All new* C130 delivered

as far as cars go, the clutch interlock is the exception rather than the rule.

I have only ever driven one car that had that (and it was a Hyundai).
I get to drive new cars reasonably regularly thanks to one of my jobs.

In petrol bike land you have the option of starting either in neutral, or in gear with the clutch in.

on amount of braking disables the engine ( actual in many cornering situations you *need* to both brake and throttle simultaneously -> only a petrol bike problem).

most bikes have killswitches, the idea being if both your throttle and clutch cables break you can still cut power to the coils/spark plugs.

on an electric bike this is pointless (unless connected to a contactor) as any failure that causes an AC controller to run full on will ignore a kill switch signal.

they're included anyway as they're usually part of the design rules.

Matt

as far as cars go, the clutch interlock is the exception rather than the rule.

In my experience clutch interlocks are market driven. None of the UK cars or bikes I have driven have had a clutch interlock that I can recall. All of the US ones have, (or at least the majority), including a US bought Triumph TT600 and a 1986 Chevy pick up truck. (To give examples of age and country of origin.)

My guess Matt is that you're in the UK, and so you rarely see such an annoying device. :)

In the US, I believe that a clutch interlock is required by either government regulation or manufacturer's fear of lawsuit.

Most US cars have other odd and annoying features - like doors locks that automatically engage as soon the car starts rolling at 10 kph or more and automatic transmission cars that only can be started in "park", not "neutral". So, a stopped engine cannot be started if the vehicle is coasting. I learned that in a Chevy Impala at the bottom of a long hill on which I was coasting engine-off as a fuel saving measure.

And I still hate power windows, but there are virtually no cars available in the US that can be bought any other way.

Manual transmission cars have become practically extinct in the US. Very few even know how to drive them. My wife drives what is probably the only M/T Hyundai Elantra in all of the populous state of Pennsylvania. We will be taking a motorcycle training/license course next month ahead of getting our CuMoCo Scooter, and it is going to be fun to watch the other riders - a motorcycle will be the first M/T vehicle most of them (especially women) have ever operated.

On my e-maxs, which have cheap Givi knock-off cases, the case and base plate and mounting hardware were adequate, but the scooters "luggage rack" bracket behind (i.e forward of) the tail light and under the body panel, onto which the base plate hardware was mounted, was too flimsy. I fixed it by adding a pair of aluminum tubing struts between the bracket and gusset plates at the rear of the scooter frame behind the tail light.

Scooters (unlike most motorcycles) are used for day to day transportation, and you need to be able to carry even heavy dense cargo in the top case.

Congrats on the 3K milestone!

On the Kill-A-Watt (KaW) front I very much agree with Mike B. A KaW is a great tool for keeping an independent measure of efficiency, health of batteries etc. It's not a requirement - but it is a VERY useful tool for any EV'er. For those on a budget (or if you're just cheap like me) keep an eye open for the P4400 model. It's the "old version" (the newer version has some non-volatile memory on it whereas this one looses it's data as soon as you unplug it). I find these work well and I recently found them for as little as $16 each... (I think I paid $40 the first time I bought the P4400 when it was just out).

For those who care about the details the charging issue seems to be a varying amount of noise coming from the charger. Unfortunately this noise is causing the voltage sensing circuitry to read high - resulting in early termination. We are fitting an opto-isolator in this circuit at the charger end - it will be a plug in upgrade sent out to existing owners. In fact there's already this circuitry in the BMS - but there's far more noise now being seen than we expected and for some yet-to-be determined reason (a) new chargers suffer it sooner than old chargers but (b) old chargers seem to begin suffering it after some undetermined amount of time.

p.s. note that this failure "fails safe". There's no damage to the battery pack. The LVC circuitry in the BMS isn't affected (because here's no charger induced noise when riding the bike)

4000 miles?!?!! Wow!

Sorry to come in this thread so late, I've only just seen it. This c130 looks like a very nice scooter and bigger than the ones from eRider. Just out of interest, how many Ah of battery is squeezed inside that bike? 72v 50Ah? Or more? Anyway. Interesting read. Cheers.

Mike,

My understand is that the main 24-cell part part of the battery pack slides out the back of the scooter. Various wiring is disconnected, but the BMS harness stays connected to the cells. The swing arm and suspension is is detached and swung aside, then the pack slides out the rear.

