Low voltage alarm for many LiPo cells?

Low voltage alarm for many LiPo cells?

Summary:

Is there a simple and cheap way to to get a visual or audible alarm, if ANY LiPo cell voltage goes under a set voltage in my electric bike during riding? (without a bms during discharge).
Something like several buzzer alarms https://www.banggood.com/buy/lipo-tester.html, about 1 USD for 8 cells in quantity of 20x, but more practical.

Long story:

My bicycle has a 250 W hub motor, a 15 Ampere "no name" controller 36 V, and a Chinese bottle battery of about 320 Watt hour capacity (8 Ampere hours). Inside the bottle there is a BMS and 40 pcs of 18650 cells "no name". A charger for LiPo batteries max 42 V (4.2V/cell), which came with the bottle battery, 2 A charge current. All these were once bought together as a kit.

I have also 6 pcs of 5 Ampere hour 18 Volt LiPo packs (Zippy Flightmax 5000 Ah 5s) discharging max 20C, or 100 Amperes. There is balancing charger Imax B6ac up to 6 serial.

During 2-4 years I have happily used both batteries SEPARATELY(either the bottle battery or all the Zippy packs in configuration 10s3p). Suddenly the bottle battery refused to take charge.

Assuming the bottle battery is now scrap, I opened it to learn about batteries, cut all the wires and cut the point welded nickel(?) connections. Most 18650 cells had 4.1 V charge , 12 cells had about 4 v. Apparently none of them was dead. But the bottle battery cannot be assembled as it was. I tested one 18650 cell capacity by discharging from 4.2V to 3V and then charging. 1996 mAh.

An idea: what if I use Vruzend plastic connectors for the 40 pcs of 18650 cells and assemble them in packs of 5s, 8 packs total? (each pack like one Zippy pack 5s). Then with thick cables (as Zippy has) I could combine 2 packs to 10s. Those pairs can be temporarily parallelled with thick wires up to 4p at ground and 36V level(nominal) to get e.g a configuration 4p10s. No bms, but always balance charging with Imax B6ac to 4.2V/cell. The original cell might have had capacity 2.2 Ah as new, and max burst discharge 2C = 4.4 A .

During riding/discharging there has been no balancing of Zippy cells, no balancing wires used. Paralleling connections are only at ground and 36V level(nominal). I have tried to be careful not to discharge under 3.6 volt in any cell. Stopping the ride when in doubt and checking with a Hobby King cell checker
https://hobbyking.com/en_us/6s-hobbyking-lipo-cell-checker.html?___store=en_us
Balancing only when charging. Zippy zells still keep their balance well.

The Lipo cell checkers from Banggood (https://www.banggood.com/buy/lipo-tester.html) would be a feasible solution, only 15 cents per cell, but I have to insert/remove the display/buzzer products before and after each ride to prevent discharging when not in use. Too awkward, if there are e.g. 14 insertions each time. The accuracy is not very good, but that is not so important. But anyhow, this example shows that there might be solutions to this problem at a very low cost. An Arduino based simple circuit to read the voltages from each cell? I have made simple Arduino boards for my own use, but how to simply read/multiplex/scale 42 Volts to an 5 Volt AD-converter. (how has the product from banggodd done it?).

If there is an easy to understand way to read and hear the offending voltages during riding (without stopping) I could later replace or remove the worst cells, and also expand the battery with used 18650 cells from Ebay. I would like to use both Zippy packs and self made packs of 18650 cells parallelly connected together.

Ideas?

Look, given the cost vs performance ratio of LiPoly, I would forget the round cells and focus on doing the LiPoly the correct way, which you are not doing now. Mixing chemistries is just a bad idea.
As is relying on an alarm for a Low Voltage Cut(LVC). An ebike running LiPoly needs to have a fool-proof, hard and fast LVC and that should be in the controller. If your controller lacks a LVC, controllers w/ LVC's are not expensive. A 36V controller w/ a 31-32V LVC is somewhat low for 10S, but is doable(9S is much safer), but a 48V controller w/ a cut @ 42V is the best match for LiPoly(12S). The best way to deal with low Voltage limits is to carry enough capacity that one never gets close to the LVC. Riding without "range anxiety" improves the experience greatly!
After many years, my system has evolved to this;

View attachment 1


4) 6S/10000 mAh Multistar bricks wired 2S/2P. That is 12S(44V) and 20000 mAh capacity. By using the largest bricks that fit my space, which simplifies wiring, but more importantly, it leaves only 4) balance taps to deal with. Ck'ing and on the very rare occassion I need to balance, is done w/ 4) Battery Medics(B.M.'s are the only tool for LiPoly, all the other checkers and alarms are useless) . I just priced my batteries and they are $280 shipped to me here in the States. No other chemistry comes close to this cost/performance ratio.
After going thru many RC balance(toy) chargers, I now bulk charge w/ the reliable industrial Mean Well HLG series of LED power supplies(no more need to "break" the packs to charge);


Here charging 2) 6S/5000 mAh bricks in series that I use for the frt. system of my 2WD.

