Honda Insight Forum banner

121 - 140 of 173 Posts

·
Premium Member
Joined
·
2,856 Posts
Discussion Starter #121
I wasn't suggesting an alternative method for testing IR; I was suggesting a potential method to see what the accuracy/precision/whatever of the PL8 IR measurements might be. I was suggesting that the PL8 has to use the cell ports and circuitry to measure the voltage response of the cells - so the IR measurements are probably limited to how good those are - and that the specs for cell balancing likely reveal that 'how good'... If the device can only balance to say plus or minus 0.003V or something like that, then that's likely the precision or whatever of the IR measurements...
AC Impedance ohms are not necessarily 1:1 correlation to DC IR resistance ohms.

I also don't know if the correlation is linear, or consistent.

AC impedance Ohms vary with frequency , shape of the conductor , nearby magnetic or electrical sources , etc.

Heck .. Aluminum windings in a transformer give a different AC impedance ohms than copper windings .. even if the two windings have the same DC IR resistance ohms.

We also don't know the software algorithm used by the PL8 as it changes the AC frequency , as the voltage changes, as the amps change , etc.

Although I agree a correlation data set / graph , with reasonable tolerance , could be made from doing a variety of tests .. I just think it would require more than just the 3 aspects of measurement of DC voltage for the PL8 .. it' accuracy , precision, and resolution.
 

·
Super Moderator
Joined
·
7,989 Posts
Really appreciate the thoughts in the last two posts. Looks like we are all agreed that measuring IR at these very small levels is a very difficult task and that the PL8 probably isn't capable of doing so with much accuracy.

I have not done a formal mean/standard deviation of the measurements I've made, but the PL8 measurements range approx. .5-1.0 mOhm. I think I'll just call the PL8 results .7 +/- .3 as normal. This will help me sort out anything at 15 mOhm or 150 mOhm - which would be a bad/questionable cell.

Thanks again guys. Putting it to bed for now and moving on:)
 

·
Super Moderator
Joined
·
7,989 Posts
You haven't addressed the real issue, IMO. I strongly suspect, from a fair amount of direct measurement experience at this point, that the significant variations in measured IR are traceable to variations in contact resistance. The fact that the measurements are unstable and "move" seems pretty good proof that something other than PL8 inaccuracy is going on. Since we know, I think, that the "true" IR of the LTO cells is very low, it seems normal to me that contact resistance would take on a major role is causing the observed erratic measurement.

Bingo! I think I finally broke this code. While testing some packs for a friend, I was again frustrated by very wide variations in the cell Internal Resistance (IR) value scatter, sometime small for an 8 cell segment, and sometimes huge. I finally noticed that if I got a large IR value in a segment it routinely occured on the first (lowest voltage) cell, which correlates to just two wires in the balance harness. I made up another harness, paying particular attention to good "hot" solder joints, but I don't think that was where the problem was actually occuring.

It turns out that tinned ring terminals acquire a fine grey oxide on their surface. I burnished the rings with some 800 sandpaper and now the IR values for the previously troublesome segment are all .65mOhm+/-.15. This is beginning to look like real data!

I can now go back to some previous troublesome IR measurement and redo them as confirmation:)
 

·
Super Moderator
Joined
·
7,989 Posts
On this Toshiba site:

https://www.scib.jp/en/product/cell.htm


The following statement is made,

"Characteristics

SCiB™ rechargeable cells can be categorized into two types: high power and high energy.
The high power cell is suitable for applications requiring the charging and discharging of large current in a short time such as regenerative braking in automotive applications.
The high energy cell is suitable for applications requiring large capacity such as electric vehicles and stationary storage systems."

I'm kinda wondering what this may or might mean for use of the Fit packs in a hybrid application - if anything.

Obviously, if the Toshiba guide was used, then the Fit packs are of the "high energy" type, while the same guide would call for a "high power" type for hybrid application.

The packs work extremely well, but it does kinda call into question the use of high levels of assist and regeneration - maybe??
 

·
Super Moderator
Joined
·
7,989 Posts
Reflecting, English units of energy are kW-hrs, and units of power are Watts. So the difference seems trivial in the case of the LTO batteries. One only has to multiply the Ahrs by the nominal voltage of 2.3V. All this seems to say that so long as the nominal voltage is known, the distinction is trivial:confused:
 

·
Registered
Joined
·
6,660 Posts
Not sure I understand your second post. On your first, on energy vs. power cells, what Toshiba says seems kind of weird, as you'd think that cells that are appropriate for a full-on EV would be fine for a hybrid. I think it mainly has to do with, simply, the size, as in the power cells are smaller and the energy cells are bigger. With bigger cells you can tweak the chemistry to make them last longer - since the size alone will handle the power. In a hybrid you'd only need and want small cells, but you'd need high power - so Toshiba makes those and tweaks them for high power...
 

