Using clone ELM327 Bluetooth OBDII adapter with Leaf

[bradbissell]

What is the correct shunt order for the app? And if it isn't different between model year, can we make it hard coded in the app? I see lots of values posted in screenshots within this thread. For example 2 posts up. One unified hard coded value would help.

 

[Turbo3]

Correct value is 4821

 

[Weatherman]

Let me know what you see when it hangs.

When connected, it shows 29 lines of numbers, headed by 00 through 28, followed by a 7BB 8, then eight two-digit hexadecimal numbers.

When it hangs (in this case on a 23 10), it flashes on and off an additional 11 lines (00 through 10).

If I disconnect the BAFX, the additional 11 lines stop flashing on the screen and it's back to showing only 29 lines of numbers, again.

 

[LEAFfan]

Correct value is 4821

Actually, Greg said it was 8421 and we had the incorrect one.

 

[Turbo3]

So when it hangs the lights on the BAFX stop flashing and the log screen starts flashing a number of lines that matches the second number in the status message?

Any chance you could grab a screenshot or picture of those lines? I would also like to see the 29 lines after you pull the BAFX.

Can you exit the app when it gets into this state or do you need to kill it.

 

[Turbo3]

Correct value is 4821

Actually, Greg said it was 8421 and we had the incorrect one.

I guess GregH and Gary Giddings need to talk ;-)

When everyone agrees on a single value then I can hardcode it. But it sounds like now is not that time.

 

[Weatherman]

So when it hangs the lights on the BAFX stop flashing and the log screen starts flashing a number of lines that matches the second number in the status message?

Any chance you could grab a screenshot or picture of those lines? I would also like to see the 29 lines after you pull the BAFX.

Can you exit the app when it gets into this state or do you need to kill it.

I'm not absolutely sure the number of extra lines matches the status message, since it extends off the screen. I'll e-mail you a couple of screen shots as it's flashing between the 29 lines and the 29-plus-whatever lines.

Once it gets into that state, the transmit/receive LEDs on the BAFX stop flashing and I have to just do a normal stop and restart the app to clear it.

 

[garygid]

I still think that the shunt order is 4812, but I have not seen as much data
as needed to verify my guess.

I still believe that the highest-voltage cells should be On,
but GregH believes that the lower-voltage cells should be On,
figuring that the current drawn by the shunt, and the wire length
to the BMS, would cause an apparent voltage drop in the shunted cells,
so the shunted cells would appear to be lower voltages.

I think that the shunt current is unlikely to be as much as 1 amp,
which would be 400 watts if all the shunts are on.
The typical cell "resistance" seems to be around 0.001 ohms.

The battery pack voltage drops perhaps 10 volts when the
car is using something like half power (40 kW, or 100 amps).
If so, the Pack resistance would be about 0.1 ohm, and each of
the approximately 100 cells (really 96 cell-pairs) would contribute
about 0.001 ohm.

So, the voltage drop in the cell due to the 1 amp current would
only be 0.001 volt, barely visible on the 50 mv scale. Of course, these
numbers are guesses, and the wiring could have some effect, especially
if very small wires were used, but I consider that unlikely.

So, I think that a shunt-on cell-pair most likely will not show a
significant loss in the measured voltage.

However, GregH says that he has observed this voltage drop
when the shunts come on, and he is a very good observer,
while I have had almost no live watching of the data.
My conclusions come from an incomplete analysis of the
available pack data, and mostly static displays.

If you find that the 4812 shunt-On status seems to be incorrect,
and that swapping each Odd-Even pair gives what appears
to be a better assignment of the Shunts to the cells, please show
your pictures, along with the Shunt order assignment.

I do observe that sometimes it doesn't make much sense
(odd-even swapped or not) which cells that the BMS has
picked to discharge with the shunts. I see medium cells
shunted and high ones not shunted. But, with the 4812
ordering, I have not seen the lowest cells being shunted.

So, that might be more than you wanted to know, but
that is my (perhaps defective) description of the source
of this "minor" odd-even-swap controversy.

We are still learning, in many areas.
Cheers, Gary

 

[DaveEV]

One small issue remains for me: when changing between landscape and portrait modes few times sometimes the screen goes dark, so always on option stops working, a bit annoying when I drive and phone slips and changes orientation and then goes dark...

Happens on my galaxy S3.

I've seen similar issues as well. Sometimes my screen will actually lock and I'll have to unlock it... Also a Galaxy S3.

Correct value is 4821

Actually, Greg said it was 8421 and we had the incorrect one.

I guess GregH and Gary Giddings need to talk ;-)

When everyone agrees on a single value then I can hardcode it. But it sounds like now is not that time.

FWIW I tried 8421 and it "looked" much worse than what I'm currently using, 4812. I thought the best one was 4812? At least 4812 seems to produce results that one would expect most of the time.

 

[91040]

Correction- It was 25k that was functioning without problem. Tried 25n3 tonight. it worked for a while but then stopped connecting, showing an error with the number 16.