I have also made LiPoly "bottle batteries".


This one holds 4000 mAh of 12S(44V).

https://endless-sphere.com/forums/viewtopic.php?f=2&t=84624&p=1242082&hilit=+water#p1242082

I luv LiPoly

Note: Zippys are ok, but Turnigy is better and Multistars, for system with low to mid discharge rates are the best.

First principle, get a handle on your battery characteristics. You'll need a watt meter, to determine real world capacity, and monitor that as you ride. The cycle analyst is ideal, but it can be a cheapie if you cant afford a ca. And you will need some kind of cell checker to show cell group voltage, which can also be an alarm. (you have one)

Your capacity changes, getting less, the higher you draw the amps too. So one slow ride actually has more capacity, as well as losing less power to the wind in the fast ride.

This knowledge of your battery allows you to know just which group of cells has the least capacity, from using the cell checker when the battery is nearly empty. Knowing the capacity, you can plan 90% of your rides so you are well in the safe zone.

Once you get a handle on which pack, or group of cells tends to run out first, you can just put one cell checker/alarm on that pack. No need to monitor cells you know will not the be first to run out. And you only need the alarm on the 10% of rides you know will use 100%. You can be out riding, and seeing the numbers on the watt meter, you will know when to briefly stop, install that alarm, and limp home careful so you don't over discharge any cells.

As you limp home, the alarm starts to beep, slow some more till it stops. Keep riding slower and pedaling harder, and when it won't stop when you stop, you are now pedaling home.

You could also use a simple voltmeter, bar mounted, and watching just that one lowest capacity packs weakest cell. Always many ways to do this, but the basic idea is watch your worst cell. unless you unbalance the pack enough to change which cell this is, you know which cell runs out first every time.

motomech and dogman dan, your replies gave me new ideas and a lot to think. Thanks.

I did not know that Hobby King has Multistar packs up to 20 000 mAh at a lower price than I had thought. A kilowatthour about 244 eur including local taxes in European Union! Unfortunately, there are no 5S packs, only 6S or 4S . For 36V I have used 5S+5S.

As an oldfashioned consumer (and software and electronics hobbyist) I am used to hate solutions, where I have to throw away a product with many good parts instead of repairing it.

E.g. if there are 40 cells in a battery and one cell has gone bad: throw away 39 good cells!? What really happens if I replace the bad cell?

Connecting used cells to serial and parallel combinations is risky. But people in youtube videos seem to build huge battery walls etc without worrying much about internal resistances, allowed amperages in charging and discharging (for each cell), Kirchhoff's circuit laws , temperatures etc. It would be nice to approximately simulate all the cell currents in a planned ebike battery during a ride (as a function of varying load(t)) with 'second order models' and then verify results with the real battery. One needs to measure internal resistance and allowed amperage range and charge/discharge curves as a function of SOC(state of charge)for each intended cell before assembling the battery. If the model gives acceptable values then one could assemble the battery and have a test ride. Possibly with magic smoke .

Re magic smoke: the damages could be limited during a ride when a person is present all the time and there might be a warning from 'Banggood' cell checkers some seconds (?) before the catastrofe. Cell temperature could be sensed without stopping with one's hand when the battery is open in a basket in front of the handlebar.

Charging with balancing could be done with the genuine imaxB6ac outdoors where a fire would be contained. Balance charging e.g. one pack or even one cell at a time, if necessary. Thus 4.2 V per cell would not be exceeded.

About calculating battery currents:

http://web.mit.edu/sahughes/www/8.022/lec08.pdf (pages 77-79). Hundreds of cells could be calculated!
('the first order models').

Comment containing craigsj's comment (pity that craigsj left and removed all his writings when this topic started to become really interesting ):

https://endless-sphere.com/forums/viewtopic.php?f=14&t=75705&hilit=parallel+matched+cells#p1277899

Thoughts?

Here is an image of 6 packs of Flightmax 5000 mAh, connected 10S3P.They give 540 watthours (= 0.5 kWh, = 15 Ah at 36V) in theory, weight 4.2 kg.
The motor controller ( its display at right) has an undervoltage protection (might be 31.5 V) for the whole battery but not individual cells.You see that SOC display has only 4 states.
The multimeter in the middle (mock up so far, only 5A) could show the amperes being used each moment and thus the power in watts would be roughly the Amperes times 40 Volts. (poor man's power analyzer...).
At the left there is a usual speedometer.
The yellow thing is Hobby King's cell checker, showing voltage in one pack total and 5 cell voltages inside it. 5 more thingies like that will show all the cell voltages (ordered 10 from Banggood, about 17 USD total, with audible alarm).
The shopping basket has thick foam inside. The cardboard box inside is covered with plastic to protect from light rain.

