RavCharge, a solution to charge timer and entune woes

[fooljoe]

There needs to be an adjustment for temperature adjustments. It's all fine in mild climate California, but for somebody in one of the those really cold places where the battery heater will also be running.

Yeah I've thought about the effect of ambient temperature a bit, but with basically 0 data to go on I'm not sure where to start here. The best approach might just be my eventual plan to allow you to set your own charging efficiency value, since running the TMS during charging obviously plays into that.

...the time will be extended, both because of the cold battery and because of the increase energy consumption of the battery heater.

Won't the cold battery cause it to hold less charge, which would then have the effect of shortening charge time? There could be a bit of canceling out here, with that shortening being countered by the lengthening caused by the heater's energy consumption, and the net result might be RavCharge's "default" estimate still being pretty close. Again, we just need to see the data to figure it out.

 

[fooljoe]

I will have "tweeked" (i.e. calibrated) all settings in RavCharge to correlate as closely as possible to Entune's precharge estimates for projected range and the in-car's prediction for time to charge.

I'm not so sure why you want to do this - do you just want to see how silly your settings have to be to come up with a time-to-charge and/or range estimate similar to the car's? By the way, Entune is not involved in the car's time-to-charge prediction as far as I can tell - that's something entirely done by the car's computer and sometimes reported to Entune, sometimes not.

30.6kWh normal charge capacity (very conservative)

This is conservative for estimating your range, but it is decidely NOT conservative for estimating time-to-charge - fewer kwh in your battery means shorter charge time.

2.9 mi/kWh driving efficiency (3.1mi/kWh actual in-car)

This setting is not used for time-to-charge calculations.

Number of bars on GoM: 8 bars

I would be careful to differentiate the number of bars from the "miles remaining" display which is I think what most people assume you're talking about when you refer to the Guess-o-Meter.

Dashboard charging est: 3 hrs 50 mins

I'm a bit surprised by this. Was this before or after plugging in, and was it after setting extended charge mode? I think that number's actually pretty close, and a bit optimistic if anything, which is basically unheard of for the car's time-to-charge estimator.

Based on what you've said are your best estimates of the actual conditions (235V/32A/49% SOC), and assuming a 41.8kwh extended / 35kwh normal charge capacity battery, I (hence RavCharge) would estimate your extended charge to actually take about 4:06. Do report back with the actual time!

EDIT: Once your car began charging it reported a remaining charge time of 5:03, which is much more in line with what I'd expect the car to report. It probably would've reported something greater than this if you checked it after plugging in and with extended mode selected, since after charging starts the car will finally know the voltage and can presumably improve upon its estimate (but its estimate still sucks.)

By the way, your charge tonight is a perfect example of why you should use RavCharge and not trust the car's timer! Because of the car's overestimate of time-to-charge, your charge started before midnight, which could mean you're paying higher energy rates. You let the car's timer kick in instead of setting it a few hours later than you actually want and letting RavCharge initiate the charge.

I have observed if you shutdown the car with the climate control still in operation, the range indicated by Entunes will differ from the range indicated by Ravcharge, if they would have been otherwise the same. For this reason, I'm now in the habit of making sure my climate control system is turned "off" BEFORE I shutdown the car. This helps keep Ravcharge and Entune's range estimates in sync.

Are you mixing "Entune range" with "dash range"? These are not the same. From what I've seen, the Entune range is equal to the dash range with the CC off. And RavCharge's display of Entune range is exactly the same as what the Entune app will display (except perhaps for some rounding differences.) RavCharge's "real" range prediction, on the other hand, is unrelated to the dash/Entune, aka GoM, range, and is the only range I recommend anyone pay attention to (in concert with their expected efficiency for a planned drive.)

 

[Dsinned]

Here is the link to Enerati HEM plot of power consumption of my Leviton 7.7kW EVSE charging station while doing a scheduled extended charge to my RAV4 EV last night. If you zoom in on the graph there is a noticeable taper during the last 22 minutes of bulk charging, followed by 27 minutes of equalization that took place for 2 hrs and 53 mins after the bulk charge completed.

https://www.enerati.com/Electricity...=ODAwMDAxMDAxMDAwNDAwMjkwMDAwMDAyMDAwMjAwMDA=

My car's extended charge took place between 23:38 and 03:38 for a duration of precisely 4 hrs 0 mins based on my HEM graph (above). A delayed departure was scheduled for 06:50, so the equalization charge took place from 06:30 to 06:57 for a duration of 27 mins. The pre-charge "time to charge" calculated by Ravcharge was 3 hrs 49 mins, so as far as the bulk charge, it was fairly close, but still optimistic by 11 minutes even with the EVSE voltage artifically set lower to 220V. Thus, if I had set the EVSE source voltage to 235V, closer to actual, I think Ravcharge's "time to charge" calculation would have been more accurate.