I'm nearing 2000 miles, with the pack still healthy. Early on, I added a connector so I can check individual cell voltages. I still can't figure how pack balancing is supposed to work, assuming my BMS _is_ working. The pack balancing I'm familiar with (like the kit balancer that works with the stock charger on my old e-max) clamp each cell voltage at the same value in turn as the cells come up to charge. The charger is throttled back as this occurs to prevent swamping the 1/2 amp balancing shunts. The cells are held at a constant voltage until charge current declines to certain value. Charging is then terminated. So, when checking cell voltage during the final minutes of charging, all cells at a constant voltage and within a few .01 volt of each other - 3.67 to 3.70.

However, on my C124, the cell voltages rise continuously through the "balance mode" - initially very slow - so slow that that completion of charging takes about 1 1/2 hours longer than would be needed with charge protocol that didn't abruptly switch from 8-10 amps to a 1/2 amp trickle. The cell voltages are initially very close, but start to diverge as they accelerate exponentially upward (normal behavior when a cell hits the full point- that's why a BMS is needed. The lowest cells reach about 3.63 to 3.65, the highest cells reach 3.74 volts or so, with charging ending just as the voltages on the high cells are about to hit a worryingly high value. I don't see any balancing going on - the charge shutoff seems to be occurring based simply on reaching a certain pack voltage - (88.4 volts on my voltmeter, a volt higher on the BCU). The only reason things are staying in reasonable balance is that the cells are well-matched, and the very low 0.5 amp charge rate during balance mode keeps the higher cells from "running away."

I've brought this issue up with them but have encountered a bit of evasiveness on the issue. I personally think that CuMoCo does need to work on their charging and pack balancing system a bit.

Paul

Yes,it need active BMS,which moves energy from high cell to low cell without loss.

PCM with energy "running away" from high voltage cell will caused heating even 0.5A current.Efficiency and energy are burnt too.

I've brought this issue up with them but have encountered a bit of evasiveness on the issue. I personally think that CuMoCo does need to work on their charging and pack balancing system a bit.

Hi Paul,

I'm sorry that you feel we're evading the issue. We believe that your BMS is functioning as designed and that it is keeping the pack balanced. Remember that a BMS should be balancing state of charge to within an acceptable tolerance - not voltage. As we all know, voltage is a poor measure of state of charge for lithium based chemistry.

We do agree with you that the balancing phase could be quicker if we were to implement multiple stages of ramp down - and we are looking at that possibility. This is the sort of incremental improvement that any normal vehicle development program will go through. On our end it's just a matter of balancing the available resources and the relative priorities of different aspects of the bike.

Balance mode is not controlled by pack voltage - but by individual cell voltage. Furthermore most other BMS's that we've seen fitted to bikes are shunt style like ours - but not as capable as our design for two reasons:

1) We shunt at 0.5A most others shunt at 0.1A. Starting from the same conditions when entering the balance phase this means that we can achieve balance 5 times faster than these other systems. This means that the owner is much more likely to achieve a balance with our system.

2) We communicate with the charger based on a per-cell voltage. A lot of the systems we see where they mount a module per cell don't appear to communicate with the charger. Thus you have two completely independent algorithms at play. The charger must decide to throttle back either on pack voltage (a bad idea for a balancing system). Most of these chargers appear to not actually reduce charge current to the level that their BMS's can shunt but go into a very coarse duty-cycle mode (x minutes on / x minutes off). This may be "adequate" in an idealized situation but in the field there is much more opportunity for the systems to be out-of-sync and thus not balancing the pack. In the worst case they may in fact damage the pack (by overloading the shunt and transferring more current into the cell).

This is the 30,000 foot view of some of the issues. I was planning on writing a longer description of our BMS and a more detailed examination of the pros and cons of various aspects of the system.

Hi Mike,

To address Mike's concerns - we don't have any plans at present to move away from shunt balancing. So this will result in a small amount of energy that is "bled" off during the balancing phase. We do however plan on a more sophisticated BMS with its own microprocessor control communicating back to the BMS. This will allow for finer grained control and, as mentioned above, we're also looking at multiple current levels from the charger. This will allow us to speed up the balance phase and reduce the amount of shunted current - thus increasing the efficiency of the system.

