PINNED - Lithium battery conversion/info

[LROBBINS]

Those hex screw heads look a bit lower than socket heads. Are the screws brass or copper? Having everything basically in Cu, even if the screws are brass, should keep the resistance of the pack about as low as possible. I like that.

[expresso]

Looks nice - what volt is that pack and how big is it - what will you use it for - Wheelchair also ?

i might have spoken too fast about making another two part pack - i have to figure that out first - layout - connections - arrangement of cells for 12V -

[shirley_hkg]

Those hex screw heads look a bit lower than socket heads. .

They came with the battery , and yes , they are copper .

The head is only a mm thick , aim at keeping battery ends close.

[shirley_hkg]

I love that too .

[ex-Gooserider]

My recollection is that if you use the 12AH cells, there is a combination that works out to almost exactly the dimensions of a standard Group 24 lead brick.... IIRC, it is 4P 6s, arranged 6 cells long by 4 high (If I'm wrong, I'm sure BM will let us know...) One then wires up the main power just like the lead bricks, and then you just need to figure out how to do the balance wires... My first inclination would be to have a main balance connector harness that spits into two smaller connectors for the harness on each pack....

Given the way they do the batteries on the Storm chairs, if / when I do lithium on those, I will have to do a setup of this sort, since I basically just have a plastic box for each brick that slides into the frame and connects with Andersons.

ex-Gooserider

[Irving]

Think its 4S5P

[Andrey]

4 weeks ago I charged cells to 100% until charging current dropped to 0 (~30 min after hyperion says it's done). After that I discharged 1000mAh from each cell.

Now 4 weeks later I'm charging the cells again to 100% and it only puts in 950-960mAh. How is it possible?

TL/DR: Charging cells after 4 weeks of rest and it puts in less Ah than was discharged. Where did I make a mistake?

[Scollard]

Here's the problem Andrey. There is some error in all the measurements the charger makes. By only waiting 4 weeks the self discharge is probably well under 100 mAh. I waited 6 weeks and my measurements were all between 0 and about 75 mAh. You discharged 1000 mAh, and now your trying to measure say 25 mAh against the 1000 mAh you removed. The margin of error for your process is larger than the number of mAh lost to self discharge. Contrary to what BM suggests, unless you are willing to wait longer for the self discharge (maybe 8 weeks), you should just fully charge the cells and let them sit 6 weeks. Yes it's not ideal for the cells, but then again the factory ships them fully charged and they have already been sitting like that for longer than it will take you to determine the self discharge rate. So my advice is to charge them right up and let them sit 6 weeks fully charged. Then recharge them and you will measure the self discharge rate.

[Burgerman]

Its not that it makes an error, but that the charge ends with the beep, at a small current. And stops counting. But remains on a 3.600v level adding more charge. Soyou removed 1000mah when the cell was completey full. But you only charged to say 95% at the beep... And thats when the count stops.

Makes no difference to the results, as the figures you are interested in is the DIFFERENCE returned to each cell.

[shirley_hkg]

It is a 12V 9P4S block, measures 250 X. 250 . Two of these 12V blocks will be put into an INVACARE ARROW G24 battery compartment.

[expresso]

nice looks huge - thats a direct fit in a group 24 size ?

would you be able to take a picture of the under side of that ?

thanks

[Scollard]

When I mention error what I am actually referring to is uncertainty. These chargers are not very precise. At best they probably have no better than 5% precision (e.g. a high quality Fluke multimeter runs around 3% uncertainty). That means when you remove 1000 mAh from the battery it is somewhere between 950 and 1050 mAh. Now if you let the batteries sit for 4 weeks and they self discharge between 0 and 50 mAh, what you are measuring when you charge them up is the uncertainty (variable error) of the charger. So your best way to minimize that uncertainty is to charge them fully and then let them sit 6 weeks and then charge them fully again. If you take out 1000 mAh you will probably have to wait 12 weeks before the self discharge becomes significantly larger that the uncertainty of the charger.

[Burgerman]

When I mention error what I am actually referring to is uncertainty. These chargers are not very precise. At best they probably have no better than 5% precision

Definitely not true.

