Sanyo 2170 battery NCR20700B

has anyone started a build using the Sanyo NCR20700B 2170 cells yet?

they’re still kind of expensive but would be nice to reduce parallel packs to make a more compact battery.

their continuous discharge rate is rated at 10a, is that too low for a eboard?

this is a pretty good writeup on the cells:

They are rated for 15 A continuous discharge.

They do have an equal energy density as with high capacity 18650s.

They are currently cost / kWh wise worse than the 18650s available at the moment.

To put it simply. Not too sensible choice before the prices drop.

A comparison for a bulk purchase of 100 pieces of (prices taken from NKON, without VAT): Panasonic/Sanyo NCR20700B 4250 mAh 15 A Samsung INR18650-25R 2500 mAh 20 A Samsung INR18650-30Q 3000 mAh 15 A LG INR18650-MJ1 3500 mAh 10 A

Sanyo: 492 €, 696 Wh/l, 322 €/kWh Samsung 25R: 244 €, 540 Wh/l, 271 €/kWh Samsung 30Q: 298 €, 641 Wh/l, 275 €/kWh LG MJ1: 306 €, 756 Wh/l, 242 €/kWh

At the moment there are cheaper options available in the 18650 selection. Once the price for the Panasonic/Sanyo/Tesla 20700 comes down, then it’ll make sense.


@SimosMCmuffin Good comparison of cells. IT does point out that MJ1 is currently the best cell cost wise, eur/per kwh (when ordering from nkon)

It’s one of the best “bang for your buck” cells available at the moment, but is restricted by it’s output current, so you might need to add more parallel cells to get your rated/wanted current and that’ll end up in a bigger battery pack overall, but you’ll also go further.

It all really can be boiled to:

  1. How many cells do you put in your pack, so it fits on your board. (overall pack size)
  2. How many cells do you want in series. (voltage)
  3. How many amps does the battery need to provide. (current)

And then you go around looking for the cells, which can provide the current you need with the X parallel cells you decide to build the pack with and then by price point select the cell capacity, for how much you want to pay. Of course makes sense to go to the low cost / kWh options so you’ll get most range for your money.

It’s actually not 2170 they are 2070. 2170 is the new cell design by Tesla and Panasonic. :slight_smile:

yeh, that needs to be considered that MJ1 is also 10A cell.

I think the best option to ‘‘value’’ cells power output is in watts.

If you take 3.6v nominal times 10A for example, you get 36w…

Now, if you normally wish to draw 500w and at peaks 1500w, you need at least:

42 cells

In 6S 7P config, this would give massive range… and max power output of about:

70A / 1512 W.

– 10S 4P would also be possible and would still ‘‘yield’’ about 1400W of power, with ‘‘comfortable’’ output of still 400w I believe.

(even though some would consider this a bit low maybe)… but range should still be quite good.

Considering 400w ‘‘normal usage’’, with 10S system, that would translate to about 400w/36 = 11.11 A total per pack, with 2.77A per cell.

This gives us about 11wh per cell (when really drained), so maybe ‘‘usable’’ capacity could be closer to 10.5 wh per cell.

with 40cells:

440wh (close to max) 420wh (with some ‘‘reserve’’)

Would probably give ‘‘longboard’’ about ~42km of range (~26miles)

For mountainboard (15-20wh per km) This might give 21-28km of range, but then again power output might be a bit on the low side, especially if 2 motors need to be fed from the same battery.

So yeh, for crazy longboard range these cells might fit but otherwise for high power setups they might be a bit on low side.

1 Like

I’m waiting for Samsung and LG to release their 2170/2070 cells. Hopefully the price drops then. These cells aren’t really anything special tbh.

Samsung release one of them, capacity wise is “bad”, but look at the IR and discharge capability

For a 10S1P or 12S1P and a range of 12Km it would be awesome and light, but with no power compromises

Yea capacity wise that sucks. Hopefully they will have another version.

Geez 3000mah, what’s the point of the larger cubic area?

Look at the discharge current no other cell (maybe A123 26650) can do that and not suffer massive sag

All cells are a tradeoff between energy, discharge capacity and cycle life, in this case the pushed a lot towards discharge capacity, my guess is the market is tools

@Pantologist I’ve read on candlelight forums that they have a 5300mah version almost ready to market, but it sounds to good to be true and has no sources

I doubt that. I don’t think that is even possible with the same chemistries… it will have really bad discharge output too.

@Pedrodemio @Pantologist if the 21700 cell has a capacity of ~5200 mAh, it has the same energy density as the LG MJ1 3500 mAh cell, so capacity wise it’s absolutely doable, but it will most likely be pretty low discharge rate as Panto mentioned, but Tesla has so many cells in their cars overall I don’t think it’ll be a problem for them to still get crazy cyrrents.

Exactly, by even so they are still limited by battery discharge, if I’m not mistaken they peak around 20A per cell, if using a different chemistry they could break the 1MW barrier, but at cost of lower autonomy

Beside the NCR18650GA cells, I started to design battery packs with Panasonic Sanyo NCR20700B. Are more powerful than shows specs from data sheets!

1 Like

I used these cells a few months ago on a 12s2p config. Work beautifully

I use them in a 10S3P config now for around 4 Months

The big advantage over the 18650s is the spaceconsumtion: a 12Ah needs with this cell 3P compared to 4P with 18650s

if you dont need more than 40A batterycurrent then these cells are the best in my mind

Using these in a 10s6p configuration right now. Very little voltage sag.


@chaka: like your style - just double everything

whats you range with 6P? probably around 70km/45miles?

I have yet to do a full drain on the pack. I’m sure it has plenty of range. 45 miles should be reachable but I would be happy if it gave me 30 miles of aggressive riding.

1 Like