Honda Insight Forum banner
41 - 60 of 116 Posts

·
Linsight Designer
Joined
·
3,904 Posts
Discussion Starter · #43 ·
Finally got the dual cell long term test software working properly:
I've never used TDK's instrument driver with more than QTY2 supplies at once... and certainly never with supplies that weren't programmed in lock step. Turns out TDK's driver sucks... but I've now rewritten it to not suck.
 

·
vote 4 mawah
Joined
·
314 Posts
I no longer have any evidence that suggests these cells are brand new.
It turns out the amp-hour change over time is strongly correlated to time-of-day:
View attachment 98041

If we change the X axis to time-of-day and then overlay these first two days:
View attachment 98042

We can see that the capacity does in fact vary by time-of-day... almost certainly due to daily outdoor temperature:
View attachment 98043
View attachment 98045

Note: This test setup is located in the shade, and does not receive ANY direct sunlight throughout the entire day.

So basically the ambient cell temperature impacts cell capacity substantially more than I expected: nearly 2 Ah of the 47 Ah rating (~4% of total capacity).

FYI: I went ahead and verified the test equipment itself isn't causing this uncertainty: Based on the Lambda Gen specifications, the current measurement error is less than 0.3 Ah (out of 2.0 Ah)... so based on the datasheet at least 1.7 Ah of this error is from the cell under test (not the test equipment). One way to verify the measurement equipment isn't at fault is to connect an additional temperature-controlled current sensor at various times throughout the day, and then make sure the charge current isn't changing as a function of temperature. I'll do this a few times tonight and tomorrow to verify the charge current is a constant 75 amps.
If only running the tests for a month or two, currently the ambient temp range appears to be withing the Lithium sweet spot. 80 - 90 degrees F. ?
Is there already documentation on how the dynamics of Lithium cycling change as temps move further away from the sweet spot range? Or are we all still studing those changes?
 

·
Linsight Designer
Joined
·
3,904 Posts
Discussion Starter · #45 ·
There's certainly a general knowledge in the literature, but for these cells we know very little... and are unlikely to get that info from the manufacturer.
 

·
Registered
Joined
·
238 Posts
I'm 95% sure that the Ford C-MAX 5.5 Ah cells are the same as the EHW5. They're both said to be made by Panasonic. Early EHW5 cells were rated at 5.5 AH per cell (The 18S G3 module says 352 WH). The physical dimensions appear to be the same as the EHW5.

Ford C-MAX cell (info from GreenTec)
Length 4 3/4" (120mm)
Width 1/2" (13mm)
Height without terminals 3 3/8" (85 mm)
Height with terminals 4 3/16" (106mm)
Weight 9 oz

Ford C-MAX cells are a lot more common and are sometimes available new. There are aftermarket bus bar and case options designed for the C-MAX.

It appears the now discontinued generation 3 Honda Insight is the only Honda vehicle with 12S EHW5 modules. The CRZ is EH5 and all the other cars moved up to a 72 cell system. The NSX uses 4x18S welded modules and most, if not all of, the other Honda hybrids are now using 2x 36s welded modules. The NSX appears to use the same engine based motor as the other IMA style hybrids but then they added an electric front differential with 2 motors for torque vectoring.

Integrating Ford C-MAX cells into the LIBCM system might be the best way to overcome the limited number of 12S modules, especially when we look at the 60s upgrade.

There might be aftermarket parts for the C-MAX that could possibly help in building 12S packs out of the spare cells from 18S conversions.

Can we adapt this 14S C-MAX (brand new cells) kit to 12S:

 

·
Linsight Designer
Joined
·
3,904 Posts
Discussion Starter · #47 ·
Great info. I would love for someone to experiment with CMAX modules... if they actually use EHW5 cells, then we could certainly manufacture steel compression cages (just like the EHW5 modules we're using now. That's entirely in my wheel house, and I already have a sheet metal fabricator that can make them... the only difficulty is cost, but sheet metal is cheap in volume.

One great thing about making our own module is that we can get rid of the BMS connector Honda includes on their EHW5 modules... assembling the mating connector is super annoying. With CMAX-sourced EHW5 cells, we could bolt a single PCB on top that would serve as both bus bars and BMS sense leads (with a press-fit ribbon header).

If someone wants to send some CMAX cells my way, I'll tinker with them as I have time. This will remain a back burner project for a while, but it's important to have a sustainable line of cells for the LiBCM project.
 

·
Linsight Designer
Joined
·
3,904 Posts
Discussion Starter · #49 ·
That certainly looks like an EHW5 cell, but the threaded posts are quite a bit longer. Wonder if they'll fit inside the OEM battery enclosure?
 

·
Premium Member
00 MT
Joined
·
48 Posts
Moving the 47ah ordering saga to this thread to avoid muddying the waters in the LiBCM thread.

My 2 replacement modules arrived last week and I finally got a chance to look at them today. One is perfect, the second has an (admittedly smaller) nick through the black insulation.

Cuisine Dish Ingredient Rectangle Gas

This is how this particular order was shipped. I didn't remove any peanuts.

Rectangle Wood Composite material Gas Metal

Bottom of one of the modules.