After the extended charge, the range projected by Entunes in Ravcharge (and displayed on the GoM) was much lower than I expected at only 118 miles. The outside temperature this morning, before 9am, was about 58F.

 

[Kohler Controller]

Kohler - Thanks for the interesting charts. Interesting that it tapers down to about 10A, but only for the last 15 minutes of a nearly 6 hour charge. I would also be interesting to see where the taper starts on a 40A EVSE. I'm assuming the charts are from your AV EVSE. It looks like the different charts have a different draw - one is clearly over 7kW and the other is barely over 6kW. Why is that?

The Enerati monitoring system only measures current and assumes 230V, hence the lower over all power. I found out later how to add a correction-factor to get it to show 7.4kW, but didn't have access to those screen shots for this thread.

Here is an extended charge with the correction factor.



The start of taper occurs when the voltage of the pack reaches 444V, regardless of charge rate (40 or 30A).
My taper usually lasts exactly 15 minutes but I suspect will grow with time. Seems to stop when current drops to ~6-8A.

Here is a close up of the taper (not the same charge as shown above).

 

[Kohler Controller]

Exporting the data to a spreadsheet would be much more useful for real calculations. I suspect that Kohler just did a screen shot and cropped and uploaded to PhotoBucket.

Miimura is correct Dsinned. Hit Print Screen and cropped it using MS Paint, then uploaded to PhotoBucket. Not elegant, but gets the job done.

 

[Kohler Controller]

Here is my HEM plot of last night's power consumption of my EVSE (Leviton 7.7kW) charging station while my RAV4 EV was scheduled to do an extended charge.

https://www.enerati.com/Electricity...=ODAwMDAxMDAxMDAwNDAwMjkwMDAwMDAyMDAwMjAwMDA=

Here is a close up of Dsinned's taper (20 minutes).

 

[fooljoe]

My car's extended charge took place between 23:38 and 03:38 for a duration of precisely 4 hrs 0 mins based on my HEM graph (above). A delayed departure was scheduled for 06:50, so the equalization charge took place from 06:30 to 06:57 for a duration of 27 mins. The pre-charge "time to charge" calculated by Ravcharge was 3 hrs 49 mins, so as far as the bulk charge, it was fairly close, but still optimistic by 11 minutes even with the EVSE voltage artifically set lower to 220V. Thus, if I had set the EVSE source voltage to 235V, closer to actual, I think Ravcharge's "time to charge" calculation would have been more accurate.

Again, I'll note that the time-to-charge figure you saw on RavCharge should be discounted completely, because you used inputs for voltage and battery capacity that were both pretty far off from what you considered them to actually be. The time still came close because the two incorrect inputs had the effect of canceling each other out. Also, you stated that the car estimated the time-to-charge at 3:50, but that's obviously incorrect because the car started your charge at 11:38pm when you had a departure time of 6:50am set. That means the car actually predicted your charge to take 6:52 (we know the car actually targets 20 minutes before your scheduled departure), an overestimate of 72%!

Had you just left everything in RavCharge at the defaults (except for the clearly-known amperage), i.e. 240V, 32A, and 35kwh capacity, RavCharge would've calculated your charge to take 4:01, precisely how long the charge actually took (you can't do any better than a +/- 1 minute rounding error.) If the voltage was actually 235, then that along with assuming 2% battery degradation (or 34.3kwh normal charge capacity) would've given you the same result.

Thanks for the plots. I noted a couple of interesting things: Your power consumption steadily grew from about 7.4kw at charge start to 8kw at the peak before the tapering started (Kohler's show a similar growth in power consumption.) That's a fairly significant change, and also quite strange given that your EVSE consumption shouldn't ever exceed its rated 240V@32A, or 7.7kw, especially if your voltage is really 235, which would give you 7.5kw max.