There is ample space in the basket for Arduinos, more batteries, cellphone, heart rate monitor etc.

The green part at the right is my recumbent bike, with electric motor also.

Of course all is very temporary, easy to change arrangements. Experimental lab. It will never be ready

Dogman dan might now see how I really agree with him. And motomech's images are very informative, therefore me too...

Everybody is showing here well packaged elegant battery arrangements. But here are just the opposite, hastily thrown together parts and then riding...

I was eager to test Vruzend kits with the simplest possible battery: 10s1p cells of 2 Ah (type 18650, unknown manufacturer). Instead of a BMS I had 2 cell checkers each monitoring 5 cells, and supposed to give a loud alarm when any cell voltage is out of range. A multimeter (10 usd, max 5 A) shows the instantanious current from the battery.

ESC took 60 milliamperes when idle. Pedaling without any significant force took about 100 milliamperes (pedelec). The inrush current from the motor took about 4 amperes and the cell voltages dived to around 3.3 V. Nice to see those things in real time! What is the feeling if 1 ampere (41 wats) is used? Oh, on level ground the bike went forward, not high speed, but I was in no hurry...What about 80 W? ... About 3 km test ride and then home to make notes about each cell. I had charged every cell with imax6ac to 4.1 V before the ride. 2 Ah at 36V gives about 72 watthours. Typically I use at most 10 watthours / km. But if leaving at least 20 % of the charge unused could give about 5-6 km with this tiny "battery".

The cheapest cell checkers from Banggood (about 1.5d usd each in lots of 10) had +-0.15V errors, but repetitive measurements of the same cell were within 0.02 V mostly. It is easy to tape a correction table to the device, if needed. When the lower limit of the voltage can be e.g. 3.3 V , 3.2V, 3.5V almost arbitrarily then high accuracy is not necessary. The balancing charger takes care of the high voltage limit very accurately.

The feeling was like moving magically with the aid of a small lamp (not exceeding the speed of light, of course). The "battery"cells weighs only 0.45 kg , 0.045 kg each.

I must admit that I have never understood what BMS really do. With this system I am solely responsible for all the actions with the battery. Nice to be in 100 % control. (or is it 100 % ?).

Vruzend kits made it possible to put together configurations like this. Test it. Change it. Add more cells. Remove it. Etc. Thanks Vruzend!

Everybody is showing here well packaged elegant battery arrangements. But here are just the opposite, hastily thrown together parts and then riding...

Click to expand...

Well, you said it, not me And I'm sure it's great learning experience and the last thing I would want to do is slam anyone curious and energentic to do what you are doing.
But, yes, there is a more elegant solution.
My goal has always been to make using LiPoly as simple and autonomous as possible, to work towards a "Plug and Play" solution.
The first big "breakthough" came several years ago after replacing several RC (Toy) balance chargers due to failures. These were Quality chargers too, they simply don't last for Ebike usage in the long run. By going to "bulk" charging, I was able to stop "breaking" the packs to charge, avoiding accidents. The side benefit to this is the cells stay balanced better when their series connection is not always being interrupted. By using a large capacity pack relative to avg. discharge usage, it's easy to observe to 80% discharge rule, so these thing together means I no longer need to balance the cells(But I always ck. them).
My first attempt at bulk charging was w/ a 2S/3P system using first, Zippy Flightmax and then, the better Turnigy 20C. This required paralleling the balance leads, which turned out to be tricky and messy.
The next "Breakthough" was coming to understand that the best way was to stay @ 2P, which reduced the brick count, allowed me to use 4 Battery Medics and elimiting balance wiring altogether :lol:

I did not know that Hobby King has Multistar packs up to 20 000 mAh at a lower price than I had thought. A kilowatthour about 244 eur including local taxes in European Union! Unfortunately, there are no 5S packs, only 6S or 4S . For 36V I have used 5S+5S.

Click to expand...

2S + 3S bricks can be series together to create a 5S brick. Of course, their balance taps need to be joined. There are adapters to make it easier. Also 5800 mAh and 6200 mAh 5S brick are avail. in Zippy Compact and Turnigy 20C. If a 11000 mAh or 12000 mAh pack(2S/2P) is not large enough, build 2) 2S/2P packs and keep them apart. Run one down to LVC, then disconnect/connect @ the controller.

As an oldfashioned consumer (and software and electronics hobbyist) I am used to hate solutions, where I have to throw away a product with many good parts instead of repairing it.

Click to expand...

When the MultiStars became available and I was able to move up to larger bricks, I had 6S/5000 mAh bricks I wasn't using. I repurposed them into other projects. I now have LiPoly powered lights, heated grips and heated jacket. And a LiPoly car batt. jumper.

