Xyster's X5

[xyster]

First row done; second row in place. The green arrow shows a placeholder battery to keep the above, overhanging battery in place while the silicon dries.

[xyster]

Finished soldering. Next up is finishing the box, and securing the new pack with homemade shrink-wrap -- tape and cardboard of course!

[Lessss]

Am I misreading those batteries or do you have them in Parallel and series as one large 31 amp hour 7.4V battery?

[fechter]

Xyster, on the side of the battery pack where the cells are jumpered to put them in series, the wires running parallel, will theoretically carry no current, and are therefore unnecessary, at least from a power standpoint. I can see they help hold the pack together.

I was thinking for a massively parallel pack, it might be a good safety feature if the individual series strings each had their own fuseable link to protect against a single cell shorting. Fuses would be too expensive, but a calclulated length of a specific wire could be used as a fuse. I think they make special wire for that.

[xyster]

Am I misreading those batteries or do you have them in Parallel and series as one large 31 amp hour 7.4V battery?

Yep, it's 2-series, 15-parallel for 7.4V nominal and 33ah (15 cells X 2.2ah). This will be added to the other 18-series 15-parallel cells I've been using for about four months now on the bike.

Xyster, on the side of the battery pack where the cells are jumpered to put them in series, the wires running parallel, will theoretically carry no current, and are therefore unnecessary, at least from a power standpoint.

1) do you mean vertical parallel or horizontal parallel ? I'm not clear on of which wires you speak.

2) can you explain why these wires would not carry current? That's a shocker to me! This configuration works in other parts of the pack. But 35 amps is the max I can make this pack draw, most of the time it's around 33-34 amps. Those wires making the S curves are the wires that jumper the two packs together in series, and also connect to the charger. They are 18 gauge. The wires running straight are 12 gauge. Both ultra-flexibility silver spaghetti wire coated in silicon -- solder sticks much better to this wire I've found and it's a real joy to work with in comparison to copper.

After much research at rcgroups.com I learned I probably damaged some cells during the overdischarge event such that they've retained capacity, put lost some current capability permanently. I figured that was the reason for the ~1C max current I'm seeing. Patrick M has the same cells in a 12s8p config and reports 1.8C max.

it might be a good safety feature if the individual series strings each had their own fuseable link to protect against a single cell shorting.

I'm recall Aerowhatt having hundreds of little fuses built for his hybrid SLA/18650 pack. Problem with both fuses and wires:

1) The pack is basically done and until hell freezes over I'm not going to again desolder and resolder it!

2) Space. Where would you put this calculated length of wire?

3) What's the likelihood of an apparently good cell shorting closed circuit while operating within spec? The only cells I've blown up (so far) were via overdischarge from my own ignorance.

[/quote]

Thanks guys for the feedback and suggestions!

[fechter]

Now that I think about it, the low current wires will help balance the cells, but since the current is very low, you could use much skinnier wire, which might be easier to deal with.

I guess I was thinking some of the jumpers (ones that say "most current") could be replaced with fuse wires. Of couse, too late now for that one.

I just keep remembering that Valence video of the lithium battery that has a fairly spectacular "thermal event" after being hit. Having a few cells go off would be better than the whole pack. See the video here:
http://www.valence.com/SafetyVideo.asp

[xyster]

Now that I think about it, the low current wires will help balance the cells, but since the current is very low, you could use much skinnier wire, which might be easier to deal with.

The wires labeled in your pic as high current are the smaller 18 gauge. But, I was thinking by looping it up and down 14 times to connect the 30 batteries, the effective gauge is more like 14x18gauge, whatever that is.

I guess I was thinking some of the jumpers (ones that say "most current") could be replaced with fuse wires. Of couse, too late now for that one.

Fuse wires...like a gauge small enough that it'll burn up if there's a short more than X amps? Wouldn't that greatly increase resistance?

I just keep remembering that Valence video of the lithium battery that has a fairly spectacular "thermal event" after being hit. Having a few cells go off would be better than the whole pack. See the video here:

I watched those videos and studied this issue in-depth before choosing to construct this pack sans BMS.
I perceive the risks of a catastrophic meltdown or damaging fire much less than the risks of me being run over by a car (seriously). I've worked to minimize both risks, but just like I wear a helmet and keep on ridin', I'm not going to add a BMS or other safety system now, unless, upon new evidence, my perception of said risks should change drastically for the worse.

