[Ovmsdev] Roadster cooldown
Stephen Casner
casner at acm.org
Sat Nov 3 14:40:53 HKT 2018
Mark,
I think the initial abort due to "charging stopped" may just be a
consequence of initial system state. When I reboot OVMS, if the
charge cable is plugged in and the car is waiting for a timed charge,
I get a "charging stopped" notification on the iPhone app. It seems
plausible that this instance was related because I had just loaded
the newly built firmware and the car was in the state of waiting for a
timed charge.
One curiosity from the cooldown test is that the car was awake with
the coolant pump running an hour after the test and perhaps the next
morning as well. This even though the battery was cool, about 19C.
The cooland pump runs continuously if the battery temp is above a
threshold, but does it also run if the battery temp is below some
threshold?
-- Steve
On Thu, 1 Nov 2018, Mark Webb-Johnson wrote:
> Steve,
>
> The cooldown itself seems correct, but I don't like that initial abort.
>
> I suspect a bug. Perhaps we send the charge start message, but then get a CAN bus message saying charging not started, followed by another to say it has actually started. If we picked up the first, we could treat it as an abort. We handle the 'Cooldown: Aborted as car stopped charging' in the ticker60, so that is not very likely, but still possible.
>
> Perhaps we should not test for a charge stopped until at least 1 minute into the cooldown? That would not be hard to code.
>
> Regards, Mark.
>
> > On 30 Oct 2018, at 1:45 PM, Stephen Casner <casner at acm.org> wrote:
> >
> > Mark,
> >
> > As I proposed last night, I pulled your changes this evening to build
> > the firmware and test cooldown. Since my battery was already down to
> > 20C I changed the limit to 19C for this test. Here's the result:
> >
> > OVMS# config set xtr cooldown.templimit 19
> > Parameter has been set.
> > OVMS# charge cooldown
> > Cooldown has been started
> > I (120425) v-teslaroadster: Cooldown: 0 cycles 20C/19C 12A (with 60 minute(s) remaining)
> > I (120425) v-teslaroadster: Cooldown: Aborted as car stopped charging
> > I (124555) ovms-server-v2: Send MP-0 PAStandard - Charge Stopped|SOC: 83.0%|Ideal range: 197M|Est. range: 183M|ODO: 53271.4M|
> > OVMS# charge cooldown
> > Cooldown has been started
> > I (162265) ovms-server-v2: Send MP-0 PIRange - Charging|247.0V/1.0A|SOC: 76.0%|Ideal range: 228M|Est. range: 212M|ODO: 53271.4M|
> > I (180425) v-teslaroadster: Cooldown: 0 cycles 20C/19C 12A (with 60 minute(s) remaining)
> > I (240425) v-teslaroadster: Cooldown: 0 cycles 19C/19C 12A (with 59 minute(s) remaining)
> > I (240425) v-teslaroadster: Cooldown: Cycle 0 cooldown completed
> > I (240745) ovms-server-v2: Send MP-0 PARange - Charge Stopped|SOC: 76.0%|Ideal range: 228M|Est. range: 212M|ODO: 53271.4M|CAC: 184.4Ah|SOH: 83%|
> > OVMS# config set xtr cooldown.templimit 25
> >
> > Note that I needed to issue the command twice. The first time it said
> > the car stopped charging, but that may be related to the hack you
> > needed to put in to say charging stopped to work around a limitation
> > in the v2 server (if I'm remembering that correctly).
> >
> > The second command took effect as expected. The VDS indicated
> > charging at 12A during the cooldown, and when it completed charging
> > stopped and the VDS went back to preparing for timed charge at 32A.
> >
> > Thanks for the fixes.
> >
> > -- Steve
> >
> > On Mon, 29 Oct 2018, Mark Webb-Johnson wrote:
> >
> >> Steve and I worked through this, and found a couple of things:
> >>
> >> Tesla Roadster US firmware doesn't seem to want to set the charge limit below 12A. Our CAN bus work seems fine - just the VMS ignores anything below 12A on Steve's car. Maybe a 1.5 thing, or perhaps US firmware. Anyway, OVMS v2 used 13A for cooldown, so I think we can change to 12A without issue. I added this as a default, and made it configurable as 'xtr cooldown.current 12'.
> >>
> >> A bug meant that cooldown didn't stop properly after it's work, even if the previous state was that car was not charging before cooldown. Simple fix.
> >>
> >> These fixes will go out in the next nightly build.
> >>
> >> Regards, Mark.
> >>
> >> P.S. ovms_shell worked amazingly well for this.
> >>
> >>> On 29 Oct 2018, at 11:13 AM, Stephen Casner <casner at acm.org> wrote:
> >>>
> >>> Mark,
> >>>
> >>> I could do this test any time in the next four hours if that is
> >>> convenient for you, or during your afternoon in the next couple of
> >>> days.