So at this point I have 6 modules, 3 are perfect. #4 has several small dents, #5 has dents and a large area of damaged black insulation, and the above pictured #6 with a small damaged area of insulation.
I emailed them again to see if they'll send another (!!) module. Honestly unsure what the response will be, maybe they'll want this one back. Who knows!
 

·
Registered
Joined
·
238 Posts
Yes, you have a good eye, CMAX terminals are reported to be 1/8" longer then G3 ehw5 terminals. Is that long enough to create issues?

It's easier to shorten the terminals then lengthen.

From the jag35 ad
"G3 cells in 18s configuration.

3.6V 5Ah lithium cells, comparable to CMAX cells, with the G3 terminals being ~1/8" shorter."
 

·
Premium Member
Joined
·
2,865 Posts
Yes, you have a good eye, CMAX terminals are reported to be 1/8" longer then G3 ehw5 terminals. Is that long enough to create issues?

It's easier to shorten the terminals then lengthen.

From the jag35 ad
"G3 cells in 18s configuration.

3.6V 5Ah lithium cells, comparable to CMAX cells, with the G3 terminals being ~1/8" shorter."
A grinder will solve that problem...
 

·
Linsight Designer
Joined
·
3,904 Posts
Discussion Starter · #53 · (Edited)
Yes, you have a good eye, CMAX terminals are reported to be 1/8" longer then G3 ehw5 terminals. Is that long enough to create issues?
Based on my testing just now, if the CMAX terminals are less than 0.320" taller than EHW5 terminals, then they will fit in the OEM battery enclosure.

So yes, they will fit... but remember we do still need to:
-design all the mechanicals to compress the modules and secure them in place, and;
-design a high current PCB that connects the cells in series and also has a BMS connector, and;
-verify this all works with the existing LiBCM mounting brackets.

One benefit with this custom module is that you can fit QTY22 cells in Bay1 & Bay2 (furthest from the junction board), and QTY16 cells in Bay3 (closest to the junction board)... so that's 60S inside the OEM pack. That's exciting, but I must stress that designing the above is a LOT of work... and certainly isn't my top priority right now.
FYI: @retepsnikrep and I previously theorized such a custom module, but at the time EHW5 modules were flooding the market, so it didn't make sense then. It probably still doesn't make sense now, given that the additional parts required to make the above modules will cost several hundred dollars in raw materials... plus my paltry "will never recover the engineering effort" markup. I suppose the mechanicals for each module would end up selling for around $200/bay.
 

·
Linsight Designer
Joined
·
3,904 Posts
Discussion Starter · #54 · (Edited)
Last night it rained hard enough that muddy water bounced off the ground into the lowest power supply (GEN 20-165). I had only thought to protect the supplies from water coming from above... I certainly didn't think water would bounce up nearly three feet vertically and four feet horizontally into the electronics.
The mud deposited itself on the serial communication PCB, which corrupted the data sent to the downstream power supplies.
This caused the tester to discharge the EHW5 cell under test all the way down to 0.000 volts. Therefore, collecting additional data from this specific cell has limited usefulness. Fortunately, we were able to log QTY297 complete cycles on that cell prior to this event, so we can use that data as a compare on future runs. I haven't fully analyzed this data yet, but will soon.

Fortunately, the 47 Ah FoMoCo cell under test was not affected.

Right now I'm repairing the serial communication PCB... assuming that's successful, I'll resume testing later tonight.

Edit:
We're back up and running. Gonna go ahead and run the tester overnight on the old EHW5 cell, and then switch over to the new one tomorrow. I am well aware that having been at 0.000 volts, this increases the potential that the cell might catch on fire... that's why I'm performing this test outside.
 

·
Premium Member
Joined
·
2,865 Posts
Last night it rained hard enough that muddy water bounced off the ground into the lowest power supply (GEN 20-165). I had only thought to protect the supplies from water coming from above... I certainly didn't think water would bounce up nearly three feet vertically and four feet horizontally into the electronics.
The mud deposited itself on the serial communication PCB, which corrupted the data sent to the downstream power supplies.
This caused the tester to discharge the EHW5 cell under test all the way down to 0.000 volts. Therefore, collecting additional data from this specific cell has limited usefulness. Fortunately, we were able to log QTY297 complete cycles on that cell prior to this event, so we can use that data as a compare on future runs. I haven't fully analyzed this data yet, but will soon.

Fortunately, the 47 Ah FoMoCo cell under test was not affected.

Right now I'm repairing the serial communication PCB... assuming that's successful, I'll resume testing later tonight.
OH NO. That does sound like Murphy's law though...
 

·
Premium Member
2001 5S "Turbo"
Joined
·
12,107 Posts
Fried earthworm...yummmmy. LOL
 

·
Linsight Designer
Joined
·
3,904 Posts
Discussion Starter · #60 ·
Thanks to covid, I'm never in a hurry when ordering anything anymore...

...

I've been meaning to analyze the data from these first cells... been super busy these past few days, so haven't had a chance to even switch over to non-damaged ones.

Rough overview:
-EHW5 cell has QTY938 cycles on it and is still going strong.
-47 Ah FoMoCo cell has QTY260 cycles on it and is still going strong.
 
41 - 60 of 116 Posts
Top