Next Step

SETTING UP FOR TESTING

Because I'm vary wary of lithium batteries I've added two more temperature sensors which give a digital readout of the exact temperature of the cells at all times. The temperature sensors I've added are called thermocouples and they use a standard multimeter to show the temperature (the thermocouple units were kits bought from dick smith electronics some years ago I think they are discontinued nowadays). The batteries will be charged off the bike (in an area where if they happen to catch fire nothing will be damaged) and temperature monitored. Also two multimeters will be sitting on the handle bars to show the temperature while riding.
I think its especially important to have an accurate idea of the battery temperature when riding as if the motor is under load for a long period (say going up a long hill) that will be the time when battery temperature should start to rise and any sign of overtemperature conditions hopefully can be avoided. If the current drawn from the batteries while riding is kept below its maximum discharge rate ( unfortunately i'm not exactly sure what it is for these batteries probably 10 to 15amps.....I hope!) there should be no problems.

Below shows the components used for monitoring temperature:

I'll be using the batteries in series to give 48volts 10ahr ( hoping I can do this as it will be a very short experiment if they cant be used in series I'm hoping it wont have any effects on the bms circuit).
I'll use the GL-1 motor on a mountain bike. If any one knows a good reason why I should not try them in series please tell me.
In between the batteries I'll use 3 or 4 car alternator diodes to stop any current going back from one battery to the other. During charging the batteries will be isolated from each other by a high current switch from a jump starter unit.

Sunday 16th July 06

Battery box used is an old globite school case (over 20years old) and still in good condition, they certainly last.
In the case are the two lipo batteries 24v 10ahr each connected in series. The circuit has a heavy current switch from a jump starter unit which turns off power and also isolates the batteries when charging. Also 3 car alternator diodes are used in parallel between the batteries to stop any current flowing backwards from one cell to the other.

Pictures above show the batteries ready to go on the back of my small trike. The motor is a brushed P3 running at 48volts.
The controller can run on 36v or 48v it has a variable power screw so I can adjust the maximum amps the motor draws, I'll do the first test ride with the maximum amps on about 10amps.
Just as a test to see if everything works ok I ran the motor using the lipos in series (48v) with the front wheel off the ground, under low load the battery temperature raised by about one and half degrees. When I applied the brake to put load on the motor the battery temperature went up about 4 degress from 14degrees celcius to 18degrees celcius. I haven't set up an ammeter yet so I'm not sure how many amps I was drawing probably less then 5amps or so.
So it appears the bms circuit is not effected by running the lipo cells in series which is very good news. I tried out the chargers after the short test (first isolating the batteries by turning the high current switch off), all seems fine the chargers registered the batteries as fully charged. Haven't measured battery voltages yet.
Whilst installing the charger connectors I did accidentally short the red and white wires a couple of times producing a decent spark, so it appears that current can flow from pos. to neg. via the charger circuitry without damage (I hope!!!). Though this does mean there is potential for a short to occur via the charger plug when raining or high humidity.

Heres a schematic of the circuit:

Since the red and white wires (coming out of the batteries) are always live there needs to be low amps switches on those wires (which go to the charger plugs) so no shorting could occur. The charger plugs are on the outside of the case and if a short occurred due to moisture the battery could discharge and possibly catch fire.
The high current switch when 'on' connects the two batteries together to give 48v for the motor. When in the 'off' position it means the two batteries are isolated from each other and the chargers can be connected.

Next Step: test rides and get some data.
My small trike already has a 48v 17ahr set of sla batteries, so I'll be able to ride till the lithiums start to get low (low voltage cut off is about 24.5volts for each battery) and then return home using the sla batteries. Fortunately I have a bike track very close by which goes for about 5km through some very nice bushland and NO CARS!!!
So I guess I'll be riding up and down the bike track until the batteries give out.

Tuesday 18th July:

Still setting up for test rides, have placed temperature display, current meter and voltage panels on handlebars so I can monitor while riding.
Did some testing under load simply stopping trike from moving and using throttle.
Some very surprising results: temperature increase appears to be very high indeed under load. Going up to about 12amps gives an apparent temperature increase up to 70degrees celcius or higher. I suspect that the outer case of the lithium cells has current flowing through it and is effecting the results of the temperature probe. The temperature bimetal joint is sitting directly on the outer casing of the lithium cells, if there is current flowing through that casing it could be giving false temperature readings (I hope so!!!). I will have to recalibrate the thermcouples with boiling water and ice (100degrees C and 0 degrees C) just to make sure they are accurate and put some insulation on the bimetal joints and see what happens.

