ScanGauge II Extended PIDS: Capacity and Specific Energy

[AnthonyW]

So after all 7 reasons that I listed that I would not buy a ScanGauge II in this thread:

https://www.insideevsforum.com/community/index.php?threads/odbii.1110/#post-20812 and then, of course, I went out and bought a ScanGauge II. I contacted ScanGauge and they offered me some free accessories and free updates for life in exchange for me being a beta tester.

The list of extended PIDs are here: https://www.scangauge.com/support/x-gauge-commands/honda-acura/ and they are based on the Honda Accord PHEV. Some still work for the Clarity and the ones that don't are being worked on as we speak.

Note: This post will be specifically about ScanGauge II X-Gauge (Extended) PIDs that relate to specific energy and capacity. I will add others related to power, temperature and performance as they become available.

 

[AnthonyW]

The first PIDs I would like to examine are SOC. There is the SOC that we all know from the app and then there is another one behind the scenes which is called
HV Battery Cell Max State of Charge (%) (MXS).

When SOC is at 100% MXS is at 96%. I took observations at various states of battery usage and noticed that SOC declines at a faster rate than MXS.

View attachment 1590

*Note: I don't yet have observations under 30%.
**Since I don't have observations at each and every percentage point, I determined the slope and the intercept of the observations I had and made this graph.

This seems to confirm what I suspected. Previously, I would charge to 100% SOC and it would take me about 50% on all electric to get to work. When I started charging only to 85% it would take me 55% to get to work. This also seems to support KentuckyKen's observation from the other day that he seem to lose charge level bars faster at the bottom than at the top. It looks to me that Honda is manipulating our perceptions as they conclude that most people will do most of their driving in the higher states of charge. Therefore with the perceived slower rate of decline at the top, we feel that the car is performing better than it is.

 

[Atkinson]

Any updates on the new gauges?

Battery Coolant Temperature (°F)
Battery Module Power (kW)
DC-DC Converter Temperature (°F)
Generator Motor Speed (RPM)
HV Battery Cell Max State of Charge (%)
HV Battery Current (Amps)
HV Battery Fan 3 Speed (RPM)
HV Battery Max Cell Voltage (mV)
HV Battery Usable Capacity (%)
Motor Inverter Current (Amps)
Outside Air Temperature (°F)
State of Charge (%)
Temperature of Air In Vehicle (°F)

 

[Viking79]

Thanks for this!

As for the gauge, I imagine that they did that to make it behave like most gas gauges. I have noticed the same. The EV range number is the estimate that I use.

 

[AnthonyW]

Any updates on the new gauges?

Battery Coolant Temperature (°F)
Battery Module Power (kW)
DC-DC Converter Temperature (°F)
Generator Motor Speed (RPM)
HV Battery Cell Max State of Charge (%)
HV Battery Current (Amps)
HV Battery Fan 3 Speed (RPM)
HV Battery Max Cell Voltage (mV)
HV Battery Usable Capacity (%)
Motor Inverter Current (Amps)
Outside Air Temperature (°F)
State of Charge (%)
Temperature of Air In Vehicle (°F)

[/QUOTE]
So the information in the first post relates to HV Battery Cell Max State of Charge (%).

The graph above is an interpolation of the data because I only I had a few data points. I have since gathered more data points and here is the graph of the raw data:

 

[AnthonyW]

Okay so we got another one working. It is HBV Hybrid Battery Voltage (Volts). In regards to voltage here is what we know:

-168 cells between packs A and B connected in parallel so basically we have 84 working cells. Source: Design Specs Sheet
-Nominal voltage is 311 volts. Source: Design Specs Sheet
-Max voltage is 352 volts Source: Emergency Response Guide.
-Nominal voltage per cell 3.7 volts. Source: Design Specs Sheet and 311/84=3.70
-Max voltage per cell is 4.19 volts. Source: 352/84

My ScanGuage showed 340 volts after a full charge and ran all the way down to 286 volts before the car turned the ICE on automatically. That would mean that our usable ranges is 4.05 volts to 3.40 volts. Considering that Lithium Nickel Manganese Cobalt Oxide have an operating range of 3.0v to 4.2v (and cut off voltage at 2.5v), this information would suggest that Honda reserves very little buffer at the top and a whole lot at the bottom. This is backed up by the ScanGauge MXS state of charge reading at 96% when the battery is fully charge.

Might also explain why the car has so many unexpected turn on's right after a full charge. With so little buffer at the top, it does what ever it takes not to accept charge any further.

 

[insightman]

Might also explain why the car has so many unexpected turn on's right after a full charge. With so little buffer at the top, it does what ever it takes not to accept charge any further.

