Rebuilding Gotway Battery Pack. Battery Suggestions?

[Chaz Witt]

I have a pair of 84v Panasonic battery packs that I've put a few thousand Km on. They last about 25% as long as they used to, so I am going to rebuild them. Should I just go with the same exact batteries? I was thinking of going for something a little cheaper like Samsung 18650s. I am aware that I will need unprotected cells, but I don't know much about drain rate or where to find genuine cells. Any advice would be greatly appreciated.

[Will R]

Try going to a decent vape shop. They will order big boxes of cells regularly so may know how to help you get your hands on what you need.

[mrelwood]

 

On 3/6/2020 at 7:26 PM, Chaz Witt said:

a few thousand Km on. They last about 25% as long as they used to, so

In terms of charge cycles a few thousand km is nothing to a li-ion cell. Unless the packs have been severely misused, they might just have a dead cell pair.

I’d consider investigating this angle further in case you’d be good just by replacing the few bad cells instead.

[mike_bike_kite]

Is there a guide anywhere showing how to test identify batteries plus testing, replacing or upgrading them? I recently watched a youtube video on doing this for the Toyota Prius which I found quite interesting but I'm guessing it's not the same.

[mrelwood]

There might not be a step-by-step guide, since working with live batteries like this is generally a task that shouldn’t be done if one would need a detailed guide to do it. The energy stored in the battery pack is enormous, and besides burning one’s fingers and welding tools useless, it’s a bit too easy to cause an actual battery fire.

But since you already planned on rebuilding the whole packs (a lot of work), I would expect you to have decent soldering skills and understanding of what you’re dealing with.

I would first charge the individual packs to full and then check the pack voltage. If it stays at 84V or close, there are likely no dead cells, and the batteries have indeed just lost their capacity. But if the voltage remains below 82V, some cells are at a noticeably lower voltage, possibly even dead.

Once you have opened the packs with non-conductive tools, it’s easy to measure the voltages of each cell pair. My guess is that some of them are at 0V. If the rest are within a reasonable range (like 0.5V), just get the new cells within the same voltage range and replace the old ones. Pay extra attention to have the new cells in the same orientation as the old ones, as the correct orientation isn’t always immediately obvious.

 

If the above guide is not detailed enough, it might be best that you wouldn’t rebuild batteries or replace the cells yourself.

[Hansolo]

I advise Sanyo NCR18650GA 3350mAh - 10A

 

[yuweng]

i presume you have some electronics knowledge & some experience in handling lithium ion batteries bcos any mistakes will results in sparks that you would normally see at arc welding, speaking from experience  i've rebuild two packs recently, one mine, one for my buddy, a total of 160 pcs using this. So far so good, they perform as advertise. AFAIK, they do ship to US nowadays, it'll take awhile as batteries uses sea-freight or you could also source it at your local EV shop ? Saw Jehu's video few years ago, remembered it was not cheap though.

[zoomr]

On 3/12/2020 at 8:22 AM, Hansolo said:

I advise Sanyo NCR18650GA 3350mAh - 10A

 

I'd love to see the flip side of these series units if you have a picture. Could one swap out the cells and salvage the PCB? I think Sony VTC6 cells would actually provide more USEABLE capacity (10% @ ~3.4v/cell) and a higher margin of safety (due to the 15 amp continuous discharge rating) especially for aggressive/high speed riders. Source: https://lygte-info.dk/info/batteryIndex.html

[Nic]

Have you tried sourcing replacement packs and compared the cost to building one yourself? Manufacturers get cells at really low prices, so they can build a pack cheaper than you can. I've seen replacement packs for scooters that cost less than simply buying cells and building your own.

[hyperair]

I got mine for a decent price ($200+ for 50x Sanyo GA) at batterybro.com. Rebuilt my V8 battery pack and range-tested it at 49.15km according to EUC World from 100% to 0% battery tiltback.

 

[mrelwood]

 

Quote

I think Sony VTC6 cells would actually provide more USEABLE capacity (10% @ ~3.4v/cell) and a higher margin of safety (due to the 15 amp continuous discharge rating) especially for aggressive/high speed riders.

Titan recommends the Sanyo over the Sony for applications using less than 20W per cell. 84V 1600Wh EUCs have 120 cells, putting the threshold at 2400W. If you have watched how much power the battery provides when you ride, you’ll know that the threshold is difficult to reach even for a short moment.

https://www.gettitanpower.com/blogs/news/selecting-the-right-cell-for-uavs-sony-vtc6-or-sanyo-ncrga

Also, the VTC6 costs up to 50% more than the Sanyo.

Edited March 21, 2020 by mrelwood

[zoomr]

@mrelwood Thanks for the link, that provides some good context. The important consideration seems to be the capacity of your batteries. For example, a Tesla with an 84v, 1020Wh, 80 cell battery would hit the "threshold" at 1600W, well below its 1900W nominal motor rating. You certainly have more knowledge about this than me: What exactly is the limiting factor for EUCs? I'm sure this is make/model specific, but I would think it has to be peak motor output or temperature somewhere in the system. 

[xorbe]

11 minutes ago, zoomr said:

What exactly is the limiting factor for EUCs? I'm sure this is make/model specific, but I would think it has to be peak motor output or temperature somewhere in the system.