You can see that for yourself, test the same cell 10 times over and not wait between cycles and get the same figures for recharge within say 3 to 5mAh over many 2 to 10 Ah cycles. If theres an error its low and the same every time. So for these tests its pretty accurate. If it was only measuring Ah to within 5 percent then theres no point doing the test at all!

Cant find the data for the Hyperion now, but its similar to the PL8 below and repeatability willbe better than the figures below.
Worst case:

Voltage calibration: Cell voltage measurements are factory calibrated to a standard traceable to NIST; calibration is to +/- 5 mV
Current calibration: Charge current is factory calibrated on a 4A standard; calibration is to +/- 1mA
Measurement accuracy:
Voltage resolution: 78uV (16 bit)
Voltage tolerance: +/- 5mV
Charge current: +/- 1%
Capacity added to pack: +/- 1%

[Scollard]

I am not wrong. It's very complicated and only a very small number of people on the planet will understand why I am not wrong because calculating uncertainty is just not something the average person knows about. That +/- 5 mV is, at best, determined on a 100 mV scale (+/- 5%). But more likely it is determined on a 10 mV scale (+/- 50%). Uncertainty (error, coefficient of variation, whatever) becomes larger as measurements become smaller (Horowitz Trumpet, for those of you who want to Google it). The hobby charger does not count electrons, it is measuring Amps (with an uncertainty of 1%, probably underestimated because scale is undefined. If I had to guess that 1% is probably at full scale when it's pushing 40 A, so we're talking +/- 400 mA) and it is measuring voltage with an uncertainty of ?. Not only that but it is making 2 measurements (Amps and Volts) on differing scales all with differing uncertainties and then calculating mAh. A 5% uncertainty would actually be quite impressive. It's probably larger. But hey, it's a hobby charger, it's not a life and death precision instrument.

[Burgerman]

If I take out 1000mAh from my headway cells with the PL8 and replace it repeatedly I get 1004mAh returned +/- 1mAh time after time. Which is one hell of a lot better than a 5 percent accuracy level! The actual amount isnt relevant. Its the repeatability.

Test it yourself.

[JoeP]

Best way to punch busbars is to rent or buy yourself a hand operated punch.

Something like this?:
'Neiko' 8600 Hand Power Punch
https://www.youtube.com/watch?v=MVR9aEvgG5w

[shirley_hkg]

I need 70 bars @ 6 holes, so that is 420 strokes. Besides, the hardest part is the exact spacing ; the more holes on each bar, the more demanding job.

To my surprise that they use CNC to make me just 70 pieces. The precision astonished me , that I should have been more picky to order 6.5 mm hole instead of 7 mm.

Price is only $0.8 a piece for pure copper. Things get easier here in the FACTORY OF THE WORLD.

Any one need these ? I can help.

[Irving]

To my surprise that they use CNC to make me just 70 pieces. The precision astonished me , that I should have been more picky to order 6.5 mm hole instead of 7 mm.

Why surprised? It's CNC that makes this sort of thing possible - doing it by hand would be prohibitively expensive and no where near as accurate. Setting up a CNC machine to drill holes at regular intervals in a bar is a very simple exercise, probably no more than 15min to write and test the code; it'll take longer to set-up the workpiece and run the program. My own CNC mill would take about 2min per bar plus set-up time and I'd probably clamp 10 bars up at a time, taking may be 15min, so 25min per 60 holes

[JoeP]

Hi all,
I am starting to assemble a sourcing list for the materials required to start an add-on LiFePO pack for my aging TDX-SP.

If the plastic black covers shown in the link under 1) below are to be used, it seems to me that the proper type of Allen bolt to buy for LiFePO cells would be the plain barrel type shown under item 2) below. Am I right?

On the other hand, it looks like the flanged button type (3 below) would give the greatest contact.

Joe P
(sorry, for forcing you to click links, but I have not figured out how to add images to my posts)
(Shirley, those CNC cut busbars look terrific)
********************

1) BLACK SOCKET HEAD COVER CAPS - For Headway LiFePO4 Cell 4mm hex screw http://www.ebay.com/itm/182087848890?_t ... EBIDX%3AIT

2) Plain barrel type Socket Allen Head:
M6 (6mm) DIN912 Allen Bolt Hex Socket Head Cap Screws A2 Stainless Steel?
http://www.ebay.com/itm/M6-6mm-DIN912-A ... xlcPAa0swg

or
3) Button Head Socket Cap Screws?
http://www.ebay.com/itm/121633184230?_t ... EBIDX%3AIT

or
4) M6 A2 STAINLESS STEEL FLANGED BUTTON HEAD ALLEN KEY BOLTS HEX SOCKET SCREWS?
http://www.ebay.com/itm/321231731068?_t ... EBIDX%3AIT

[Burgerman]

I use stainless because it doesent corrode, button head because they are shorter and neater.