Tuesday 18th July continued

Have electrically insulated the thermocouple bimetallic joint in case that was causing temperature reading problems just using heatshrink plastic.

The thermocouple wire for each battery is sitting next to the topmost cell at the positive end of the battery where its likely to get hotest.

The instrument panel (dont laugh!) below (just one more voltmeter to put on).

Some surprising results:

[postscript: the temperature readings I was getting were caused by interference from the bms circuit in the lipo batteries, I'm yet to resolve how to accurately measure the temperature of the cells ( I will need to put some shielding around the temperature sensor probe wires), suffice to say that the batteries will only get warm to the touch after prolonged use under load...........very promising indeed........ but below I'll leave up my test results just for anyone whos interested in how I went about it although the temperature readings were not anywhere near accurate)]

I calibrated the temperature sensors. In boiling water the temperature sensors read 99 degC-100 degC. In ice they measure approximately zero degrees, so I know the temperature sensors are working correctly.

By applying the brakes on the bike and then turning the accelerator I can measure the amps and control the amount of amps being drawn from the motor. When the motor is off amps is zero of course, and lipo cells were both 12 degC. (air temperature).

Applying a small amount of amps (the motor not being able to move because the brakes are applied) gave an immediate increase in cell temperatures. At 2.31amps Lipo Battery 1 reached 46degC and Lipo Battery 2 reached 29 degC. Batteries are in series so a temperature difference was expected. [postscript: this temperature reading is not the temperature of the battery it is due to interference by the bms circuit inside the battery case]
Also it was possible to hold the throttle in position giving a constant current of about 2amps, the temperature of the batteries did not vary when the current was constant, they remained at the temperatures of 46 and 29 degC.(approx).

I'm very surprised at this large temperature increase of the batteries for such a small current.

I then increased the amps (just by turning the throttle more). At about 12amps constant (12.13 amps in the photo below) Battery 1 reached
99 degC (!!!!!!!) and battery 2 reached 84 degC . When power was cut off (release throttle) the temperatures fell back to 12degC very rapidly. [ postcript: when power is turned off this temperature measure of 12degC is the actual temperature of the batteries!!!!!

The outer casing of the batteries were not warm at all to touch [ postscript: quite important to note that the temperature had not risen at all really], I think the thermocouple is giving a very accurate idea of the temperature of the outer skin of each Lipo cell. I didn't leave the batteries at this temperature for very long probably 10seconds or so. [ poscript: wrong it wasn't measuring the temperature really at all, it was due to interference of the bms picked up by the temp probe!!!]
I really have no idea at what temperature thermal runaway will occur ( and yes I will be running away very quickly if thermal runaway starts to occur): hence the whole idea of having temperature sensors to see if this starts to occur. If I see the temperature just continually rising its time to disconnect any power quickly and runaway!!!

I'm actually quite concerned now about using these batteries at all!!! [ postscript: turned out to be nothing to worry about at all it was rf interference from bms circuit causing anomalous temperature readings, looks like I might have to go back and try again and use some shielding around the temperature probe wires to cancel any interference]

I now have some small understanding why lithium batteries are so dangerous. Their temperature rises rapidly the more current that is drawn, even a small amount of current gives significant temperature increases ( I guess there is some sort of exothermic reaction occuring in the batteries).
I think I will be putting a metal shield around the battery box before going for a test ride just in case.

20th July 2006 Test Ride No.1 (early morning, slight constant rain, air temp. about 8 degrees at start)
Note Battery1 is the first in series (its negative lead goes to the controller/motor) and Battery 2 is the second in series (its positive terminal goes to the controller/motor).

Track was very slight constant gradient to almost flat. Total distance travelled was 11km before motor cut out, reason it cut out I'm not sure until I recharge the lithiums, they may have an auto cut off when they reach a certain voltage but I'm not sure as yet why the power cut off.

Before Test Ride: Battery 1 (24v 10ahr): temperature 8 degC and voltage was 26.4volts
Battery 2 (24v 10ahr): temperature 8 degC and voltage was 26.3volts.
There is a diode between the battery packs which drops the voltage by about 0.3volts so total voltage was 52.4 volts (two batteries in series).