My understand is that the main 24-cell part part of the battery pack slides out the back of the scooter. Various wiring is disconnected, but the BMS harness stays connected to the cells. The swing arm and suspension is is detached and swung aside, then the pack slides out the rear.

This is correct. Unfortunately Mike has an older bike (one of about 10) that have an older style of battery box mounting. Unfortunately these don't slide out of the back of the scooter. With this older style the cells are loaded with the battery box "in place".

John wrote:

As we all know, voltage is a poor measure of state of charge for lithium based chemistry.

Not true! The discharge voltage curve between a SOC of 97% and 3% is very flat (as it is for other battery chemistries, like lead acid, too). But, voltage is still the only way to verify that an individual cell is full (say, 3.65 to 3.7 volts under a charge current of 0.2 amps), And also the only way to tell when the criteria for fully-discharged is reached (say, 2.5 volts under the max. working current). This also is no different than lead-acid. One alternative would be to maintain a precise integration of ampere-hours up and down (per your fuel gauge and 20% minimum allowable SOC concept) but if the voltage is not at the fully charged level, you are not taking advantage of the full capacity of the cell, or could be over-charging and over-discharging the cell.

With regard to your comment number two, the Goodrum/Fecher BMS (which I have in my e-maxs in conjunction with the stock Pb-acid charger) throttles the charger back by cycling a switching FET on the negative side of the charge circuit. The cycling is typically in the hundreds of Hz, so the charger simply "thinks" this is an increase in pack resistance and lowers its current. No communication is needed between the BMS and the charger and the charger can be a simple power supply with a voltage limit or an ordinary CC-CV Pb-acid charger. The FET is driven based on the duty cycle of the highest shunting channel. As cell charge current decreases, the shunt current increases (the shunt current is variable, because of the cyclic interaction: shunt on-cell voltage down-shunt off), this automatically provides the recommended smooth CC-CV charge profile for each cell without any switching or coarse-stepping the charge current. When charge current (i.e. 0.5 amp full shunt current + cell current) reaches a set low value charging is shut off.

Under older Goodrum Fecher versions, the problem was that the shunts would cook away for a while during the CV phase of charging generating a lot of heat. But then it occurred to them to simply set the charger voltage just below (instead of above) the sum of the shunt "on" voltages. This allows the shunts to only go on if cell 'pokes its head' above the shunt voltage value, which is just slightly above the charge voltage/#cells. This eliminated the heat handling problem, and an ordinary CC-CV charger provides the proper charging. They call the newest implementation the Zephyr BMS. I understand that Gary Goodrum is looking into making a SMT version of it. By the way, it can shunt at up to 1.0 amp, and works for any number of cells.

With regard to my situation, can you answer these questions:

1. What is the GBS cells' charge voltage that should not be exceeded without degrading the cell? A value of 3.74-3.75 at end of charging makes me uncomfortable.

2. If the balancer is shunting at up to 0.5 amp during "balance" mode why do the cell voltages continue to rise, at different rates, some racing far ahead of the others, when the charge current is 0.5 amps? Such behavior is exactly that seen with a healthy pack with no BMS being charged at a low rate, so I am having trouble seeing if or how it is working.

3. You implied that the BMS does not use cell voltage as the indication to initiate shunting. What indication does it use?

4. To test the function of the pack balancer, I can stop charging at the start of balance mode, introduce a slight extra charge to a cell (using a benchtop power supply) then resume charging. If the balancer is working, this cell should not reach a excess voltage over the of the others, nor should charging shut off prematurely (due to the high cell). Correct?

Thanks in advance for reading this and answering my questions...

Paul D.

John wrote:

As we all know, voltage is a poor measure of state of charge for lithium based chemistry.

Not true! The discharge voltage curve between a SOC of 97% and 3% is very flat (as it is for other battery chemistries, like lead acid, too). But, voltage is still the only way to verify that an individual cell is full (say, 3.65 to 3.7 volts under a charge current of 0.2 amps), And also the only way to tell when the criteria for fully-discharged is reached (say, 2.5 volts under the max. working current). This also is no different than lead-acid. One alternative would be to maintain a precise integration of ampere-hours up and down (per your fuel gauge and 20% minimum allowable SOC concept) but if the voltage is not at the fully charged level, you are not taking advantage of the full capacity of the cell, or could be over-charging and over-discharging the cell.

