Another DIY battery pack question

[wurly]

DIY battery pack questions

I was going to use 3 SLA 12V 4.5Ah as i have a fairly short commute (price £44 the cheapest i can find).
The bike i am in the process of converting is a mountain bike with a tongxin front motor. I figured i wouldn't be discharging the battery any more than 50%for my normal ride ....but, the temption to twist and go will always be there, i am beginning to think it might be false economy.

So now i am thinking 10 lithum ion cells in series will give me 37V @2.5 ah and another 10 paralleled will give me 5ah (price approx £50), which should be more than enough and no real worries about shortening the life span of the batteries should i discharge them too much.

Now the questions: How does BMS work? if i were to buy protected cells over non protected, do i need BMS? are there any circumstances where BMS is not needed. I am using the same cells for my lighting at the moment and the charger outputs 4.2V for each individual cell to be charged.
Would i be correct in assuming a 42V charger would be required for charging 10 cells connected in series?
Thanks
Mel

 

[flecc]

Yes, you absolutely must have a BMS system if you don't want to end up with a fire while charging or premature failure of the cells while using the bike.

Lithium cells are potentially very dangerous and have to be continuously managed during both charging and use, some of the battery cost being in that included circuitry. The BMS monitors the cells during charging, preventing imbalances and providing limiting where necessary, shutting off cells one at a time if necessary until a full balanced charge is achieved. During riding, the BMS carries on the monitoring to avoid any damage occurring and also provides two kinds of cutout, the regular cutout before a damagingly low cells voltage is reached, and a cutout for any unusual circumstance of discharge runaway.

If you use the link below to have a look at the internals of an eZee lithium battery on my website, you'll see how all cells have their own individual monitoring under the control of a master board. Some batteries have all the wiring into a single enclosed BMS box, but the principle is the same.

In short, unlike NiMh and NiCad, lithium batteries are not suitable for amateur recelling, though complete content packs can sometimes be used.

eZee lithium battery

A 42 volt charger is correct for 10 cells of 3.7 volts nominal rating.
.

 

[wurly]

Ok thanks for that info Flecc.
I figured it might not be that easy. I can't find anywhere selling a BMS pcb ready to go, so i'll knock that idea on the head.The thought of only having to use 20 cells seemed a good reason to explore the idea.
So now i turn my attention to using NiMh instead. I found some 1.2V 8000mah D cells. 30 cells to give me 36V 8ah. Total cost with delivery £72 (hopefully!). Seems reasonable for a 8ah battery pack and i figured i'd charge them 10 at a time with a 12V charger. My only concern with having so many cells in series would be a dud cell possibly reducing the output and if one did...how would i know and find it???
Also, i read on one site that it's advisable to 'zap' the batteries and recharge them to assure they are of all the same consistancy. Is there anything in this? And is it something i should do before i use them?
Again...i'm looking for more advice here. Anything i should be taking into account?
(I'll use SLA's as a last resort)
Mel

 

[Footie]

If you use Ni-Mh batteries you won't need a BMS.
I started tinkering with building a battery using AA cells. I went for blocks of ten cells that can be charged separately using a Ni-Mh charger (£14) designed for charging RC packs. The chargers low cost means one could buy three and still pay less than a charger. These chargers will charge blocks of ten cells at 12 volts max, so they should work with 10 D-cells (you would have to build your own blocks). Proper Ni-Mh pack chargers are about £100+.
The main disadvantage with my pack is it has to be dismantled to be charged. But as a spare battery it's not a problem. Just connect to a block, wait for the green light, then connect to the next block - easy. If a block won't charge then a cell has died, so its easy to spot and swap dead cells. My posts on this are tagged onto this thread -
http://www.pedelecs.co.uk/forum/battery-faqs/1398-making-your-own-battery-pack.html

flecc has also built an AA battery and his fully working/tested battery is far better than my project. It even uses his Ni-Mh charger.
http://www.pedelecs.co.uk/forum/battery-faqs/642-using-aa-cells-e-bike-battery.html
.

 

[flecc]

Most of todays NiMh cells no longer have adequate discharge rates for our e-bikes and tend to starve the motors of current, giving lower power. The special high discharge rate 8.5 Ah to 10 Ah cells that were previously used have largely gone out of production since lithium took over the e-bike market.

The AA cells discharge rate problem is a different one, and it arises from the pursuit of ever bigger capacities right up to about 2800 mAh now. The higher the capacity, the greater the cell content density, in turn giving high internal resistance which impedes current flow.

That said, the best 8.5 Ah cells will still allow reasonable performance, but the lack of available information means almost all purchasing is a shot in the dark. Batteryspace in the USA are still able to supply high discharge rate 10 Ah cells and packs, but the present exchange rate makes their purchase very expensive.

One aside following your mention of 20 cells with lithium. A 37 volt lithium battery only needs 10 cells since each cell is 3.7 volts nominal, and a 26 volt batery uses only 7 cells. There's also a different type of lithium cell of 4.1 volts nominal, but these are rare.
.