After Test Ride: Battery 1 (24v 10ahr): temperature 29 degC and voltage was 23.5 volts
Battery 2 (24v 10ahr): temperature 26 degC and voltage was 25.1 volts.
There is a diode between the battery packs which drops the voltage by about 0.3 volts so total voltage was 48.1 volts (two batteries in series).
(yup I know it doesn't add up but voltmeter read 48.1 across both batteries and diode measured 0.3volt drop., theres 0.2 volt missing somewhere maybe my dodgey connections).

[postscript: the above is probably the only usable info on temperature readouts that I gained from the testing, as the temperatures were taken when the power of motor was off, so there was no interference from the bms system. So basically one battery had cells that reached 30degC which is quite a modest temperature considering the amps being drawn]

Description of the ride: First part of the ride for about 4km I set the controller on minimum power setting max. amps drawn was roughly 10amps though the meter did peak at 15amps briefly a couple of times. Temperature of battery 2 was generally higher than battery 1 if the amps drawn were less than about 8 amps . But once over about 8 amps the temperature of battery 1 would rise sharply and peak at over 100degC whilst battery 2 would stay at about 60degC or less.

The second half of the ride I put the power setting on controller to maximum. Its a gradual uphill slope the second half of the ride of very constant gradient (an old railway track converted to a bicycle path). Was drawing a constant 10 to 12amps, temperature of both batteries was around 70degC until I drew more power (more throttle) and again battery 1 would rise steeply in temperature peaking at about 150degC whilst battery one would remain at about 70-80degC. Maximum current drawn very briefly was just over 20amps (meter limit is 20amps so it just registers a '1' when over 20amps). I tried to keep current draw below 15amps by adjusting throttle to suit.

The temperatures remain steady if the current is held steady (quite hard to hold a perfectly constant current while riding though).

Importantly though, as soon as I would take off throttle power the temperatures immediately drop back to what I gather is the true battery temperature. For example at the start the temperature probes were registering 8 degC and they would not go below this, so I gather that is the true temperature of the batteries (same as air temperature at that time). But as soon as current is applied the temperature shoots up, letting off the throttle it drops back to a minimum(very quickly almost instantaneously) which reflects what I think is the true battery temperature. It took some time for the minimum value of 8 degreesC to start to creep up, by the end of the ride the minimum temperature the probes would drop down to was 30degC and 35degC (for battery 1)(when the throttle is let off i.e. current equals zero being used).

[postscript: 35degC appears to be the maximum temperature one of the batteries reached, the place where the thermocouple was placed was on the cell that would get hottest (the last one in series), so it could be assumed (probably) that the other cells in the pack would have lower temperatures than this]

After sitting for some ten to fifteen minutes after the ride the temperatures were 29degC for battery2 and 27degC for battery1.

I was constantly letting off the throttle to check the true battery temperature while riding. So what I had in front of me while riding were basically high temperature readings getting up to about 150degC to 160degC but true battery temp readings seem to be between 20degC to 30degC whilst using the motor.

Here is a video showing the meter readings as I was on the second leg of the journey (uphill). The right hand meter is the temperature of battery1 (3 digits showing a peak breifly of 160degC but generally holding around 150degC), the middle meter is the current (note right at the start it goes to '1' where the motor is briefly drawing over 20amps, it has 4 digits and shows the current drawn from the batteries it is generally around 15amps in the video, well it really keeps dropping as my speed increases). The left hand meter is not really visible but shows the temperature of battery2 ( it holds steady at about 70degC and peaks at about 80degC about half the temperature reading of battery1).

http://video.google.com/videoplay?docid=7160445582054465758&hl=en

This is very puzzling indeed, why are the temperature readings so high when current is being applied?? The thermocouple probes are sitting next to the top most portion on the outer casing of a lithium cell. Is it possible the outercasing of each cell gets quite hot whilst the overall temperature of the cell remains quite cool??? Or is there something else going on???? I really have no idea!!!!

Can anyone else verify getting similar readings?? The batteries are lithium polymer type so i assume the outer casing of each cell is a polymer.

Picture below shows the battery temperatures after the bike had been sitting for about 20minutes to half an hour after using the motor. It seems very slow for the temperature to drop back down to air temperature again.

After the first test I put the lithium chargers on the batteries.