But, as you say yourself "(say, 3.65 to 3.7V under a charge current of 0.2 amps)" - this points to supporting my statement that voltage (alone) is a poor measure of state of charge.

With regard to your comment number two, the Goodrum/Fecher BMS (which I have in my e-maxs in conjunction with the stock Pb-acid charger) throttles the charger back by cycling a switching FET on the negative side of the charge circuit. The cycling is typically in the hundreds of Hz, so the charger simply "thinks" this is an increase in pack resistance and lowers its current. No communication is needed between the BMS and the charger and the charger can be a simple power supply with a voltage limit or an ordinary CC-CV Pb-acid charger.

My second comment wasn't directed at the G/F BMS design but at the types of BMS we see in common use on bikes such as the EMoto G6 and Xero (two bikes we have direct experience of) and probably the ZEV system (just by looking at pictures - so we could well be wrong). Furthermore and speaking in the abstract that switching FET is a communication to the charger.

With regard to my situation, can you answer these questions:

1. What is the GBS cells' charge voltage that should not be exceeded without degrading the cell? A value of 3.74-3.75 at end of charging makes me uncomfortable.

2. If the balancer is shunting at up to 0.5 amp during "balance" mode why do the cell voltages continue to rise, at different rates, some racing far ahead of the others, when the charge current is 0.5 amps? Such behavior is exactly that seen with a healthy pack with no BMS being charged at a low rate, so I am having trouble seeing if or how it is working.

3. You implied that the BMS does not use cell voltage as the indication to initiate shunting. What indication does it use?

4. To test the function of the pack balancer, I can stop charging at the start of balance mode, introduce a slight extra charge to a cell (using a benchtop power supply) then resume charging. If the balancer is working, this cell should not reach a excess voltage over the of the others, nor should charging shut off prematurely (due to the high cell). Correct?

I'll get Erik to answer these for you and post the answers here. Aside from number 3. I must have written something poorly. BMS does use cell voltage to initiate shunting. It uses the first cell that reaches the threshold voltage to switch on shunting. What I meant was that the BMS does not use pack voltage to calculate an average cell value. With an independent charger (not the G/F design) it must use pack voltage / average cell voltage to make this decision.

Thanks in advance for reading this and answering my questions...

You're most welcome. I think the both of us are actually much closer to being on the same page with regard to BMS's - I think it's just a communication thing.

But, as you say yourself "(say, 3.65 to 3.7V under a charge current of 0.2 amps)" - this points to supporting my statement that voltage (alone) is a poor measure of state of charge.

OK, it is a function of voltage and current. But whether it is 3.6 or 3.7 volts, and at what current criteria is splitting hairs. If voltage is not regulated, there will be a point at a low charge current (0.5 to 0.1 A) where voltage will accelerate upward as dramatically as running into a wall. This is the indication of a full pack - the difference between 3.6 or 3.8 volts or 0.1 or 0.5 amps being the last 0.2% of the cells capacity. The general consensus is that "full" for LiFePO4 cells is 3.65 volts at 0.2 amps (or 4.2 volts for LiCoO2 type cell). Higher terminal voltage may cram an bit of additional charge in the cell, but shorten its life an risk damage, while tapering the current down lower than 0.2A just prolongs charging time with little added capacity.

By the way, is any manufacturers literature and specifications available for the GBS cells? Thy don't seem to even have a web page.

Regardless, Current should find a way to charge the packs using the proper CC-CV protocol. The existing setup dos not do this. Go here:

http://batteryuniversity.com/learn/article/charging_lithium_ion_batteries

But back to the matter at hand, what I am noticing now is that near the end of charging, one of the highest cells, when it reaches 3.74 or so, it voltage suddenly gets "knocked down" to 3.68 volts or so, but other cells continue to climb. I tried to knock a few of the high cells down using my string of four 1157 bulbs used for this purpose in the past to get new packs balanced, but overshot a bit so a couple cells driving them down to 3.52 or so. Upon continuing balance charging, cell #1 was now the high cell, but its voltage keep rising and rising - up to 3.79 volts, at which point I stopped charging to save the cell. I knocked it voltage down, and with a little more of these whack-a-mole operations on other cells got a charge termination with the cells ranging from 3.71 to 3.63.