Edit: Error was "Leaf Trying 16." This version worked fine yesterday.

 

[garygid]

To better report errors, best to get the latest version, v0. 25n3
of 11 hours ago that has extended (usually 4-digit) error codes.

 

[GregH]

So, the voltage drop in the cell due to the 1 amp current would
only be 0.001 volt, barely visible on the 50 mv scale. Of course, these
numbers are guesses, and the wiring could have some effect, especially
if very small wires were used, but I consider that unlikely.

I would say the harness wires are the primary reason for the observed voltage drop when shunts are turned on. Most balancing decisions I've seen were made at the end of a full charge. The balancing load is certainly much less than 1A, probably on the order of 100mA (just imagine how much heat that little box would have to dissipate if most of the shunts were on). The BMSs we designed at EnergyCS were only 40mA and the BMSs I've designed for EDrive and Trexa were 80mA. I know of several other systems that balance at about 80mA for 100Ah vehicles.

Anyway, decisions made for balancing stick for several hours of "ON" usage (driving or charging, but not when the car is off). So if the balancing resistor only draws 100mA and is on for 2 hours (over a 24 or 48 hour period when the car is on) it will only draw about 0.2Ah from your 50-66Ah pack. Most of this balancing will happen at SOCs much lower than full, therefore the cells where shunts turn on may not look like the highest cells anymore. It doesn't matter. They were at some point and the BMS is just carrying through with its task of pulling several mAh out over a period of time. When the shunts turn on, you can see the cell voltages drop if you're watching it in real time (like on a LeafDD).

 

[Weatherman]

With the latest version, I got a halt in data transmit/receive on the following "Retry Leaf"s:

23 05
23 10
23 04
23 12


I've only tried these four times, so I don't know if there's any pattern to the sequence.

Each halt requires the app to to shut down and restarted to get data flowing again.

Jim, just confirming that when the BAFX gets into a state where it's no longer transmitting data, we want the app to send an ATZ to see if it can kick-start it again (or whatever the equivalent is of a restart of the application).

 

[kevinleaf]

I have been trying to figure out:
When you get a call and the phone talks bluetooth to the car does the bluetooth to the ELM327 get interrupted?
Can you play bluetooth audio to the car while still communicating to the ELM327 at the same time?

I have dropouts in my data logs and I'm trying to determine why.

 

[LEAFfan]

I have been trying to figure out:
When you get a call and the phone talks bluetooth to the car does the bluetooth to the ELM327 get interrupted?
Can you play bluetooth audio to the car while still communicating to the ELM327 at the same time?

I have dropouts in my data logs and I'm trying to determine why.

Probably would interfere, if you're using the same phone for both. I just use the Kyocera E so there is no interruption with my other phone.

 

[UkrainianKozak]

Tried 25n3 today on Galaxy S3, could not connect at all
Always getting stuck at Looking 4 ELM, had to revert to 25k...
Restart, reboot, clear data didn't help... no connection at all...

 

[davewill]

I have been trying to figure out:
When you get a call and the phone talks bluetooth to the car does the bluetooth to the ELM327 get interrupted?
Can you play bluetooth audio to the car while still communicating to the ELM327 at the same time?

I have dropouts in my data logs and I'm trying to determine why.

I've done both phone calls and streaming music at the same time as the app is running. It seems like the communications are more difficult, with less time spent with a green data status. The system does multiplex the use of the bluetooth between multiple connections, but some dropouts seem inevitable although my adapter has ALWAYS dropped data. I typically only get 60-70% good reads in the LBA stats at the best of times. It hasn't been a problem for me, but I don't try to log the data, just use the display.

The bigger issue(for me) is that my phone usually ends up back asleep if I take a call, which means no LBA display until I can pull over and wake it back up (or I break the law and try to wake and unlock the screen while I'm driving).

 

[TomT]

FYI, 25n3 is working perfectly with my Samsung S3 and BAFX adapter.

 

[gbarry42]

I was contemplating buying one each of the three favorite ELM327 devices (though from what I read I should buy two each, so that's six!) to find out which worked reliably. But I got to wondering, (warning--heretical idea), maybe instead of sending my money to buy poorly-built blatant rip-off products, I could instead get one that's built here with parts actually made by Elm...it would only cost me three times what one of the clones would, but it should be stable. I don't see where anyone has tried that approach.

One question is whether we'd expect the "real" device to work better in this situation.

The other is whether having the latest firmware in the ELM module would cause trouble with the leaf app, given that the clones use a much earlier firmware version.

 

[kevinleaf]

Seems I sort of answered my own question about reliability of simultaneous bluetooth connections: phone to car for a call and phone to the ELM327.

In the beginnings of my LEAF log viewer I can see a big drop between points 5 and 6:

I know I was on the phone at this point. Now I don't know if the dropoff is because of the phone went to sleep while I was on the call (screen turned off) or bluetooth connection was sketchy. I'll have to play with it some more.