Traction Battery SOC "Meta-States"

Actually, I think what you are experiencing is just the nonlinear relationship between the MFD battery display and the actual SOC, along with some of the hysteresis built into that relationship. That is, the 6-bar and 7-bar states actually cover a wider range of SOC% than the 1-5 bar states, and they will stay in those state to lower SOC% when coming from the top as opposed to coming from the bottom. See the attached graph I copied from another thread. From the graph you can see that about 2/3 of the total available range of SOC (40-81%) is covered by the top three bars of the MFD display (55-81%).

That being said, I drive around a lot in EV mode since I have a PHEV aftermarket kit (Enginer) installed, and I keep a close eye on the OEM battery SOC% on my ScanGaugeII. I have noticed that the SOC (% now, not MFD bars) seems to drop faster when it gets below 60%. Sometimes the SOC% will even continue dropping even during a glide, when the battery is actual being charged by the Enginer battery (again, I know this for sure from the SG). That's never made sense to me, since if it's just Coulomb counting, the SOC should never drop when it's being charged.

Note that I do not however fall into the "old Prius" category really ('09 with ~70k miles).

 

I'm glad someone tried to put a name to this phenomenon! I have a 2nd hand 2007 with 123,000km (76,428mi), where I've driven 56,000km (34,796mi) of it on my watch since 2010.


I've been experiencing this phenomenon for more than 3 years already and it's only about the past 2 years that I've managed a way to avoid it. This usually happens when I encounter unusual traffic on my morning commute at the 8th km(5th mile) going down the hill from 160m (524ft) to 15m and going back up to 100m (328ft) in slow-moving traffic.



This section is usually driven 60kph (37mph) for the first downhill part and moves to 80kph (50mph) for the steeper downhill part up to the 100-meter peak. When it's traffic all the way down and up, I couldn't use B-mode to slow down the charging rate, so i'm stuck with normal regen braking where SOC would sometimes be 80% just before I hit the bottom. Thus, going back up, i'm practically stuck with a system-default EV mode (unless I really floor it).

So now I'm practically using most of the juice going back up the hill. The first few times I encountered this going up, I noticed the bars went from green to blue to purple in about 2-3 minutes (I think). It started falling faster from about a little over half of the blue bars (55% SOC). When I checked my SOC reading on the ScangaugeII, it was literally falling about 0.5% per 1-2 seconds even when I wasn't moving. It would get me down to 45-46% midway up the hill so it was a pain seeing those gas used in charging the battery AND moving the car uphill at snail's pace. The more I encountered it, the more I thought my HV battery was on its way and losing capacity faster on high-demand duties.

But somehow throughout my whole experience driving this car, I think it's just the way it's designed. I think for long durations of EV use (automatic or manual), the SOC being made available is actually less than one may think. I may be wrong, so enlighten us further, please.

So one solution I found out is to force charge the HV battery ahead of time or when you start seeing the bars or SOC % values start to go down rapidly on high-demand use of the traction battery and prolonged use of EV mode. It doesn't need to be prolonged stationary force-charging. You can charge bit by bit as you accelerate and keeping an eye / making sure that some juice is flowing to the traction battery as well. This preventive solution of mine works for me and I no longer see purple bars or go below 48% if I can help it.

Another solution that ties in conjunction with the first one is for me is to avoid, as much as possible, reaching the 65% and up SOC territory going down that hill. A few miles before I go downhill, I monitor my SOC already and make sure that my SOC will be at a level that's not too high before I go down. Upon going down, I use B-mode when I can. If it's unavoidable that I'll reach 75-80%, and traffic is slow, i'll go into neutral and use friction brakes in the most efficient way I can.

So I'm not really sure if my traction battery on its way as I'm still getting my expected mpgs for the season down here.

 

I don't see why that is so. You can select B mode anytime.

 

sorry i didnt explain it further... even i continued to use B mode all the way down, i'd still fill up the traction battery to 80% which i was trying to avoid. therefore, i'd switch to N and just use friction braking to cut out charging.

 

I'm going into my 120k maintenance here soon, and Capitol Toyota does a nice 20 some-odd check point. Everything checked out "green" (on a scale of green-yellow-red) including a traction battery check when I purchased this Prius at 100k. Well, actually the cabin filter was at yellow, haha. Anyway, how much stock can I put into their ability to know the condition of the battery? I plan to put a newer HV sooner than later ($800 at our wreckers!), are there signs that a very indicative to a soon to fail battery?

 

Okay, thanks for the info I'll be keeping and eye for any signs on SOC. I had no idea that there was a "red" state, which I assume the bar would be red.

I haven't seen any noticeable SOC signs that make me worry about the battery's abilities thus far, and will use your post as reference. Thanks again!