To cut to the chase, the BMS is not working. It is not pining for the fjords, it is not working!!! Or, I hope it is not working as intended.

I don't understand why Current can't simply send me a BMS and I will return the old one. They have nothing to lose, if they are confident that the BMS is actually working, they can simply use my BMS on another bike.

Thanks,

Paul D.

PJD -- a few answers below.
And thanks for the link to the Battery University article. It's generally good, although it is written for "generic" lithium ion batteries and the specifics are not right for GBS nor for other LiFePO4 batteries.

With regard to my situation, can you answer these questions:

1. What is the GBS cells' charge voltage that should not be exceeded without degrading the cell? A value of 3.74-3.75 at end of charging makes me uncomfortable.

GBS spec sheet says 3.80v.

2. If the balancer is shunting at up to 0.5 amp during "balance" mode why do the cell voltages continue to rise, at different rates, some racing far ahead of the others, when the charge current is 0.5 amps? Such behavior is exactly that seen with a healthy pack with no BMS being charged at a low rate, so I am having trouble seeing if or how it is working.

The BMS should limit each cell voltage to 3.8v or less. If you see over 3.8v then the BMS or some other part of the system is indeed broken, and we will fix it.

During the time before the shunt comes on, each cell voltage will naturally be different due to variations between cells. When the cell gets full, the shunt comes on and holds the cell voltage somewhere between 3.7 and 3.8v. This is not super precise due to component tolerances. But as you know there is very little difference in state of charge between 3.6v and 3.8v so the exact voltage doesn't matter much.

A healthy (but slightly out of balance) pack with no BMS does not act like this. It takes very little energy to drive a cell from say 3.8v to 4.2v or beyond and damage it. So with no BMS you either leave some cells not fully charged (and worse and worse with time) or you damage some cells by driving them way over 3.8v.

3. You implied that the BMS does not use cell voltage as the indication to initiate shunting. What indication does it use?

I will expand on John's answer a bit. The key to managing a lithium pack is to count amp*hours in and out of the pack. We do that. We use voltage *only* to determine when a cell is full (turn on the shunt), and when a cell is empty.

4. To test the function of the pack balancer, I can stop charging at the start of balance mode, introduce a slight extra charge to a cell (using a benchtop power supply) then resume charging. If the balancer is working, this cell should not reach a excess voltage over the of the others, nor should charging shut off prematurely (due to the high cell). Correct?

Mostly correct. The cell that you added charge to might indeed have a higher voltage than its neighbors for a while. But its shunt should turn on before 3.8v just like the others.

Here is how I would define "BMS working correctly":

A.. No cell ever goes above 3.8v. (Don't overcharge)

B.. All cells reach at least 3.6v before charging stops. (so we know they are past the "knee" and very nearly full)

It is not necessary, nor even desirable, to drive all cells to the exact same voltage, say 3.700 volts. It gains you nothing, and results in holding some cells at high voltages longer than needed. In fact the "ideal BMS" might stop charging the cells a little *before* they reach the knee. So that some of our 20% un-used DOD is at the top, and some is at the bottom. This would be better for the cells, even though the cell voltages might vary widely. But sadly, neither our BMS nor any other than I have heard of has this capability now.

It's an interesting challenge, eh? Lithium batteries continue to improve, making a moving target. GBS are not the same as the older Thundersky for instance. And our understanding of charge algorithms continues to evolve, both due to changes in the batteries and due to better understanding of the art. There is still debate about this stuff, even at the level of Phd scientists, chip manufacturers, and EV system designers at very large companies.

- Erik

Erik,

Thanks! If the GBS cells are are fine with voltages up to 3.8 volts, then my concerns have been addressed. Watching the charging with a voltmeter, I have seen how the BMS "knocks down" the cell voltage when it reaches the 3.73 to 3.79 range, so it does seem to be working. I agree that if the balancing was not working at all, I would likely have seen voltages well above 3.8 volts.