Oh and one more thing: Can you report back on how the timestamps on the entune charge start/complete emails compare to the start/stop times you saw from your HEM? Most people would only have the entune emails to determine time-to-charge, so it's worthwhile to figure out what kind of a lag, if any, there is between the actual charge event and the email notifications. Thanks!

 

[Kohler Controller]

Thanks for the plots. I noted a couple of interesting things: Your power consumption steadily grew from about 7.4kw at charge start to 8kw at the peak before the tapering started (Kohler's show a similar growth in power consumption.) That's a fairly significant change, and also quite strange given that your EVSE consumption shouldn't ever exceed its rated 240V@32A, or 7.7kw, especially if your voltage is really 235, which would give you 7.5kw max.

Oh and one more thing: Can you report back on how the timestamps on the entune charge start/complete emails compare to the start/stop times you saw from your HEM? Most people would only have the entune emails to determine time-to-charge, so it's worthwhile to figure out what kind of a lag, if any, there is between the actual charge event and the email notifications. Thanks!

I assume Dsinned has a correction factor installed on the graph viewer as well. This explains the 8kW output using a 7.7kW charger.

As for the start/stop times vs Entune's email, yes, there is a lag time but I don't remember by how much. Will have to run a test. I'm thinking only 1-2 minutes.

 

[Dsinned]

According to Entunes email notifications, my start time was 11:39pm last night and my completion was 03:36am this morning, for a total duration of 3 hours 57 minutes. With the latency in the system, that turns out to be rather close to the car's prediction of 3 hours and 50 minutes. The latter was determined prior to being plugged in. The actual timestamp start/stop times in the car were way off (as usual) even with the latest Gateway (Tesla) computer firmware update, although it did complete the extended charge + equalization before 7am as was my expectation.

However, with only ~4 hours required to complete the (bulk) charge, there was no reason for it to start 21 minutes before midnight. If I only did a "normal" charge, I suspect it would have started just after midnight as scheduled by Timer1 in Ravcharge. For an extended charge, with my battery's pre-charge SoC >50%, or a scheduled departure time after 7:11am, the car might have started charging on its own just after midnight as well.

As I recall, I have my clamp on current transformers' sensor/transmitter set to "120V" used at my Leviton L2 EVSE charging station rated for 7.7kW maximum AC output power. The default is "115V", but I can step it up (or down) in 5V increments. When there are two CTs used in the sensing circuit, as is the case for a 240Vac EVSE, this adjustment factor is doubled by the EnviR HEM receiver, which in turn is used by the Enerati web based software to plot the graphs of power consumption data.

Unfortunatlely, there is no power factor corrections reflected in any of these AC measurements. However, Enerati also allows adjustments in the amplitude of its graphic and table data in predefined "configurations". Configurations consist of either individual or multiple AC circuits added together, with set adjustment factors (e.g. x 97%). Thus, if you switch between the graph for the dedicated sensor of "Leviton EVSE L2" and the graph for the "EVSE L2 (corrected)" CONFIGURATION, which is still just for the EVSE power consumption, you will see how these various adjustments interact to affect the data plotted in each of these graphs.

https://www.enerati.com/Electricity...=ODAwMDAxMDAxMDAwNDAwMjkwMDAwMDAyMDAwMjAwMDA=

As for why the plot line slowly increases over time, maybe its because while the battery is under charge, it is slowly heating up and tempco for internal resistance of the battery pack is inversely proportional to temperature, so the EVSE's charging current slowly increases.

 

[fooljoe]

According to Entunes email notifications, my start time was 11:39pm last night and my completion was 03:36am this morning, for a total duration of 3 hours 57 minutes. With the latency in the system

That's quite odd - so that means Entune sent the "charge completion" email before the charge actually completed according to the power consumption plot. No latency at all - it's predictive. :?

that turns out to be rather close to the car's prediction of 3 hours and 50 minutes. The latter was determined prior to being plugged in. The actual timestamp start/stop times in the car were way off (as usual)

Yeah, any prediction the car makes before plugging in is meaningless - the time you saw was probably assuming a 40 amp EVSE (or who knows what.) The only time-to-charge prediction that matters is the one that makes the car decide when it will start charging. If you check the dash after plugging in (you've got to plug in fast or just wait to turn the car off until after you plug in), the start time the car displays will accurately reflect when the car will start charging (if it does in fact start...)