I believe your new ScanGauge data point could help Bobcubsfan. If the ScanGauge indicates his battery always shows 4.2v, it might indicate his car needs a replacement battery-full sensor (whatever that is).

Could it be possible that the Clarity burns off excess energy by starting the engine and then running the starter motor/generator IN REVERSE (well, trying to slow down the engine)? That would use more energy than simply having the starter motor/generator spinning an unfueled engine with the valves closed (like the Accord Hybrid does). It's the only scenario in which it makes sense to me to start the ICE.

 

[Steven B]

So the information in the first post relates to HV Battery Cell Max State of Charge (%).

The graph above is an interpolation of the data because I only I had a few data points. I have since gathered more data points and here is the graph of the raw data:

View attachment 1624[/QUOTE]

So the blue line is what it shows via the HL app that the SOC is but the orange line is what the actual SOC is. Looks like I'll be changing my stop times to end no more than 94% per HL app SOC.

 

[AnthonyW]

So the information in the first post relates to HV Battery Cell Max State of Charge (%).

The graph above is an interpolation of the data because I only I had a few data points. I have since gathered more data points and here is the graph of the raw data:

View attachment 1624

So the blue line is what it shows via the HL app that the SOC is but the orange line is what the actual SOC is. Looks like I'll be changing my stop times to end no more than 94% per HL app SOC.[/QUOTE]

Correct, the blue line is the HondaLink State of Charge %.

 

[Steven B]

X axis is just numbering the data points?

So based on the Scangauge II data, the ACTUAL remainder (battery capacity unavailable for EV use) on the low end is ~20% and high end is ~4%? At 20% (HL App 11%) it will switch to HV mode and may continue to pull battery level down below that if doing hills (I've seen it drop as low as HL App 8%).

 

[AnthonyW]

X axis is just numbering the data points?

So based on the Scangauge II data, the ACTUAL remainder (battery capacity unavailable for EV use) on the low end is ~20% and high end is ~4%? At 20% (HL App 11%) it will switch to HV mode and may continue to pull battery level down below that if doing hills (I've seen it drop as low as HL App 8%).

Yes, you are correct on all points.


Sent from my iPhone using Inside EVs

 

[Atkinson]

What gauges do you have working so far?
Besides Battery Voltage.

 

[KentuckyKen]

Okay so we got another one working. It is HBV Hybrid Battery Voltage (Volts). In regards to voltage here is what we know:

-168 cells between packs A and B connected in parallel so basically we have 84 working cells. Source: Design Specs Sheet
-Nominal voltage is 311 volts. Source: Design Specs Sheet
-Max voltage is 352 volts Source: Emergency Response Guide.
-Nominal voltage per cell 3.7 volts. Source: Design Specs Sheet and 311/84=3.70
-Max voltage per cell is 4.19 volts. Source: 352/84

My ScanGuage showed 340 volts after a full charge and ran all the way down to 286 volts before the car turned the ICE on automatically. That would mean that our usable ranges is 4.05 volts to 3.40 volts. Considering that Lithium Nickel Manganese Cobalt Oxide have an operating range of 3.0v to 4.2v (and cut off voltage at 2.5v), this information would suggest that Honda reserves very little buffer at the top and a whole lot at the bottom. This is backed up by the ScanGauge MXS state of charge reading at 96% when the battery is fully charge.

Might also explain why the car has so many unexpected turn on's right after a full charge. With so little buffer at the top, it does what ever it takes not to accept charge any further.

AnthonyW, thanks to Atkinson sending me a Dropbox full of Honda Tech files I found some great information on the battery from the PDI (pre delivery inspection?)!!!! (And losts of other goodies I will share in a separate post)
It says the dealer is to use the i-HDS to go into the Electric Powertrain Data List (a veritable treasure trove of info if you can get the ScanGauge to access it) and check the Battery Pack Capacity signal. If it’s below 36.6 Ah, the battery is eligible for warranty replacement. That is a valuable data point you can use in your calculations and the first time I’ve seen a metric for when Honda will replace the HV battery under warranty.

I’m a retired molecular biology/clinical research lab mgr, so I’m am not trained in EE, but doesn’t this mean that the cut off for battery replacement is much higher than has been previously thought.

See if I’m on the right track here.
You report a nominal 311v for the 17Kw battery pack, so 17,000w/311v = 54.7Ah total capacity. And the screen shot in the PDI (SB 17-093) shows 55 Ah. So good agreement.
Then the cut off for replacement of 36.6Ah divided by 54.7 or 55 Ah total capacity would be replace At 67 or 66.5% of original capacity (or if battery degrades by ~33%).

This sounds a whole lot better but have I made a big mistake???
This ain’t my field of expertise and I’m reading the docs Atkinson sent me at 2 in the morning. Please work your magic with this new data and let us all know If it means what I think it means.

