Honda Insight Forum banner

1 - 20 of 154 Posts

·
Moderator
Joined
·
6,918 Posts
Discussion Starter #1 (Edited)
With less than a week until I install Ron's newly developed prototype pack, I'm trying to gather as much information about the stock pack as I can. With the help of Peter's BCM Gauge, I have been logging the performance of my pack during a specific test run.

The run is a 4,000' climb over about 15 miles. I try not to use regen while going up, and assist while coming down so as not to complicate the numbers. Using MIMA, this is pretty easy.

Wireless BCM Gauge Receiver and Data Logger

Ready to log data

What I'm finding is very interesting, and I wish I knew if it was normal or not. We really need more people logging data like this. My pack is about 1.5 years old, replaced under warranty. It has never been grid charged.

During this first run, I experienced a recal for the first time. I was quite surprised. The recal happened at about 1/3rd left on the SoC meter. I was using PIMA, and was very aggressive with assist overall.

Data:

Mountain Run #1 - Up - Volts, Temp, SoC, Amps

Mountain Run #1 - Up - Loaded Volts Vs. Amps With Voltage Trend

Mountain Run #1 - Up - Amp Hours With SoC Overlay

I was able to get a whopping ~2.5Ah out of my pack.

Trip Down Data:

Mountain Run #1 - Down - Volts, Temp, SoC, Amps

Mountain Run #1 - Down - Amp Hours

As you can see, I inputted about 3.18Ah back into the pack.

Test #2:

This trip I was much more gentle with assist. The SoC gauge behaved normally with no recals. It dropped steadily down to 3 bars, hung there for a bit, then dropped off to one bar before limiting MIMA - just like normal. The graphs look very similar, however the gauge behaved how I have come to expect it to behave. The SoC drop off is when the 3 bars finally went to two.

Mountain Run #2 - Up - Volts Temp SoC Amps

Mountain Run #2 - Up - Loaded Volts Vs. Amps with SoC And Voltage Trend

Mountain Run #2 - Up - Amp Hours

Despite "feeling" like I got much more out of the pack, and the SoC gauge behaving normally, I was surprised to see that once again I only got ~2.6Ah out of my pack.

Mountain Run #2 - Down - Volts Temp SoC Amps

Mountain Run #2 - Down - Amp Hours

Once again, just over 3Ah inputted. I'm beginning to notice a trend. But this can't be normal.. can it? Is my pack already that deteriorated after only 1.5 years? I haven't noticed any anomalous behavior with it in the year and a half since I got it. Very strange. So.. is this normal?

I'll continue to update this thread with data that I get. I would like to do a grid charge and BCM Reset on my pack to see what that does to these numbers.

If anyone wants to look at the raw data and my calculations etc, feel free:

Mountain Run #1 - Up XLS

Mountain Run #1 - Down XLS

Mountain Run #2 - Up XLS

Mountain Run #2 - Down XLS

Edit: More Data:

Mountain Run #3 Up:

Mountain Run #3 - Up - Volts, Temp, SoC, Amps

Mountain Run #3 - Up – Amp Hours Out

Mountain Run #3 - Up – Volts Vs Load with Voltage Trend and SoC

Average: -9.35A over 1226 seconds

Mountain Run #3 Down:

Mountain Run #3 – Down - Volts, Temp, SoC, Amps

Mountain Run #3 - Up – Amp Hours In

Average: +8.57A over 1464 seconds

Mountain Run #4 Up:

Mountain Run #4- Up - Volts, Temp, SoC, Amps

Mountain Run #4 - Up – Amp Hours Out

Mountain Run #4 - Up – Volts Vs Load with Voltage Trend and SoC

Average: -9.23A over 1045 seconds

Mountain Run #4 Down:

Mountain Run #4 – Down - Volts, Temp, SoC, Amps

Mountain Run #4 - Up – Amp Hours In

Average: +8.84A over 1464 seconds
 

·
Administrator
Joined
·
10,623 Posts
Excellent idea to start a new thread for this.