 

I'm having the same exact problem on my 06' with 50,000 miles. When the charge gets maxed out (76-79% SOC) and then it'll favor electricity to get rid of that extra, then once it hits 6 blue bars in that "normal area" it then continues to fall quickly down to 5 or sometimes 4 blue bars for me. Then it'll charge up to 6 blue bars again and drive and act normal. I have a video of this on youtube.

 

Uart, I have seen the rapid drops in SOC when my car was also new, so your observation may be correct but not unique. A change in battery behavior for the same car, routes and season is probably a more useful reference. If I was more analytical and less lazy I would grab a baseline on our new v(agon) and follow up with yearly testing.

 

The scenario I remember best and repeated many times went this way:

1. Leave home with a cold engine and drive down a hill about 1 mile
2. Drive 3 miles on level ground. ICE is warmed up by now
3. Drive down a hill. Battery SOC is full green
4. Drive some 3 miles of mostly small decline and timed signals. It was in this section I would be tempted to use the traction battery to stay in EV and keep a reasonable speed. My MPG would sky-rocket to well above 100 MPG for the trip ...
5. Alas, at some point the SOC would plummet from 3-4 bars to purple and then I would spend the next mile or three bringing the battery SOC back to it's happy state.

I stopped driving this way after I realized that I was wearing the battery prematurely, and the period where I was rejuvenating the battery was inefficient.

 

Yes!! That's exactly what it does!! I'm looking into The Hybrid Shop. Apparently they can restore it's capacity. Check it out and see what you think. I've called and the price seems steep though. They have happy customers it seems.

 

You might consider a grid charger as another option for 'restoring' degraded battery capacity/performance. Requires a couple hours of DIY, but much less expensive than what other places are charging for the same deep cycling of the battery pack.

 

My guess is that you may be encountering what's been called SOC Drift, in this case brought on by a weak battery. As mentioned above, SOC is tracked by coulomb counting. The controller knows how many Ah the battery is supposed to hold, and integrates current in and out over time to figure out its current SOC. To my knowledge the Prius does not use any sort of adaptive capacity in its calculation, just a fixed capacity number. This is set pretty conservatively so typically you are still well within its limits even as the battery degrades over time, though you are using more of its total remaining capacity as it degrades. When the remaining capacity gets low enough, you are starting to enter into territory where the normal SOC excursion (based on the original capacity minus a significant buffer) is in danger of over charging or over discharging the battery. As a guard against this, Toyota added a high and low voltage limit well outside normal operating conditions that tells the controller it's estimated SOC is significantly off and it needs to revise its estimate of actual SOC. Since it seems to be unable to change its assumed total usable capacity, as a battery becomes increasingly weak it would seem likely that it would encounter these limits more often, and eventually end up sort of bouncing back and forth between the high and low limits. I believe this is what you are describing as your top and bottom referenced states.

The upper voltage limit is pretty well documented to be around 242V, as this trick was used in some of the early PHEV hacks to get the Gen 2 to use the extra capacity in an add on pack.

There is a brief description of this SOC Drift phenomena from some of that early work here:
State Of Charge Drift - EAA-PHEV

 

There's also a mention of the lower limit perhaps being ~200V and of the "dropping SOC while charging" phenomena that seems to match some of the descriptions above on this page from the hybrid interfaces controller used in some of these Gen 2 DIY conversions:

NOTE: if you revert to the original battery ECU when the packs have been exhausted down below 200v, the battery ECU will initially show an alarming drop in SOC despite a charging current as it tries to resolve the two conflicting parameters: it will be raising the SOC as it does coulomb counting on the charging current that also follows a low SOC but it will also be dropping the SOC because of its SOC drift mode that tries to correct for a low voltage by setting a low SOC. This conflict should resolve in a few minutes as the very low SOC will cause a much higher charging current and the coulomb counting from this should more than offset the reduction due to low voltage SOC shift. It should settle out within 10 minutes back to normal voltage and normal SOC. If you revert when the voltage is still above say 210v then none of this off conflicting behaviour will occur as the SOC drift at this higher voltage is now much smaller than the increase due to charging.
​

BMS+ ECU screens

I can't recall now, but there may be some temperature dependency on these voltage limits and certainly the battery's voltage, resistance and capacity characteristics will be different at the extremes of the temperature range. Driving conditions that result in large charge / discharge currents will also probably be more likely to push a weak battery into one of these limits, particularly if the internal resistance of the battery is getting high as I believe it often does as it weakens.

 

If anyone that regularly sees this issue has a way to monitor the HV battery voltage, it would be very interesting to see if these events seem to correspond to the voltage hitting these limits. Don't have any experience with torque/scangauge, but the pre-boost high voltage should be about the same as battery voltage? The sum of all the block voltages should also equal battery voltage as a sanity check.