 

[AnthonyW]

KentuckyKen, thanks for the compliments, but I am no electrical engineer either! I just fumble and stumble until I end up falling on the ball in the endzone! I think you are right though. The SAE article that I posted in this thread https://www.insideevsforum.com/comm...-state-and-battery-longevity.2555/#post-22709 states that the capacity of the cells is 27.3Ah. Since the cells are connected in parallel, that would double the capacity to 54.6 (connecting in parallel doubles capacity, connecting in series doubles the voltage) so your calculations are correct. Great find!

P.S. Because I am an accountant, I am a stickler for details. I calculate that the the pack is not 17kWh, but 16.970kWh and the true nominal voltage is 310.8v. I know, I know, that's just me. It's awesome that we are finally able to triangulate and confirm a lot of the stats we have been wondering about!!!

 

[Wayne Wilson]

I would be interested in seeing the battery voltage graph at various SOC. Almost all SOC numbers are 'calculated' based on some function. I see from the technical documents that Honda has both current sensor's and voltage sensor's available to them, so they could either be calculating SOC from a battery voltage/SOC current function or from a subtraction of actual KWH consumed versus a KWH (max) value.

 

[KentuckyKen]

Does that mean that the warranty replacement of less than 36.6 Ah means that our batteries get replaced after they lose ~33% of their original capacity and not the 75% that was mentioned somewhere according to a graph I didn’t understand. This is what I wanted you to verify with the new # I found thanks to Atkinson’s docs. Thanks.
Using your more exact numbers, I still get 36.6Ah/54.6Ah = 67% or 33% loss of capacity to trigger a replacement. Is this right??

 

[Viking79]

I would be interested in seeing the battery voltage graph at various SOC. Almost all SOC numbers are 'calculated' based on some function. I see from the technical documents that Honda has both current sensor's and voltage sensor's available to them, so they could either be calculating SOC from a battery voltage/SOC current function or from a subtraction of actual KWH consumed versus a KWH (max) value.

Most Li-ion use "coulomb counting" to measure SoC, combined with voltages. So yes to both It is important to occasionally "fully" charge and discharge a battery to calibrate the SoC. By fully, I mean what the car thinks is full and what the car thinks is dead (so in Clarity roughly 10-85% or whatever it uses). If you constantly run partial charges (like between 20-60%) it won't hurt Li-ion, but it might cause the SoC to drift off and be less accurate and might also prevent cell balancing. Depends on what algorithms they use for all of that.

My understanding is they charge or discharge to a specific voltage, then use coulomb counting (integrate the current into and out of the battery) to know the SoC. They might also check that against voltage, but maybe only at certain times (like at car start). The battery is dead when the weakest cell reaches the minimum voltage (or maybe cell pair since they are wired in parallel) or when the current integration shows you have used all your capacity.

 

[AnthonyW]

What gauges do you have working so far?
Besides Battery Voltage.

Atkinson, I will put together a list what is working and what is not working later today. Its a slow process as he sends me codes to try and then I have to report back whether the Scanguage is A.) reporting data and then B.) whether the data is accurate. Based on my response he sends me new input codes to try. He has been on vacation for the last couple of weeks so that has slowed our progress.

 

[Wayne Wilson]

might also prevent cell balancing

Cell balancing during charging is an important consideration as variances among cells impact overall pack performance vs theoretical performance based on ideal batteries. So when, exactly, in the charge process, is cell balancing taking place? I think we don't know much about Honda's battery management system and all we can do is infer by reverse engineering from things we do know. One of my concerns about just SOC data is that it is already synthetic, i.e it's already be calculated somehow, batteries don't directly report SOC....

 

[AnthonyW]

Most Li-ion use "coulomb counting" to measure SoC, combined with voltages. So yes to both It is important to occasionally "fully" charge and discharge a battery to calibrate the SoC. By fully, I mean what the car thinks is full and what the car thinks is dead (so in Clarity roughly 10-85% or whatever it uses). If you constantly run partial charges (like between 20-60%) it won't hurt Li-ion, but it might cause the SoC to drift off and be less accurate and might also prevent cell balancing. Depends on what algorithms they use for all of that.

My understanding is they charge or discharge to a specific voltage, then use coulomb counting (integrate the current into and out of the battery) to know the SoC. They might also check that against voltage, but maybe only at certain times (like at car start). The battery is dead when the weakest cell reaches the minimum voltage (or maybe cell pair since they are wired in parallel) or when the current integration shows you have used all your capacity.

Viking79, is correct. Lithium Ion SOC can't be determined based on voltage alone. See chart below (ignore the voltage numbers, that does not represent the NMC type battery that we have, but the shape of the curve is representative). SOC can be calibrated using voltage but only when getting close to fully charged and close to completely empty of the usable range.