Also, thanks for the additional explanation behind your approach to charging - it makes a lot of sense. So, it appears the approach on my other scooter's BMS, leaving some cells to be effectively "float" at 3.65-3.70 until everything is balanced, may not be the best approach for cell life. I will be building a latest-version Goodrum Fetcher Zephyr BMS for the other E-max scooter and will take you advice when adjusting it and the charger.

The Norwegian Parrot is alive after all. So close out my complaint!

It's always nice when the bad cell is so obvious - and at an easy to access location too! That doesn't happen very often.

Do you have that glued-down flat black ABS bottom-panel under you seat? Cutting it away at the glue joint and accessing the cell that way will probably be a lot easier than all the stuff that needs to be removed to even start removing the whole seat/storage compartment assembly. The adhesive for this panel is called E6000 Maybe current can send you a tube with the new cell.

Mike (and John or Erik),

Interesting that it was that cell. Because on my bike, the DC-DC converter in the same area (beneath the BCU?) would produce a noticeable amount of heat. I converted to all-LED lighting, including the instrument panel, and only run the small headlamp bulbs (upgraded to brighter LED's) in daytime riding (added a headlight switch). Now the DC-DC barely gets warm. I understand that you have extra lights, plus the summers are a lot hotter down there, so cell #30 may have failed due to excessive heat?

You may want to look at a way of installing a heat barrier between this cell and the BCU/DC-DC converter area (or put ventilation holes in the cover?), and/or switch to LED illumination for everything except the headlights. The upgrade was not cheap (about $170). I had some trial and error involved in getting the proper color and brightness, but I can put together a supebrightLEDs.com bill of materials of the bulbs that seem to work best.

Paul D.

Paul,

Any chance you could post your LED upgrades? I would be extremely interested in what you swapped out your various lights for.

Aaron

Bulb location, Superbrightleds.com part No., Quantity needed, Price Ea.

Tail, 1157-R18-T, 2, $17.95

Turn, 1156-A18-T, 4, $17.95

Lic Plate, WLED-W-120, 1, $1.39
_or_ WLED-NW5, 1, $3.94 if brighter than stock desired

Little Bulbs in Headlight, WLED-NW5, 2, $3.94

Bulbs in rear view mirror, WLED-A-120, 2, $0.79

Inst Panel, WLED-W-120, 5(I think), $1.39

Turn Signal Indicator WLED-G-120, 2, $1.79
_or_ WLED-G4-90, 2, $2.59 - Brighter than stock - helps remind you to turn them off!

High beam Indicator, WLED-B-120, 1, 0.79 (don't use a brighter bulb here - it will dazzle you in the dark road situations you use the high beam on.

With LED's, an electronic turn signal flasher is needed, to get the proper 90 flashes/min. The stock flasher will only flash at 30-40 flashes/min due to the much lower current draw from LED lamps. I used the following:

Turn Signal Flasher, G215-EB, $8.95
This is a 2 wire flasher. The stock flasher is a 3 wire flasher - 1)hot, 2)normally open load, 2)normally closed load. But the normally closed load lead is only used to operate the rear view mirror lights in their alternate mode of on all the time, blinking off when the TS blinks on (achieved by switching some connections - ask if interested). So this 2-wire flasher will work in the normal mode the bikes are delivered in. Find the correct connections with a voltmeter. Red-hot, Black-normally closed load.

Also, note that with LED's always get a bulb with the same color as the lens. So the tail lights should be red, not white. Turn indicator green, HB indicator blue.

Does the pack tester do the charging? I believe the BCU doesn't do anything "fancy" here, it just indicates balance mode if the charging current is below 1 amp. My understanding is that charging "decisions" are made with the BMS communicating with the charger. For example, when I had a bad charger that could only charge at an amp or less, I would only get a "balance charge" indication.

As far as the cell, my Thunderskys would swell noticeably if discharged completely flat (when testing an old sell). There would also be audible sloshing or electrolyte. Recharging would remove the swelling and the electroltyte would be re-absorbed. The completely dead cell was acting like a big resistor and heating up? Does the scooter shut down when a LVC condition is detected? Was the melted side facing the DC-DC? The heat from it could have softened the plastic enough so it deformed under the swelling pressure and bot actually melted?