Below is a closer close-up with labels more clearly showing which wires go where. Do you think the loopy 18gauge is OK as far as resistance for carrying 35 amps?

Please excuse the cat fur...curious creatures they are

[safe]

So let me get this straight.

You really are not running ANY kind of protection with your battery pack?

The only "meltdown" was when you allowed the voltage to drop below the safety zone and it burned out about 25% of your batteries. (as you were riding?) Did the fire occur in such a way that you could smell it or feel the heat as you road?

Do you use any protection in your charging or are you simply dividing things up into small units and then praying each one is okay?

(I was just talking with my home insurance company today... is your home insured for a lithium battery fire?)

[xyster]

And here it is all "shrink wrapped" with vinyl electrical tape, sittin' pretty in its almost finished home.

Purpose of the electrical tape wrapping is to:

1) Press backside of wires against their tabs to decrease resistance, and chance of connection failure.

2) Secure bits of solder that sometimes flake off here and there under long-term vibration, preventing one mechanism of short circuits.

3) Provide more strength to the structure.

4) Insulate the wires.

5) Save $ and time finding, buying, cutting and joining special-sized sheets of shrink wrap

I make certain to tape only over the tabs so that the battery's vents are never occluded, and the cells have plenty of breathing space underneath.

Next up: installing pack into finished box.

[fechter]

The wires labeled in your pic as high current are the smaller 18 gauge. But, I was thinking by looping it up and down 14 times to connect the 30 batteries, the effective gauge is more like 14x18gauge, whatever that is.

That's right. It's the equivalent of welding cable. The 35 amps will be divided among all the wires, so each one only carries a bit over 1 amp.

Fuse wires...like a gauge small enough that it'll burn up if there's a short more than X amps? Wouldn't that greatly increase resistance?

If the length of each fuse wire is as short as possible, the resistance will be negligible. Again, since they're all in parallel, the total resistance will be divided by the number of jumpers.

[patrick_mahoney]

That is a much neater soldering job than I did on my, Xyster. Very cleanly done. I'm impressed. Soldering sideways is also much smarter than how I did mine. If I was going to do mine again (which I am definitely not), I would copy your method. It's safer and better than the rather goofy way that I soldered my 12s8p pack.

I did 12AWG on the power wiring and 18AWG x 2 for the charging wires.

[xyster]

You really are not running ANY kind of protection with your battery pack?

There's a main 40-amp fuse, there's the controller's 35-amp current limit, there's the voltmeter and ammeter on the handlebars, there's the DVM I take with me on long rides with which I can check any of the now 20 individual subpacks via their respective charging port. In other words, I'm the BMS...now go ahead, crack a joke

The only "meltdown" was when you allowed the voltage to drop below the safety zone and it burned out about 25% of your batteries. (as you were riding?)

2 cells ruptured, the heat from overdischarge (these cells tank fast below 3.7V resting!) also killed (or tripped the PTC) on two dozen others -- about 10% of my pack at that time. I reconfigured and resoldered the entire pack, going from 19s12p to 16s15p. Since then I've added 4s15p more batteries.

Did the fire occur in such a way that you could smell it or feel the heat as you road?

There was no fire, just heat and some smoke. I could tell something was wrong because the voltage as read by the meter on my handlebars was drooping far more than normal. According to reports on rcgroups.com, lithium cobalt fires are almost never caused by overdischarge, but by charging at too high of a voltage, charging as the wrong type of battery, physical damage like when wrecking an RC airplane, and defective manufacture like with the recalled Sony's. This fit with my experience where there was no fire.

Do you use any protection in your charging or are you simply dividing things up into small units and then praying each one is okay?

What protection would I use? Each subpack is charged individually by a single cell 1.5 amp charger -- which is a very low charge rate. Each and every single cell is balanced with every charge this way. The chargers themselves only have one voltage setting: 4.2V. It seems virtually impossible to overcharge the batteries because as 4.2V is approached, the current trickles down to zilch.

[xyster]

It's safer and better than the rather goofy way that I soldered my 12s8p pack.

Ah, perhaps, but my bike looks goofier for sure!

And the award for most prolific overuse of duct tape goes to...envelope please....Xyster!

And the award for most individual lithium cells on a bicycle goes to....Xyster again!

And finally folks, the one we've all been waiting for....And the EV Academy Award for ugliest small vehicle in the world goes to...the Producer of the new hit show Ugly EV.... Xyster!