·
Super Moderator
Joined
·
7,989 Posts
Not sure I understand your second post. On your first, on energy vs. power cells, what Toshiba says seems kind of weird, as you'd think that cells that are appropriate for a full-on EV would be fine for a hybrid.
I was just saying what you said, with a bit of tortured math - that the distinction that Toshiba is making in their statement seemed "kind of weird."

I think we are agreeing, the statement is quizzical. It might be an appropriate consideration if one were designing cells from scratch, but the Fit converters aren't doing that.

Further analysis of the four options they offer just adds to the confusion:



SCiB™ Cells Characteristics of SCiB™ Cells
2.9Ah High power type
New 10Ah High power type
20Ah High energy type
23Ah High energy type

Only the 2.9Ah cell is significantly smaller and lighter than the other three, so for the 10Ah "high power" for hybrids, there is no advantage over the 20Ah "high energy." Toshiba seems confused in their product line.

Perhaps it is we who are somewhat confused in what we are trying to do, but after all we have to work with what is available in the existing market. We can hardly start designing a pack from scratch.

I think I've beat it to death ;-)
 

·
Premium Member
Joined
·
2,856 Posts
Discussion Starter #129 (Edited)
All 3 are the same size / volume .. each of the 3 examples are slightly tweaked contents inside for different applications.

-----

10Ah
size mm 116 × 22 × 106
510g
~47 Wh/kg
Max 10 second peak 62C charge Rate
Max sustained 20C charge rate 0-80% SoC in 3min
Max 10s Peak 75C discharge rate
Retains ~65% of capacity under 1C charge/discharge rate at -20C temps.
At a 5C charge and discharge rates retained ~92% capacity after 20,000 cycles.

Higher power density .. Faster ~3min charge time.
Better cycle life.
Worse cold temp performance.
Lower Energy Density

-----

20Ah
size mm 116 × 22 × 106
515g
~93wh/kg
Max sustained 10C charge rate 0-80% SoC in 6min
Retains more ~80% of capacity under 1C charge/discharge rate at -20C temps.
At a slower 3C charge / discharge rate retained only ~75% capacity after 20,000 cycles.

Lower power density ~6min charge time.
Higher energy density
worse cycle life
better cold performance.

----

23Ah
Size mm 116 × 22 × 106
550g
~100wh/kg

(I could expect the above trade off trend to continue on this cell as well.)


-----

The 2.9Ah is a different size than these 3 .. but it too has been tweaked .. in favor of more power density .. more cycles .. etc .. 0-80% SoC in only ~1min .. ~90% Original capacity after ~45,000 cycles at a charger 10C charge/discharge rate.
 

·
Registered
Joined
·
6,660 Posts
^ Right. But like I said, the key issue/distinction between what's right for hybrid vs. EV, why Toshiba is drawing a distinction between power/hybrid on the one hand vs. energy/EV on the other, comes down to the size:

-In a hybrid you don't need big cells but you need high power, so the chemistry has to be tweaked to handle high power with small cells.

-In an EV you need lots of energy, big cells (or a big pack) and you also need high power. The high power comes by virtue of having big cells - so you can tweak the chemistry to accommodate the other things while still having the power density to handle EV needs.

In other words, Toshiba seems to be assuming that the appropriate cells for a hybrid would be small. And that makes sense since most hybrids do indeed have relatively small cells. If you decide to use big cells in a hybrid, like the Fit 20Ah ones, you still have the power capability plus you have a lot more energy than the typical hybrid would ever need...
 

·
Premium Member
Joined
·
2,856 Posts
Discussion Starter #131
^ Right. But like I said, the key issue/distinction between what's right for hybrid vs. EV, why Toshiba is drawing a distinction between power/hybrid on the one hand vs. energy/EV on the other, comes down to the size:
Agreed.

Besides ;)
how crazy would it be for someone to try and do the 0-80% Soc in ~1min power density rates of those Toshiba power cells .. in something with as big of a battery as like a ~85kwh Tesla Model S .. that would be like over 4MW of electrical power .. man that would be a crazy big/heavy wire.
 

·
Super Moderator
Joined
·
7,989 Posts
To me the conundrum is the 10Ah "high power" cell. It weighs the same and is the same size as the 20Ah "high energy." And, since the charge time is in % of capacity, it doesn't even charge faster than the 20Ah cell - 3 min vs. 6 min for half the capacity, same charge rate. Maybe there is some slight tweak in the chemistry but I don't see it reflected in performance.

Anyway, I agree with eq1, it really boils down to size and weight for the two applications. In that context the 2.9Ah makes sense, but the 10Ah does not, unless perhaps they can sell it cheaper.