E.g. if there are 40 cells in a battery and one cell has gone bad: throw away 39 good cells!? What really happens if I replace the bad cell?

Click to expand...

Although it's very rare these days to get a bad cell in a LiPoly brick, it is easy to "jump over" the bad cell. If it the first or last cell, it can be removed(Although it's not neccessary). Of course, this reduces the cell count, which may preclude using it again in a specified system.

Re magic smoke: the damages could be limited during a ride when a person is present all the time and there might be a warning from 'Banggood' cell checkers some seconds (?) before the catastrofe. Cell temperature could be sensed without stopping with one's hand when the battery is open in a basket in front of the handlebar.

Click to expand...

"Smoking" a LiPoly while riding is rare and can only happen bricks become shunted due to worn, chafed wiring that gets "Crossed over". Over-discharging or allowing the cells to get way out of balance, while bad for the service life, does not lead to a reaction w/ low-volatile LiPoly like the MultiStar. In 8 years, I have had only one "smoking" event and it was my fault. I accidently pierced a pouch and could smell it right away. I was curious, so I opened the hole to look inside. It was glowing orange like a mini-Mordor and then it starting smoking till the energy was depleted. Rather anticlimatical actually.

Here is an image of 6 packs of Flightmax 5000 mAh, connected 10S3P.They give 540 watthours (= 0.5 kWh, = 15 Ah at 36V) in theory, weight 4.2 kg.
The motor controller ( its display at right) has an undervoltage protection (might be 31.5 V) for the whole battery but not individual cells.You see that SOC display has only 4 states.
The multimeter in the middle (mock up so far, only 5A) could show the amperes being used each moment and thus the power in watts would be roughly the Amperes times 40 Volts. (poor man's power analyzer...).
At the left there is a usual speedometer.
The yellow thing is Hobby King's cell checker, showing voltage in one pack total and 5 cell voltages inside it. 5 more thingies like that will show all the cell voltages (ordered 10 from Banggood, about 17 USD total, with audible alarm).
The shopping basket has thick foam inside.

Click to expand...

The LVC of a 36 V controller is a little low for 10S(it's great for 9S), but is better than nothing(This is why 12S is great for Ebikes-a LVC of 42V, means stopping the cells @ 3.65V, perfect). There really is no substitute for a hard and fast LVC in the controller when using LiPoly on an Ebike. I have found that trying to monitor the batteries while riding to not be practical. The small text requires attention that would best be on the road and lacking backlighting means nothing can be seen at night. The biggest deal breaker for me was having to dis-connect all the passive devices to keep the pack from dis-charging when not in use, a major hassle that I did not want to deal w/ every ride.. I use only a LCD Voltmeter that only draws mini-Amps and can be left hooked up all the time.


two shown here, w/ backing switched on.
.
.

I will ocassionly ck "sag" under acceleration and knowing what to expect as various states of discharge, any unexpected sag, is all the "red flag" I need while riding. The time to monitor cell health is during the charge cycle, to what degree do the cells stay together, especially after a low-discharge use. If there is a concern, a dis-charge test should be run and I keep a balance charger around for just this purpose.

Bottom line; Big bricks for a brick count of four. Bulk charge using Battery Medics(The good Blue ones, not the blacks).

First principle still applies. Find out which group of cells has lowest capacity. Put your cell checker on the balance plug for that group.

Again, if you know the ride is short enough, you don't need any checker. You just need to run the alarm when you are knowing you are going to ride it far enough to go below the low voltage limit you choose to use as your stop now point.

The cheap cell checkers can even be set to your desired low voltage to start the alarm beeping.

Do tape up your parallel connections at the plug. I have "smoked" lipo packs by having one disconnect during a bumpy ride. Then I'm riding along thinking I have 10 ah, when half the 20s assembly has only 5 ah. Whoops.

dogman dan, I agree with you in every item..

With an alarm connection for every balancing wire I'll see, if a wire gets loose or burnt or its threads are cut partially or if a screwdriver falls onto cells (!) etc. I'll be interested rather to know too much of the battery, at least to start with. To have a peace of mind that most probably nothing is terribly wrong, if all the alarms are quiet and display some values. Later I might become tired to have too much information available. Then a loud alarm in the suspected spot might be enough, if anywhere.

In fact the main purpose of (pedelec-) biking for me is not to get from A to B, but to measure new or unusual things. E.g. I am dreaming of connecting my heartbeat sensor to the motor controller so that the motor automatically adjusts its assisting power to try to keep my heartbeat within desired range at any moment. There is a pseudonym "jenkie" at German pedelecforum, who did that: https://www.pedelecforum.de/wiki/doku.php?id=elektrotechnik:forumscontroller