It's a clean sweep!!! rah rah cheer cheer....


Next up tomorrow: hookin' up and ridin'

[RatoN]

...

What protection would I use? Each subpack is charged individually by a single cell 1.5 amp charger -- which is a very low charge rate. Each and every single cell is balanced with every charge this way. The chargers themselves only have one voltage setting: 4.2V. It seems virtually impossible to overcharge the batteries because as 4.2V is approached, the current trickles down to zilch.

Sorry Xyster, but that is very confusing to me. I have a big interest in Lithium pack charging...

1. You charge your pack(s) with a SINGLE CELL charger? How could this be, since many 4.2v constitute more volts, there for the charger will detect LOTS more volts, and not even begin to charge.

2. Your pack(s) don't have a charge-discharge electronic board, so how do you know when the batt pack has reched it's discharge limits?

[fechter]

If that thing bursts into flames, look what gets roasted.
Got Nomex underwear?
I didn't see any PTC's. Was that a previous pack? Those work like a fuse, but suck at high currents.

They do make much more stylish duct tape. I use black.
I used to have a roll of camoflage duct tape, great for that "military grade" look.

[xyster]

1. You charge your pack(s) with a SINGLE CELL charger? How could this be, since many 4.2v constitute more volts, there for the charger will detect LOTS more volts, and not even begin to charge.

See the 20 single-cell chargers in the picture below? Each one gets simultaneously (well close) connected to each jack on the bike for charging. Each jack on the pack goes to one, and only one, parallel "subpack" of 15 cells, in well, parallel. Each of the subpacks is strung together in series. As you might surmise, there's 20 jacks on the bike.

2. Your pack(s) don't have a charge-discharge electronic board, so how do you know when the batt pack has reched it's discharge limits?

The light on each charger changes from red to green, and the charger stops charging (well, I think it stops charging) Anyway, the end charge from 4.15 V to 4.2 V goes quite slow, so I let it charge overnight and disconnect them all at the same time when all the lights turn green. Most of the time, the light actually turns at about 4.15V, so I'm only charging to 95%, which is fine with me as the pack should last much longer that way (see the tesla motors site blog for more on this...link is somewhere here I think).

[xyster]

If that thing bursts into flames, look what gets roasted.
Got Nomex underwear?

If the entire pack bursts in flames, a roasted weenie will be the least of my worries.

I didn't see any PTC's. Was that a previous pack? Those work like a fuse, but suck at high currents.

The PTC's inside the cells. Which is another point in re the BMS, safe. A BMS won't stop a cell fire caused by damage internally or mismanufacture.

See link for battery specs:
http://www.all-battery.com/index.asp?Pa ... rodID=1643
Internal Impedance: less or equal to 180 milliohm (with PTC)

See link for charger specs:
http://www.batteryspace.com/index.asp?P ... rodID=1231

They do make much more stylish duct tape. I use black.
I used to have a roll of camoflage duct tape, great for that "military grade" look.

I see you speak fluent "cool" too A regular Fonzarelli on a scooter [/i]

[xyster]

As seen in the top frame, I connected the new 2 subpacks in series with the rest of the batteries. Then turned it on, and.... it works!
The whole pack is 60% charged, and the voltmeter reads about 7.5 volts more than before! Now on the voltmeter I'm going to have re-mark the "e" for empty again....

Well friends, it's time to go pedal the pavement. Gonna have to take it easy, no hard testing yet because of that broken spoke....

[safe]

Your total disregard for the conventions of "safety" is having an effect. Rather than me being worried about Lithium batteries the lack of fear you demonstrate tends to make me also want to "risk it" and build a pack without all the PCB's and crap that makes the things overly expensive and complicated.

Do me a favor and create a new "technical thread" and just walk through category, by category all the areas that you need to watch out for and also what is "overhyped" as a fear. (when you are done with your ride of course)

If charging is the danger area, then also talk about that.

Try to pretend you are a college professor teaching a class on "Risk taking Battery Packs and how they are Cool".

Make it theoretical and not about your specific hardware...

[RatoN]

Ok, brainstorming here…

What about, a dum-charger? (Dum= no automatic cut off, no lights, just DC current).

One that would mimic a fully charged battery. One that would give constant 4.2v from the beginning, until you unplug it, like a fully charged battery that never runs out of juice. Would it charge the other batteries, therefore the whole pack?