I suspect it would have started just after midnight as scheduled by Timer1 in Ravcharge. For an extended charge, with my battery's pre-charge SoC >50%, or a scheduled departure time after 7:11am, the car might have started charging on its own just after midnight as well.

Just to reiterate, I recommend setting your in-car departure time a few hours later than you actually want so that you never give the car the chance to start charging before the start date you specify in RavCharge. By the way, your RavCharge timer was off last night - not sure if you intended that or not for this test.

As for why the plot line slowly increases over time, maybe its because while the battery is under charge, it is slowly heating up and tempco for internal resistance of the battery pack is inversely proportional to temperature, so the EVSE's charging current slowly increases.

I would suspect something like that, or the TMS needing to run more as the battery heats up, but you would think that should mean of the total 32 amps more gets devoted to TMS and less to charging over time, not that the overall consumption changes. It's especially strange that the car wouldn't start off with the full 32 amps, given that 32 < the maximum of 40 that it can handle.

 

[fooljoe]

I assume Dsinned has a correction factor installed on the graph viewer as well. This explains the 8kW output using a 7.7kW charger.

Can you guys just output raw amps instead? That would be much less ambiguous.

 

[Dsinned]

Joe, you're correct about my Ravcharge Timer1 setting last night . . . I did leave it off intentionally. This was so I could evaluation the latest firmware update to my Gateway ECU in the back of the car. Sadly, as per usual, the scheduled charge I left up to the car itself, did not work within the timeline window I want or expected when PG&E rates are lowest (midnight to 7am). The WORSE thing about charge scheduling in the Rav 4 EV, is that, unlike Ravcharge, you cannot set a definitive start time, only a departure time.

By comparison, GM's charge scheduling works flawlessly in my 2012 Chevy VOLT using exactly the same Leviton EVSE that I use to charge my Rav4 EV.

Ordinarily, I do set the departure time a few hours later (8:50am on weekends) because the lowest rates on weekends are extended. But last night I was just trying to determine what would happen on a weekday with departure set for 6:50am. I think it would have worked if I only did a "normal" charge, which I have verified using Ravcharge a few times. Regardless of that, I think a beginning SoC < 50% might have something to do with it also. My beginning SoC this time was 49% (presumably based on doing a normal charge), but in fact, I did an extended charge to reach the full 41.8kWh capacity of the battery. I think in that context, the actual beginning SoC in Ravcharge was less than 49%, correct?

Nevertheless, Ravcharge is still the best "workaround" to overcome all the weirdness and unpredictability that scheduled charging entails in the car itself.

Thanks for all you have been doing in support of Ravcharge! :mrgreen:

 

[Dsinned]

I assume Dsinned has a correction factor installed on the graph viewer as well. This explains the 8kW output using a 7.7kW charger.

Can you guys just output raw amps instead? That would be much less ambiguous.

No, the HEM system I'm using only measures power consumption in watts. It is in fact measuring current consumption (and assuming a fixed voltage) but in the self monitoring system software, this is internally converted to AC power.

 

[miimura]

Just a heads up to people in PG&E territory that are on EV TOU rates. PG&E does not implement the current Daylight Savings time schedule. They are following the old schedule and call it out specifically in the rate tariffs.

DAYLIGHT SAVING TIME ADJUSTMENT: The time periods shown above will
begin and end one hour later for the period between the second Sunday in March
and the first Sunday in April, and for the period between the last Sunday in October
and the first Sunday in November.

So, for the week between the last Sunday in October (yesterday) and the first Sunday in November (Nov. 3), Off Peak weekday rates start one hour later, 1am for E-9 and midnight for EV.

 

[fooljoe]

I did an extended charge to reach the full 41.8kWh capacity of the battery. I think in that context, the actual beginning SoC in Ravcharge was less than 49%, correct

A full normal charge is about 84% of a full extended charge, so that means 49% SOC reported by RavCharge corresponds to 41% of an extended charge.

 

[Dsinned]

Just a heads up to people in PG&E territory that are on EV TOU rates. PG&E does not implement the current Daylight Savings time schedule. They are following the old schedule and call it out specifically in the rate tariffs.

DAYLIGHT SAVING TIME ADJUSTMENT: The time periods shown above will
begin and end one hour later for the period between the second Sunday in March
and the first Sunday in April, and for the period between the last Sunday in October
and the first Sunday in November.