Anyone aware or a automotive hybrid application for their LTO?
 

·
Registered
Joined
·
6,660 Posts
^ hmm, that is weird. I didn't notice that before - when I was talking about 'size' I was mainly talking about capacity -- but assumed there was a physical size difference... Could those specs be wrong? I'm pretty sure I've seen real life photos of the 10Ah and the 20Ah and I could've sworn they were not the same physical size...

[later...] Well, I searched around and didn't find any images of the 10Ah cells, let alone 10Ah next to the 20 and 23. So I probably never saw actual images of the 10s, and they're likely the same physical size as the 20+. Maybe it's just easier to manufacture having a single can size...

Ian lists comparable sustained absolute charge rates for the 10Ah and 20Ah cells (200 amps), but undoubtedly the 10 second peak rates are different. Plus the 10Ah cell sustained charge rate is 20C - that alone tells you something is different. If the 20Ah cell could do a 20C sustained charge rate it'd be charging twice as fast... The nominal voltages are different (2.4V vs. 2.3V respectively). So, I think we can safely say there are bona fide differences...

BTW, it's possible that some of the limits are 'hardware'-based, not just electro-chemically based. For instance, perhaps the sustained charge rate is limited by the size of the terminals. If both the 10Ah and 20Ah have the same can size and same terminal size, it makes some sense that they would have the same physical limits - the terminals might only handle a sustained current of around 200 amps... When I was working with the 2.9Ah cells, I did some back of the envelope calculations on the ampacity of the terminals, and it came out to be in the ballpark of what the max charge and discharge current specs are for those cells...
 

·
Super Moderator
Joined
·
7,989 Posts
Anyone aware or a automotive hybrid application for their LTO?
My words failed me again;)

What I really meant was, is anyone aware of a COMMERCIALLY designed and sold hybrid automobile which use the Toshiba LTO chemistry?

We know of a couple of electric cars, the Fit and the MiEV, but no hybrids except for our conversions.
 

·
Super Moderator
Joined
·
7,989 Posts
SCiB Hybrid

So, by exploring the Toshiba links I found this:

https://www.suzuki.co.jp/car/technology/mildhybrid/

Apparently Suzuki is marketing a line of LTO "mild hybrids" for the Japanese market only. The link is entirely in Japanese. I tried a translator to find some technical specs, but could not. Perhaps someone who reads Japanese could explore it and see if the battery specs are discussed.
 

·
Super Moderator
Joined
·
7,989 Posts
Oh, I suppose. We were talking about LTOs in hybrid cars a few posts back. Like I said, I discovered that there is a Japanese market "mild hybrid" by Suzuki.

There are apparently several bus applications.
 

·
Super Moderator
Joined
·
7,989 Posts
More Capacity Testing

I've done more testing just to try to gain some insight into what the original capacity might have been, and how much one might expect it a proper BCM such as the Orion. The testing was done at 10A, about .5C, for both charge and discharge. It isn't clear yet and probably won't be, but they are currently measuring 19.25Ah +/- a bit, balanced test over the 1.5-2.7V range.

In doing this, several phenomena become more or less clear. First, the PL8 cannot maintain cell balance below 2V. Since it can furnish only about 3A of balance current through the small balance wires, this is to be expected. I doubt that the Orion can do much better. What this seems to mean is that the region below 2V should be treated with extreme care. Without active cell level monitoring and alarms, which no one yet has afaik, the region should be avoided. Imagine an individual cell well below 2V being driven hard by a current hack which is delivering 100A or more. Sound like a receipt for damage.

The upper end above 2.5V is looking much more stable. The tests so far have been in balanced mode, but I intend to run some cycles unbalanced, which is most folks current operating mode.

If one confines operation to the 2.0-2.5V range as currently seems smart, a large number of individual 8 cell segments have shown very predictable and stable capacity of 16.5Ah +/- about 250mAh.

It may be possible to wring out a little more capacity by charging to 2.6V. in unbalanced mode.
 

·
Registered
Joined
·
6,660 Posts
Thought this might be a good place to post a couple LTO charge graphs illustrating what can happen when your voltage threshold for charging is set too high. These are a couple charges of the high power Toshiba cells, a 5-cell battery with a nominal 12V...

The top graph is a 6A charge to 13.85V and the bottom is a 10A charge to 14V. I don't have the per cell limits in place, no balancing in play, so the charger just charges to the total voltage threshold and then implements the CV portion of the charge. You can see how one cell is a little wayward (I think its capacity is simply slightly smaller), and how at the top end it ends up getting its voltage pushed quite higher. No idea whether this matters or how much; it probably matters in the long run and as a general rule something you'd want to avoid...

83561
 
121 - 140 of 173 Posts
Top