The temperatures immediately rose to about 35degreesC when the chargers were put on. If I turn off the chargers this temperature immediately drops back down to the true battery temperature. After about one hour of charging battery2 temperature whilst charging is 25degC and battery 1 temperature is 35degC. It would seem battery 1 is accepting more charge than battery 2. Taking the charger off the temperatures immediately drop down to 17 and 18degC (about air temperature).
[postscript: once again the interference by the current (probably something to do with the bms system) the temperature probe occurs, so the temperature when the charger is taken off is the true temperature]

Picture belows shows the temperatures when first put on charger after a ride.

May have a solution to the mysterious high temperature readings. Ray Cooper of Monash Uni (rc plane enthusiast see http://www.ctie.monash.edu.au/hargrave/aerobotics.html) (via Daryl Gaspero, Monash) has suggested the high frequency switching of the controller may be inducing a reading in the temperature probe. Bulls-eye?
I think the controller I'm using runs at about 20,000Hz or thereabouts, meaning the current is switching on and off 20,000times every second, which means the electrons leaving the battery are also leaving at intervals of 1/20,000 of a second (go/stop/go/stop etc), this is how the motor speed is controlled, the interval between the on/off periods is varied so that the motor 'sees' a given voltage depending on the on/off intervals.
It is true that the high temperatures are only seen when the current is flowing (well its really flow/stop/flow/stop 20,000times a second approx.).
But how could that on/off current flow induce a current in the temp. probe???? Sorry I dont understand that part ..........any suggestions out there?
Furthermore how could the temp. probe be isolated (shielded) from the induced current/emf in the probe wire?? A small piece of aluminium around the bimetal connection? some mylar around it??

Nope Daryl Gaspero of Monash tried a thermocouple wire near a controller while running with no interference seen. Might be due to the bms circuit in the lithium pack itself is causing the strange temperature readings??
[postscript: appears to be the bms circuit or the current going through the cells causing the incorrect temperature readings]

July 30th 2006:

Have transferred the Lipo packs to a two wheel bike. And testing will be continued by Neddie in Sydney. Will post results as I get them.
Some pics of the Lipos set up on Neddies bike. Motor I've set up on his bike is the GL-1 on rear wheel 26", controller has maximum current draw of about 18amps, controller has regen. braking but wont be used just yet!!! Most likely the regen. could be directed to the batteries via the charging wires but wont be trying that for some time to come.

video of bike above being tested clink link below
( me (brett) riding at approx. 40km/hr top speed video taken after about a 10km ride: still plenty of oomph left )

http://video.google.com/videoplay?docid=1977546003028798554

3rd August, 06: update
Daryl at Monash has tried a thermocouple near a running ebike pwm circuit , no interference present. This suggests the interference may be coming from the bms circuit on top of the battery. Thanks Tim Brown (yahoo power assist group) for suggesting using a 0.01uF to 0.1uF non-electrolytic capacitor to cut out the high frequency interference in the thermocouple readings. Daryl has also tried these capacitors on a thermocouple to see if they would change the readings just of normal objects, seems no effect from capacitors (very good!! means the temperature readings should be accurate but hopefully with no strange readings due to the interference) so hoping Neddie in sydney (who is now testing the batteries) can try out a capacitor across the inputs from thermocouple into the voltmeter soon.

Have done a few tests runs before Neddie took the bike/battery setup to sydney, batteries remaining very cool after long ride and under load, thus far things looking good.

A couple of observations on the batteries and chargers: the batteries have a low voltage point at which they will shut off to protect from over-discharge. Also when charging the batteries the charger will eventually turn off if the charger is forgotten to be turned off.
On the charger are two led lights one indicates that the 240v ac power is on, the other led light turns from red to green when battery if charged, if the charger if left on after this point it will after some time turn itself off (no led light showing). That I found accidentally after forgetting about the batteries were charging and was suprised to see no led showing ( I'm assuming no led means power is off).

Update: Aug 7th 06
Neddie in sydney has been quite busy and has reinstalled the batteries into his bike. He's using a metal case (just in case of fire) and has tried to make the system very waterproof and has used some large connectors to clean up all the wiring. A few pics below. The battery case can be easily removed for charging in a safe place.

Will post more test results as I get them.

After the first test I put the lithium chargers on the batteries.

July 30th 2006:

http://video.google.com/videoplay?docid=1977546003028798554

beware of spray ebike company: more info