> >>>
> >>> -- Steve
> >>>
> >>> On Mon, 29 Oct 2018, Mark Webb-Johnson wrote:
> >>>
> >>>> Steve,
> >>>>
> >>>> I see:
> >>>>
> >>>> OVMS# charge cooldown
> >>>> Cooldown has been started
> >>>> I (649473622) ovms-server-v2: Send MP-0 PIRange - Charging|247.0V/1.0A|SOC: 23.0%|Ideal range: 69M|Est. range: 64M|ODO: 53271.4M|CAC: 187.0Ah|SOH: 70%|
> >>>> I (649500622) v-teslaroadster: Cooldown: 0 cycles 33C/31C (with 60 minute(s) remaining)
> >>>> I (649560622) v-teslaroadster: Cooldown: 0 cycles 32C/31C (with 59 minute(s) remaining)
> >>>> I (649620622) v-teslaroadster: Cooldown: 0 cycles 32C/31C (with 58 minute(s) remaining)
> >>>> I (649637002) ovms-server-v2: Incoming Msg: MP-0 Z1
> >>>> I (649637002) ovms-server-v2: One or more peers have connected
> >>>> I (649637622) ovms-server-v2: Send MP-0 S24.0,M,238,32,charging,range,70,65,70,2,0,0,5,1,3,0,0,0,184.42,274,0,0,0,0,0,0,0,0,0,0,0,-4.15,0.00,59
> >>>> I (649637622) ovms-server-v2: Send MP-0 D93,0,5,35,32,32,1965,532713,0,33239,21,2,1,1,12.73,0,12.6,128,35,0,0
> >>>> I (649680622) v-teslaroadster: Cooldown: 0 cycles 31C/31C (with 57 minute(s) remaining)
> >>>> I (649680622) v-teslaroadster: Cooldown: Cycle 0 cooldown completed
> >>>> I (649879822) ovms-server-v2: Send MP-0 PAStandard - Charge Stopped|SOC: 17.0%|Ideal range: 41M|Est. range: 38M|ODO: 53271.4M|CAC: 184.4Ah|SOH: 59%|
> >>>>
> >>>> It seems that the battery was at 33C when cooldown started. Took about 3 minutes for it to come down to 31C and for cooldown to then stop.
> >>>>
> >>>> Not sure about the 32A. It certainly seems that way. The code has:
> >>>>
> >>>> void OvmsVehicleTeslaRoadster::Ticker1(uint32_t ticker)
> >>>> ...
> >>>> if ((m_cooldown_running)&&(StandardMetrics.ms_v_charge_inprogress->AsBool()))
> >>>> ...
> >>>> if (m_cooldown_recycle_ticker > 0)
> >>>> ...
> >>>> if (StandardMetrics.ms_v_charge_climit->AsInt() != 10)
> >>>> {
> >>>> // 10A charge when cooling down
> >>>> ESP_LOGI(TAG, "Cooldown: Cycle %d fix charge current to 10A",m_cooldown_cycles_done);
> >>>> CommandSetChargeCurrent(10);
> >>>> }
> >>>>
> >>>> That should continually try to change the charge current, so long as it is not 10A (while doing a cooldown, charge in progress, etc).
> >>>>
> >>>> Your car is a 1.5, right? Perhaps we have charge limit incorrect for that model?
> >>>>
> >>>> I think the simplest would be to be able to get into the car remotely during a cooldown. Just OVMS server v2 should be fine. Probably worth setting 'xtr cooldown.templimit 25' to it doesn't stop too quickly, then arrange a time when you can start this, and I can monitor remotely. Need to see how all those metrics change and what other messages are produced.
> >>>>
> >>>> Regards, Mark.
> >>>>
> >>>>> On 26 Oct 2018, at 2:04 PM, Stephen Casner <casner at acm.org> wrote:
> >>>>>
> >>>>> After several weeks delay, I had a chance to try "charge cooldown"
> >>>>> again after driving 200 miles to return home today (after a very nice
> >>>>> drive yesterday down the rugged California coast Hwy 1 with the top off and
> >>>>> clear road ahead 90% of the time).
> >>>>>
> >>>>> As requested, with the charge cable plugged in and not charging (which
> >>>>> means the VDS is in the state of waiting for timed charge) I connected
> >>>>> the USB console cable and issued the "charge cooldown" command. This
> >>>>> did cause charging to start in Range mode, but the current was the
> >>>>> full 32A available in my garage, not 12A as I thought was the design
> >>>>> for the cooldown procedure. The log is attached; it looks like the
> >>>>> cooldown was deemed completed after a few minutes, but then the charge
> >>>>> mode switched back to Standard and charging continued at 32A for
> >>>>> another few minutes until I stopped it with the button on the VDS.