[JoeP]

Hi again all,
I think I have tracked down the SubD hardware at Digi key here in the US.

Regarding the balance wiring that will run from each of the eight "cellblocks", I tracked down a 10' piece of Belden 9 Conductor 22 Gauge Solis Unshielded Cable (BELDEN PN 9577 9). See http://www.ebay.com/itm/262348067703?_t ... EBIDX%3AIT .
I was figuring an using about 4 feet of that cable.
Am I on the right track here? Is that cable appropriate?

Joe P

Digikey Part # Description
NOR1126-ND --- IP67 25 POS CONNECTOR COVER --- $10.26
NOR1189-ND --- CONN DSUB PLUG 13POS STR SLD CUP --- $ 12.80
NOR1191-ND --- CONN DSUB RCPT 13POS STR SLD CUP --- $17.47
952-25SCE-ND --- BACKSHELL 25POS STRGHT DIECAST --- $15.08

[Sorry, the web links to the Digikey part #s did not come through.]

[Burgerman]

Cable is solid.

You really need the 36 inch pl8 balance lead extention cables.

[expresso]

Hi joe - looks like your going to go for the ADD On - very good - have you decided if you are going to do 2 rows - or 3 rows ? 12ah Cells or 15ah Cell ?

[JoeP]

Hi joe - looks like your going to go for the ADD On - very good - have you decided if you are going to do 2 rows - or 3 rows ? 12ah Cells or 15ah Cell ?

Cable is solid.

You really need the 36 inch pl8 balance lead extention cables.

PL8 balance lead extension. Got it. I had thought the wire in the ribbon cable might be a pretty small gauge. But if the Revolectrix lead is enough that's good.

If it were possible to do a full conversion on my TDX-SP I would think that would be the way to go. But I have a post type elevation actuator in the middle of the battery box and I just can't figure how you would squeeze sufficient cells in the existing battery box. Perhaps somebody out there knows something more about this.

Espresso, glad to know the Allen key heads now come with the cells. That solves that problem.
JoeP

[Burgerman]

One advantage of building custom lithium packs is that the battery can be any shape you want. This one fits my rebuild decade old BM2 chair - now green... Its an odd shape to miss seating, power module, rear light etc, while allowing the high parts to sit inside of the seating base. To maximise fitted capacity.

download/file.php?id=4241&mode=view in two halves unwired...

download/file.php?id=4301&mode=view pushed together and wired...

download/file.php?id=4460&mode=view fitted...
Much of which reaches up to the seat top plate almost (to within 5mm in places).

Its a massive 3 plus kwh lithium that allows about 50/55 miles or better range on a 6mph (7.4mph with the bigger wheels) chair.

[shirley_hkg]

It has 2 22NF lead bricks inside. Head room is enough for G24.

100Ah is quite possible , I guess.

[Burgerman]

I never bothered to sit and figure out the best solution for 22nf because I never had to. But there must be some cell, or arangement that gives at least 80Ah lithium.

[JoeP]

I never bothered to sit and figure out the best solution for 22nf because I never had to. But there must be some cell, or arangement that gives at least 80Ah lithium.

Unfortunately, my searches for a drawing showing the dimensions of a typical TDX-SP have not turned up anything.
I have techy who owes me some work. Perhaps I'll have him remove the batteries and take some physical measurements.
Joe P

[JoeP]

The TDX owners manual for a 25.5" base states that there is ordinarily room for two 22NF batteries or two Grp 24 batteries.
I have the 24" base.

[flagman1776]

(Posted above while I was typing)

Do you mean BCI Group 24 which has nominal dimensions of L= 10.5", W=6.75", H=8.75" as compared to
22NF size with dimensions L=9.04", W=5.45", H=8.15"
Sometimes the equipment has a bit of wiggle room but that can only be told by examination & measurement.