 

[wurly]

Flecc, my intension was to connect 2 banks of 10 lithium cells in series hence 20 cells.
This is the only info i can find on the D cell i am looking at. Doesn't tell you a lot as regards it's maximum discharge abilities. I guess the only way is to buy one and test it.
I should say that i'm going to be using a 250w tongxin non hall and intend to do quite a bit of pedalling. I read somewhere on this forum that someone using the same motor didn't often see current reach much above 8A. This is my first conversion so cannot comment on that, although i can also see that spare capacity in reserve would be a good thing....(SLA's maybe).
I have seen diagrams listing 1C 2C 3C etc for discharge rate. This diagram is listing 1CA. Can someone explain, does it mean temperature over current?

Footie, i've been following your battery post with great interest. How is it performing? do you have any photos?

 

[flecc]

Those cells should be ok with that motor as it doesn't seem to be too greedy from what you say. The trouble usually comes in with the high powered motors demanding up to 20 amps at times.

The discharge rate such as 1C is to indicate the effective capacity when discharging at that number of times the standard charge rate C. At 1C you get 95% of nominal capacity according to the spec. The higher the discharge rate, the lower the effective capacity you get. The temperature isn't part of the expression, though it can be included on graphs.

I think your Tongxin will still be a Hall effect motor, you may be referring to whether it's a sensorless emf feedback one or the model with Hall sensors for feedback, but I'm not clear on that from what you've said.
.

 

[Footie]

I'm embarrassed to say, progress with the battery has sadly not happened. The bad weather and no workshop means I either work outside (rain and cold) or try to take over the kitchen – much to the wife’s annoyance. The battery powers the motor but that’s as far as I’ve got. The kit has sat in the store shed since it arrived, again the lack of a workspace hampering things.
I did post a picture of the blocks on my picture site.
DIY NiMH Battery on Flickr - Photo Sharing!

Flecc's much more of an expert on batteries than I ever will be, but with regards to his comments about Ni-Mh’s having inadequate discharge rates for e-bikes. If that’s the case then what about RC packs?
Would they not be a way of over coming the shortcomings of inadequate discharge? After all these packs are designed to be drained flat.
The component-shop has some large RC packs - 12.0V 3700mAh or 12.0V 4300mAh.
Large Packs

If I’m not mistaken they can also supply them with twin connectors, so that the battery can be fast charged whist still connected. Three 12.0V 4300mAh would give a 36v 4.3Ah battery would they not? May be worth investigating for a short commute use.
.

 

[wurly]

Update on my battery pack:
I have an option of buying thirty 1.2V D cells 10000mA. Approximate is cost £86 including delivery. I'm thinking...yes it's a 'shot in the dark'...but if i find they are adequate e.g. they supply enough current for my hub motor needs, then i have found my 'el cheepo' battery solution, if not, i have learned the hard way.
Unfortunately they are not tagged, so connections between cells may suffer higher resistance than i would like. To enable connections between parallel (charging) and series (riding) I would build 3 off 12V packs of 10 cells in a tube construction with a screwtype end to push the cells together for a nice tight surface contact. 4mm sockets either end for connections.
Not sure of the exact specifications on the batteries so i may look at similar type cells for recharging characteristics. I was hoping to use a standard car battery charger with a current limiting power resister if necessary.
Does the panel have any thoughts before i make my purchase? There is still the option of SLA's...(actually,I did a weight check and i think the 30 cells maybe heavier than three 5ah SLA's, but at least i will have more power in reserve).

Flecc
It is a three wire motor.
I put a scope on one of the windings of the motor. The waveform indicated a long slope(rise and fall) squarewave shape of 89hz freq with PWM freq of 8.33khz inside that waveform. I am not sure of the characteristics of a signal produced by back emf, but there was no sign of such a signal that i could see, so i presume it is a very simple analog controller with power dictated by the mark/space ratio to the windings. From what i can figure out, because of the gearing these motors have good freewheeling abilties in one direction and not the another, so i guess the direction need not be controlled electrically?

 

[flecc]

Using a car charger wouldn't really be suitable for NiMh cells. You really need to use a thermistor in the battery and have a dedicated 36 volt generic NiMh charger which would cut the charge when the thermistor signalled the temperature rise nearing the end of charge. Relying on voltage cut-off isn't very reliable with NiMh.

NiMh is still the better bet for the battery than SLAs though, since the 5 Ah hour rating on lead acid will only deliver about half that as usable current, where the NiMh cells are capable of delivering most of the charge down to 1.1 volts per cell. The range with 5 Ah SLA would be very poor in comparison.
.

 

[flecc]

Just saw your addition about the motor. I'm not sure what model that is but it may be the Tongxin Nano type commonly used now which relies on back emf to signal the controller to allow power from a standstill. It's a rough and ready method that does cause a bit of snatch on take-up, but some designers, Cytronex for example, recommend that these motors are not used in that way but as "pedal off the mark" motors. That of course takes care of the direction of drive, and it wouldn't matter so much what controller was used in that circumstance. As you indicate, I don't think you'd have a drive direction problem anyway
.