Once you have done a grid charge on this old pack it will be interesting to see the difference.

Once the new pack arrives don't change anything with the gauge so we get consistent data into the spreadsheet. I suggest you grid charge the new pack overnight with the fan running when it is first installed to make sure it is balanced before you start.

With your data logging speadsheet just a though about how it works.

Lets assume you get 10,000 samples on a trip. If you total up the samples where the current is positive and divide by the total number of samples you will get the average Positive current (regen) for the trip. Ditto with the negative.

From the time stamp you can work out the elapsed seconds then as you say do some maths. I'm just trying to make sure we get the right result for AH.
 

·
Moderator
Joined
·
6,918 Posts
Discussion Starter #3 (Edited)
Yeah, me too. Like I said I'm not good at math at all, and usually rely on cross-checking to determine whether my answers are correct. They seem to be though. Go ahead and check out the spreadsheets, I added them to the post.

Basically I broke each sample up into even slices of time. Example: If the data set is 30 minutes and encompasses 10,000 samples, each sample is

30 minutes * 60 = 1800 seconds / 10,000 = 0.18 seconds/sample. You can see that this matches up well with the actual timestamps.

Once you have time, you can calculate Amp hours.

I wonder how much the missing tenth is skewing results though? Probably not more than a few hundred mAh overall?
 

·
Administrator
Joined
·
10,623 Posts
OK Try this vesrion 3.20 specially for the logging. The amps column is now the full resolution from the sensor to two decimal places. You need to divide the value by 100 to get the amps before the decimal point.

The no data message has been removed to fit it into code space.

Edit

I received my pair of transceivers today :)
 

Attachments

·
Registered
Joined
·
821 Posts
Pack capacity

I think 2.46 AH from a pack which has never had a gridcharger may be reasonable and representative, 2.46 AH probably represents the weakest stick in the series of 20.

April 2009 I removed my pack to benchcycle the sticks, for first discharge at 10 Amps, the best stick yielded about 2.5 to 3Ah, the worst, only 400mAh !!!!.

They all recovered, even the 400mAh duffer, though it took extra cycles to bring it into capacity line. Two years later, thanks to grid charging, my pack still performs like new.


It is staggering to see the variation in capacity and therefore balance of the pack prior to the bench cycle, I will have to dig out the spreadsheet and post on IC.

From my exercise I noted that for 10 amp charge and discharge between limits of 5.4V to full volts, for discharge / charge ratio the figures came out as 89%,
 

·
Moderator
Joined
·
6,918 Posts
Discussion Starter #6
I suppose, it's just a little surprising. I consider my pack to be healthy, other than the anomalous recal on the first data logging run - so I was expecting to see something around 4Ah, which seems to be the number thrown around as conventional wisdom.

Ron's new pack is arriving today, and I need to put together my grid charger.

I've also not played around with flashing the BCM Gauge PIC to different code. I'm very interested in having higher resolution on the amp data, so I will try this ASAP.

All this along with trying to coordinate the payment and transport of Insight #3...Too much on my plate, I need another 4 day weekend..
 

·
Registered
Joined
·
2,425 Posts
With less than a week until I install Ron's newly developed prototype pack, I'm trying to gather as much information about the stock pack as I can. With the help of Peter's BCM Gauge, I have been logging the performance of my pack during a specific test run.
Eli, can you identify whether the battery fan is on or not and its' speed? It would be important to compare how much forced ventilation is needed.
 

·
Moderator
Joined
·
6,918 Posts
Discussion Starter #8 (Edited)
I could, yes. The data won't be incorporated into the spreadsheet automatically unfortunately, so I'm not sure how much it will be worth.

You can see the fan cycling on and off during the Mountain Run 2 Down graph, if you look at the temp.

Hmmm. I could use a watt-hour meter to record the total amount of energy consumed by the fan. That won't give high/low fan speed, but it would give an overall measure of how much the fan was running during the data logging session? Might be more useful than anecdotal notes on the fan's behavior at least.
 