So, for the week between the last Sunday in October (yesterday) and the first Sunday in November (Nov. 3), Off Peak weekday rates start one hour later, 1am for E-9 and midnight for EV.

Miimura, does this mean in order to get off-peak rates, TOU EV rate payers should reschedule their charging to begin after 1am and complete before 8am (E9A)? after we go back on standard pacific time this Sunday?

 

[miimura]

Just a heads up to people in PG&E territory that are on EV TOU rates. PG&E does not implement the current Daylight Savings time schedule. They are following the old schedule and call it out specifically in the rate tariffs.

DAYLIGHT SAVING TIME ADJUSTMENT: The time periods shown above will
begin and end one hour later for the period between the second Sunday in March
and the first Sunday in April, and for the period between the last Sunday in October
and the first Sunday in November.

So, for the week between the last Sunday in October (yesterday) and the first Sunday in November (Nov. 3), Off Peak weekday rates start one hour later, 1am for E-9 and midnight for EV.

Miimura, does this mean in order to get off-peak rates, TOU EV rate payers should reschedule their charging to begin after 1am and complete before 8am (E9A)? after we go back on standard pacific time this Sunday?

You should reschedule as you said during THIS week. Next week PG&E will be back in sync with the rest of us.

 

[Dsinned]

I did an extended charge to reach the full 41.8kWh capacity of the battery. I think in that context, the actual beginning SoC in Ravcharge was less than 49%, correct

A full normal charge is about 84% of a full extended charge, so that means 49% SOC reported by RavCharge corresponds to 41% of an extended charge.

Joe, there is an fortunate downside to this thing about a "percentage of a percentage" . . .

Last Sunday night I did an extended charge solely based on the in-car's timer/scheduler. It worked like always, however the charge initiated waaay earlier than expected; about 20 minutes BEFORE midnight. The resulting range provided by Entunes to RavCharge right after this was "118" miles. Since then, my wife has driven the car 2 or 3 times to run some errands and the Entunes range dropped to 96 miles. That's about right for approximately how many miles she drove it; maybe a little on the low side, because whenever she drives, the driving efficiency goes to hell on the center console.

So, here's the rub, RavCharge still says the car has a "100% Charge", and an "Approximate Charge: 33.4 kwh". (To be conservative, I preset RavCharge to a normal charge capacity of 33.4 kwh, although I realize some prefer 35.0 kwh instead). These parameters have not changed since right after the extended charge completed Monday morning, which of course is very misleading. Now that RavCharge's range is "adjustable" with respect to driving efficiency, I decreased it downward so that Entunes and RavCharge's ranges are somewhat in sync. However, RavCharge's SoC still stays at "100%".

I guess this is just the way things have to be, since RavCharge's default operation is based on a "Normal" charge with respect to approx. charge and SoC values displayed.

To help minimize confusion, can you add a new status field to be maintained somewhere on one of RavCharge's display screens, like, "Last charge selected: <NORM | EXTD>"?

 

[miimura]

I set my car to do an extended charge tonight. RavCharge says "To extended charge: 1:52 @ 235 volts 40 amps". So, it knows that I made that selection in the car and calculates the correct time on that basis from 79% Std SOC. ((1-0.79)*35 + 6.8 ) / (0.235*40) / 0.8 = 1.88 hr

The fact that Toyota systematically hides the content of the extended charge and only reports SOC as a percent of Standard Charge limits what RavCharge can do. So, it's not RavCharge's "default operation", it's Toyota's limitation.

 

[fooljoe]

I set my car to do an extended charge tonight. RavCharge says "To extended charge: 1:52 @ 235 volts 40 amps". So, it knows that I made that selection in the car and calculates the correct time on that basis from 79% Std SOC. ((1-0.79)*35 + 6.8 ) / (0.235*40) / 0.8 = 1.88 hr

The fact that Toyota systematically hides the content of the extended charge and only reports SOC as a percent of Standard Charge limits what RavCharge can do. So, it's not RavCharge's "default operation", it's Toyota's limitation.

Right - RavCharge can use the extended charge setting info for a forward-looking calculation like time-to-charge, but it can't really ever know what your last charge was without some guesswork and complicated periodic monitoring of each user's car (which I've thought plenty about, believe me...) So Dsinned's feature request is not nearly as simple as it sounds, and unfortunately the lack of SOC information between a normal and extended charge is something we're stuck with.