below got a bit out of whack when I cut and pasted, but the headings from left to right are Time (mins) Volts under load, Volts no load, Load current, cell temp and the internal resistance in milliohms
This time I switched the load off for 5 seconds at every minute and recorded the voltage at the end of that 5 secs. That gave time for the DMM to stabilise properly. I didn't bother adding cold water or stirring the load water so the current is just what it was. The calculated internal resistance was done by taking VNL - VL divided by the current all times 1000.

This agrees fairly well with my first test and indicates an internal resistance of less than 2 mR.

LFP40AH Discharge Test - 19/07/2009                         
Time     Volts L     Volts NL     Amps     Temp     Res mR
0          3.582          16.5     
1     3.013     3.25     121     16.6     1.96
2     2.993     3.23     120     16.9     1.98
3     2.987     3.22     119.8     17.2     1.94
4     2.983     3.2     119     17.8     1.82
5     2.981     3.2     118.3     18.3     1.85
6     2.978     3.2     118.1     19     1.88
7     2.997     3.2     117.8     19.7     1.72
8     2.972     3.19     117.6     20.4     1.85
9     2.97     3.18     117.4     21.1     1.79
10     2.967     3.18     116.8     21.8     1.82
11     2.964     3.18     116.5     22.4     1.85
12     2.957     3.17     116.3     23     1.83
13     2.952     3.16     115.7     23.7     1.80
14     2.947     3.15     115.1     24.5     1.76
15     2.939     3.15     114.5     25     1.84
16     2.933     3.14     114     25.6     1.82
17     2.927     3.13     113.1     26.2     1.79
18     2.91     3.12     112.1     26.8     1.87
19     2.9     3.11     111.5     27.4     1.88
20     2.876     3.09     111     28     1.93
21     2.85     3.07     109.5     28.6     2.01
22     2.801     3.02     107.3     29.2     2.04
23     2.736     2.95     105.1     29.8     2.04
24     2.61     2.83     104.8     30.4     2.10
25          2.86          31.1     
26          2.94          31.7     
27          2.97          32.1     
28          2.99          32.4     
29          2.99          32.3     
30          3          32.3     

Image

The load was switched off permanently after 24 mins.

Note the shape of the temp curve. The delay before the temp starts rising and the fairly modest overshoot after the load is switched off indicates that the thermal resistance between the heat source and the outside of the case is relatively low.

I don't see a big change in internal resistance over the temperature range of the test.

To me this shows that I can draw 3C out of one of these cells very reliably at any state of charge above just about flat and at any temperature in the range of 16.6 to 32 deg and not expect a Vdrop across the cell's internal resistance of more than 240 mV. If I parallel 3 of them, as is my intention, I can draw 360A.

Edit: Sorry I forgot to label the RH Y-axis as temp in deg C. And the VL value at 7 mins is wrong, should be 2.977. I fixed it in my spreadsheet and it takes the wee bump out of the curve.


Last edited by Nevilleh on Sun, 19 Jul 2009, 10:56, edited 1 time in total.

Nevilleh
Senior Member
Posts: 773
Joined: Thu, 15 Jan 2009, 18:09
Real Name: Neville Harlick
Location: Tauranga NZ

Li cell discharge tests

Post by Nevilleh »

Further to all that, my spreadsheet tells me the average current was 114.7 A and the average internal resistance was 1.88 mR. The time the load was connected for was 24 minutes - 24x5 seconds = 1320 seconds.
So the energy input was 32604 Joules. The cell weighs 1400gms and the temp rise was 16 deg.
From that I calculate the Specific Heat is 1.455 J/gm/deg which is not far from the 1.5 or so that I surmised from looking at the chemical makeup.
It all makes sense to me! Image

Nevilleh
Senior Member
Posts: 773
Joined: Thu, 15 Jan 2009, 18:09
Real Name: Neville Harlick
Location: Tauranga NZ

Li cell discharge tests

Post by Nevilleh »

You know, having read and thought about all the preceding stuff, I am inclined to the view that the internal resistance is not very temperature dependent, but the rate of chemical reaction might vary quite a bit with temp. You might think that this is the same thing, but there is something going on here that is not so readily apparent.

Post Reply