·
Registered
Joined
·
2,425 Posts
Hmmm. I could use a watt-hour meter to record the total amount of energy consumed by the fan. That won't give high/low fan speed, but it would give an overall measure of how much the fan was running during the data logging session? Might be more useful than anecdotal notes on the fan's behavior at least.
It would be better than nothing. It would give us an indication of the temperature performance under load.
 

·
Moderator
Joined
·
6,918 Posts
Discussion Starter #10 (Edited)
Got the grid charger wired up last night.



To Do list:

1) 3rd data logging run up the mountain tonight
2) Grid charge overnight tonight, reset BCM
3) Install Peter's latest BCM Gauge software with higher resolution ampere logging
4) 4th, 5th, 6th data logging runs on OEM pack
5) Install Ron's new prototype beta pack
6) Begin logging data on new pack

We'll see if I can complete this by the end of the weekend.
 

·
Administrator
Joined
·
10,623 Posts
To Do list:

1) 3rd data logging run up the mountain tonight
2) Grid charge overnight tonight, reset BCM
3) Install Peter's latest BCM Gauge software with higher resolution ampere logging
4) 4th, 5th, 6th data logging runs on OEM pack
5) Install Ron's new prototype beta pack
6) Begin logging data on new pack

We'll see if I can complete this by the end of the weekend.
That's a good effort Eli you deserve a pat on back. This is very interesting research at a critical time for our cars with these new cells. A comparison is critical and your long term data will be vital.

Don't forget depending on the soc and any imbalance in your current pack when you start the grid charger it might take >18hrs to fully balance. I suggest use MIMA to regen and max out soc before you then plug in the charger which will shorten the time considerably. If you are seeing around 3ah of capacity then allow at least >12hrs to balance.

Edit
When you upgrade the gauge software don't forget to export the old code first, then if it doesnt work you can roll back to your current version. Joe made a nice video on the other thread about software flashing.

Re the battery fan the status of this may be in the BCM data stream. When you are not logging change the mode to the B3 & B4 byte display and note any repeatable changes to specific bits when the fan comes on and off or changes speed. If we can identify the bits or bytes that are changing we can then export that data as well. Others who have the gauge installed could help with this come on chaps!!
 

·
Moderator
Joined
·
6,918 Posts
Discussion Starter #12
I agree, and I'm really happy that things seem to be coming together like this.

I was going to do my best, but anecdotal notes on the new pack's behavior just didn't sit well with me. I'm a 'by the numbers' type of person.
 

·
Administrator
Joined
·
10,623 Posts
More quick fan monitoring thoughts.

If we can't quickly identify which bits in the BCM data correspond to the battery fan then consider the below.

Dirty hack

If you install a simple 18k/10k potential divider to pin 4 of the icsp on the gauge pcb and connect the input of the PD to the fan connector Negative side (I believe the fan is controlled by grounding it through two resistors and relays to control speed. i.e. it has a permanent live supply) I haven't got the manual to hand, can someone confirm that? If it's the other way round connect it to the positive lead. Then we will get a voltage whenever the fan is operating which can be measured by the gauge and then exported to add to the spreadsheet.

Fan ---> 18k ---> Pin 4 on Pcb ---> 10k ---> Ground (Pin 3 on pcb)

Does that make sense?

I'll upload some new code if you want to try this let me know.
 

·
Moderator
Joined
·
6,918 Posts
Discussion Starter #14 (Edited)
Yes, that is correct - the relays connect the fan circuit to ground.

That sounds easy enough, certainly worth a shot. Can you give me a little schematic?

With this, we would be able to tell whether the fan is high or low as well I believe?

This is contingent on whether I can flash new code to the PIC. I haven't tried that serial PIC programmer I picked up yet.. Wishing I would have just picked up one of the little yellow USB ones that everyone is using. But I can probably use the FT232 USB to serial converter that I'm using to log the data. I think it's functionally similar to the CP2101.
 