Combination of mosfets + peak voltage + motor winding.

[zoomr]

On 3/21/2020 at 8:58 AM, mrelwood said:

Titan recommends the Sanyo over the Sony for applications using less than 20W per cell.

@mrelwood It also occurs to me that there is no mention of the voltage cut off in Titan's testing. As you probably know Gotway's 10% mark is ~3.4v with full discharge ~3.3v. Based on the info from the source I mentioned earlier, the VTC6 seems superior even down to a 10W draw per cell or 1200W from a 120 cell battery. Am I missing something?

https://lygte-info.dk/review/batteries2012/Sanyo NCR18650GA 3500mAh (Red) UK.html

https://lygte-info.dk/review/batteries2012/Sony US18650VTC6 3000mAh (Green) UK.html

[mrelwood]

4 hours ago, zoomr said:

@mrelwood It also occurs to me that there is no mention of the voltage cut off in Titan's testing. As you probably know Gotway's 10% mark is ~3.4v with full discharge ~3.3v. Based on the info from the source I mentioned earlier, the VTC6 seems superior even down to a 10W draw per cell or 1200W from a 120 cell battery. Am I missing something?

No, you are correct. At around 1200W the range should be similiar on both cells. But 1200W is still not something that can be easily averaged for the full discharge. Averaging lower than that, the Sanyo will take you further.

I haven’t been able to check or graph my battery output power or amperage since I ride Gotway, but considering that on the 16S my 100kgs were only able to reach the (nominal motor maximum) 1200W on very steep inclines, I’d expect that I still average well under the limit.

[meepmeepmayer]

22 hours ago, zoomr said:

What exactly is the limiting factor for EUCs?

Short term (instant): battery current. A bit longer term (seconds to a minute): mosfets (usually just overheating which stops you, but you can fry them or the cabling if you really go for it).

Provided there are no fuses or firmware current/power/... limits that trigger first.

[zoomr]

1 hour ago, meepmeepmayer said:

Short term (instant): battery current.

How does this relate to nominal and peak motor output? Aren't the batteries and mosfets capable of delivering more power instantaneously than the motor can use for motion? Therefore the limiting factor would be the motor rating? Sorry for hijacking this thread @Chaz Witt

[Chriull]

5 minutes ago, zoomr said:

How does this relate to nominal and peak motor output? Aren't the batteries and mosfets capable of delivering more power instantaneously than the motor can use for motion? Therefore the limiting factor would be the motor rating? Sorry for hijacking this thread @Chaz Witt

A current near battery limit can only happen at very low speeds and then one has high torque but low motor output power. Efficiency is very bad in this case - most power is used to heat the coils and the batteries...

The limit is defined (in a first aporoximation) by I motor max = (U battery no load - U back emv) /( R internal battery + R motor coil) with U back emv = kv * wheel rotation per minute.

The delivered max torque is proportional to I motor max. Max output power is (proportional to) this max torque times speed. (This gives a "hill" with its maximum at half lift cut off speed and zero at no speed and lift cut off speed)

Nominal motor power "should" be the power the motor can deliver continously (but no one kniws if the controller, etc can stand this - but seems so with most modern wheels). Peak motor output power is just possible at some speed ranges at this half lift cut off speed (which is proportional to battery voltage).

Peak motor input power can be much higher (at as written in the beginning at lower speeds) - able to destroy the controller/motor/wirings...

[meepmeepmayer]

17 minutes ago, zoomr said:

How does this relate to nominal and peak motor output? Aren't the batteries and mosfets capable of delivering more power instantaneously than the motor can use for motion? Therefore the limiting factor would be the motor rating?

Disclaimer: I'm no expert! This is how I understand it.

Every electrical component is just a piece of metal. It works till it melts from the heat.

The motor is quite big hunk of metal, or at least a few copper strands equating to a quite thick wire. Naturally it will take much more power to damage a motor than some tiny (in comparison) electronic components like mosfets or thin motor phase cables. The motor is big and can internally dissipate and shed heat much more than some tiny electronic component, too.

I have never heard of a single motor being damaged from high stress, other than the hall sensors or their thin brittle wires being killed by a short when something else fried. While there have been plenty of fried mosfets and cabling...

TLDR: Small wires melt faster than big wires. Motor is a big wire, the other electronics are smaller and die first. That's what it seems to come down to.

[hyperair]

What @meepmeepmayer said is mostly correct, but don't forget that our hub motors are permanent magnet motors. They can go into thermal runaway:

1. Internal motor temperature rises
2. Magnets get weaker at ~100°C (I think, not sure about the actual specs of our motor magnets)
3. Magnetic field gets weaker, motor constant (KV) rises
4. RPM per volt increases (higher top speed, yay!), current per Nm of torque increases
5. Motor now draws more current for the same torque
6. Go back to step 1 and repeat in an upwards cycle until your magnets are dead.

Also, a motor is a relatively massive thing (relatively huge mass) and takes a lot of energy to raise its temperature. Motor windings may generate quite a bit of heat energy, but because of the huge thermal capacity of the motor, the motor temperature still takes a long time to rise. In that time, most other components (e.g. MOSFETs which are tiny) would have failed from the high currents.