Because, if one battery has more juice than the others in a pack, it will give its juice to the others, right? Then all batteries in the pack have equal amounts.

[xyster]

Just got back from a short test ride. Worked as expected. The box does not hit my thighs, though it might if I keep biking and buildup big huge quads

I'll get some hard comparison data as far as acceleration, and speed@amps after I get that spoke fixed and charge the pack fully.

Your total disregard for the conventions of "safety" is having an effect. Rather than me being worried about Lithium batteries the lack of fear you demonstrate tends to make me also want to "risk it" and build a pack without all the PCB's and crap that makes the things overly expensive and complicated.

I have great regard for the conventions of safety, but I've done enough study I think of lithium batteries to better appreciate what constitutes end-user protecting tech, and what constitutes vendor protecting tech (from lawsuits by non-diligent users). Patrick Mahoney has a pack of the same cells configured 12s8p and doesn't use a BMS either. His fancy RC charger balances the subpacks of parallel-wired cells during charge as I do with cheaper, multiple chargers.

Do me a favor and create a new "technical thread" and just walk through category, by category all the areas that you need to watch out for and also what is "overhyped" as a fear. (when you are done with your ride of course)

Do me a favor and read up and study about lithium batteries like I did, then as you have specific questions, post them to the tech thread and I'd be happy to give my 2 cents along with everyone else.

What about, a dum-charger? (Dum= no automatic cut off, no lights, just DC current).

One that would mimic a fully charged battery. One that would give constant 4.2v from the beginning, until you unplug it, like a fully charged battery that never runs out of juice.

That's basically what my batteryspace chargers are. Plug'em in and with a DVM they measure about 4.2v (range 4.15 - 4.2). Fechter would know more, but since each puts out no more than 4.2v, I don't think it can overcharge lithium batteries -- the current slows to zero as the battery voltages approaches the charger voltage (there is no electromotive force when the voltages are equal to shove in any more amps).

Would it charge the other batteries, therefore the whole pack?

If you wanted to charge the whole pack with one charger, then your charger would have to 4.2v times however many llithium cells in series you have. In my case that'd be 84 volts. But the problem there is each subpack in series will not balance with the other subpacks in series (though the cells within each subpack, because they're wired in parallel with each other, will balance each other within a subpack) so you'll end up with one subpack at a significantly lower voltage then the others, and while total voltage will look OK on a voltmeter measuring the entire pack's voltage, one or more subpacks (of cells wired in parallel) will begin to overdischarge.

Because, if one battery has more juice than the others in a pack, it will give its juice to the others, right?

For lithium and lead-acid in parallel, yes. For nicad/nimh, the situation is more complicated. But lithium and lead and every other chemistry I'm aware of will not balance in series, that's why those powercheq devices for lead batteries are popular, and why lithium RC chargers have battery balancing functions, and why good BMS's also balance batteries in series during discharge.

Balancing during discharge is slightly better, but much more expensive and/or complicated to wire. Balancing during charging as I do is good enough for me because the difference at the end of a ride between subpack's voltage is very small, like 0.02 volts or less, and gets equalized whenever I recharge.

[/quote]

[knoxie]

Yo

That bike needs a wash!! check your shock absorber on the rear ha ha, does it move?

Cheers

Knoxie

[xyster]

That bike needs a wash!! check your shock absorber on the rear ha ha, does it move?

I don't know about you, Mr. City Slicker, Mr. upstanding tea-sipping wig-wearing English gent and all that, but I live smack in the middle of mud country, and I love it here! I practically bathe in the stuff

And yes, the shock absorber usually moves. To loosen it up, I just ride through a deep puddle or two.

[xyster]

I want to know how the batteries are charged,one by one ? no BMS contol it when the batteries charged

As shown in the picture below, the batteries are simultaneously charged by 20, single-cell chargers, each connected to a single, 4v subpack of 15 cells in parallel. There is 20 of these 15-cell subpacks in series for 80 volts total (at 4 volts charge).
There is no BMS as it's normally thought of. The total current is limited by the controller to about 1C (35 amps). I watch the batteries' voltage with a voltmeter on the handlebars. Each of the cells has a PTC fuse inside.

[D-Man]

Questions :

So when you pull the batterys off the charger, how many volts does a cell read when resting?

Also, what should the cut-off voltage be per cell from your experience?