·
Administrator
Joined
·
10,623 Posts
That sounds easy enough, certainly worth a shot. Can you give me a little schematic? With this, we would be able to tell whether the fan is high or low as well I believe?.
Correct! different voltage = different speeds. We can just get the gauge to output one of three values to a new column on the spreadsheet. 0 = off, 1 = slow, 2 = fast

This is the schematic Fan Neg Lead ---> 18k ---> Pin 4 on Pcb ---> 10k ---> Ground (Pin 3 on pcb) I can't do anything else schematic or code wise until tomorrow pm. Work I'm afraid.
 

·
Moderator
Joined
·
6,918 Posts
Discussion Starter #16
So an 18k resistor in series with the fan lead, then a 10k resistor from Pin3 to Pin4 of the ICSP connector? I should be able to handle that. ;)

I'll look for Joe's video on flashing the PIC.

Yeah, that's what I plan to do.. regen until I can't anymore, then hook up the grid charger. I thought about that, and I won't have 18 hours.. only 15 hours between when I get home and have to go back to work! Maybe my fiancee can take me to work tomorrow so I can let the car charge for a whole day. I think that sounds like a good game plan.
 

·
Registered
Joined
·
1,616 Posts
Eli, what is the peak voltage of your grid charger? I have mine set at 172V but an thinking about moving it up to 175V.
 

·
Moderator
Joined
·
6,918 Posts
Discussion Starter #18 (Edited)
I just put it together last night. Have to pick up a couple of diodes and wire it up to the car tonight. Not quite sure how I'm going to do that since I don't have any of those fancy piggyback spade connectors, but I'll figure it out. How did you hook yours to the pack again?

I think setting it to 175V(or even 176V) before the diode drop is a good idea. I won't be setting it any lower than that.

That gives about 174V after the drop(assuming 2 diodes) or 1.45V/cell. I think that is what you want. Anything less is definitely not fully charged.

I have seen many, many sticks peak at over 1.5V/cell when cycling them on the Superbrain. That is with a 5A charge though.
 

·
Registered
Joined
·
1,616 Posts
I used banjo connectors that I got at the local electrical shop. One connection to the HV (lower left corner on the left side) and the other to the ground on the 144V to 12V inverter.

I have always been able to charge to the top of the SOC, about 78% with 172V at the power supply. But that isnt FULL, just full as far as the computer in the car is concerned.
 

·
Moderator
Joined
·
6,918 Posts
Discussion Starter #20 (Edited)
Just got back from Run #3. Some very interesting BCM behavior!

I'm going to run the numbers and do the charts now, but I'm pretty sure that this run will be significantly better than the last two. I could be wrong though.

I went out on a trip around town, to charge my battery to as full as the BCM will let me before the run, just like I normally do. SoC was at 72% when I started. However, this time, the SoC stuck at 72%.. For quite a period of time. Then it finally jumped to 75%, and climbed like expected to ~82%. Unfortunately I didn't capture most of this, so I don't know how many amps I inputted, as I started logging just before it jumped to 75%.

I am positive that I inputted a significant amount more than I did on previous "charge up" runs. We'll see shortly how the discharge curve looks!

If it's better, I'm hesitant to put it on the grid charger just yet? I'd like to see where this is going...

Edit: I got 3.20 amp hours out of my pack this time. Averaged 9.35A for 1226 seconds. That is a little more like it. It seems the cycling is helping? I've gotta get more of this data... no grid charge tonight. I'll try to get another run or maybe two in tomorrow.... I wish I would have been doing this for the last few months.

Average input during the down run was 7.59A over 1637 seconds, or 3.45Ah. Not as much of an improvement there?

I'll post the graphics tomorrow, I'm tired.
 
1 - 20 of 154 Posts
About this Discussion
153 Replies
23 Participants
Mike Dabrowski 2000
Honda Insight Forum
We’re the ultimate Honda Insight forum to talk about Honda’s hybrid car and its fuel economy and specs!
Full Forum Listing
Top