> >>>>> I'll let charging continue and complete during the normal timed charge
> >>>>> period.
> >>>>>
> >>>>> -- Steve
> >>>>>
> >>>>> On Fri, 14 Sep 2018, Stephen Casner wrote:
> >>>>>
> >>>>>> Mark,
> >>>>>>
> >>>>>> Not to worry. This is not a critical problem for me.
> >>>>>>
> >>>>>> I did find 'vehicle status' but nothing in the output conveyed to me
> >>>>>> whether cooldown was in progress. Sorry I didn't do a screen capture.
> >>>>>>
> >>>>>> The car was plugged in but not charging when I gave the 'charge
> >>>>>> cooldown' command, and the charge start time set on the VDS had not
> >>>>>> been reached, so I don't know why it would have been in charging
> >>>>>> state.
> >>>>>>
> >>>>>> I will try again with my laptop connected to the console rather than
> >>>>>> just trying to run this from the web UI with my phone. Since the day
> >>>>>> of the previous attempt it's been cool enough here and I have not
> >>>>>> driven far enough for the battery temperature to get above 31. I
> >>>>>> could try adjusting the templimit.
> >>>>>>
> >>>>>> -- Steve
> >>>>>>
> >>>>>> On Fri, 14 Sep 2018, Mark Webb-Johnson wrote:
> >>>>>>
> >>>>>>> Steve,
> >>>>>>>
> >>>>>>> Sorry for the delay in reply. I have been overwhelmed with work...
> >>>>>>>
> >>>>>>> You can check the status of cooldown with 'vehicle status'. In addition, an info-level log line is output once every minute while cooldown is in progress.
> >>>>>>>
> >>>>>>> The current roadster implementation was copied from v2. It is largely untested (beyond simulation on my bench). When the cooldown command is given, it looks to see if a charge is currently in progress and if so remembers the details of that charge (mode, current, etc). After cooldown has completed, it will resume any previous charge that was in progress. So, I think if you want to do a cooldown without a subsequent charge, then the sequence would be to stop charge then cooldown.
> >>>>>>>
> >>>>>>> Cooldown does use two parameters:
> >>>>>>>
> >>>>>>> xtr cooldown.timelimit 60
> >>>>>>> xtr cooldown.templimit 31
> >>>>>>>
> >>>>>>> If the cooldown has been running more than cooldown.timelimit, or the battery temperature falls below cooldown.templimit, it will terminate the cooldown.
> >>>>>>>
> >>>>>>> Can you try again, from a plugged in but not charging state, and check 'vehicle status' and the info logs to see progress?
> >>>>>>>
> >>>>>>> Regards, Mark.
> >>>>>>>
> >>>>>>>> On 11 Sep 2018, at 8:47 AM, Stephen Casner <casner at acm.org> wrote:
> >>>>>>>>
> >>>>>>>> Last evening I returned home from a trip and decided to try the
> >>>>>>>> "charge cooldown" command to cool the battery on my Roadster. I had
> >>>>>>>> not tried it before, and I think it may not have worked correctly.
> >>>>>>>>
> >>>>>>>> I issued the "charge cooldown" command using the shell page of the web
> >>>>>>>> UI. The response was a line indicating that the cooldown had started,
> >>>>>>>> and I got a notification through server v2 that the car was charging
> >>>>>>>> in Range mode. But later the status page in the iPhone app indicated
> >>>>>>>> that the car was charging in Standard mode at 32A, which is the full
> >>>>>>>> rate from my EVSE. Over the course of half an hour the SOC increased
> >>>>>>>> from 42% to 50%, at which point I manually stopped the charge.
> >>>>>>>>
> >>>>>>>> I thought the cooldown procedure was supposed to charge in Range mode
> >>>>>>>> at 12A and not significantly add to the battery charge level. I have
> >>>>>>>> not studied the cooldown code, but after a quick look I see that the
> >>>>>>>> procedure is more complicated than one might expect, stepping through
> >>>>>>>> multiple cycles as needed.
> >>>>>>>>
> >>>>>>>> Did I do something incorrectly? Or do these symptoms indicate that
> >>>>>>>> there might be a problem with the implementation?
> >>>>>>>>
> >>>>>>>> While the charge was in progress I blindly tried a "charge status"
> >>>>>>>> command hoping that it would tell me about the progress of cooldown,
> >>>>>>>> but alas that subcommand does not exist.
> >>>>>>>>
> >>>>>>>> -- Steve
> > _______________________________________________
> > OvmsDev mailing list
> > OvmsDev at lists.openvehicles.com
> > http://lists.openvehicles.com/mailman/listinfo/ovmsdev
>
> _______________________________________________
> OvmsDev mailing list
> OvmsDev at lists.openvehicles.com
> http://lists.openvehicles.com/mailman/listinfo/ovmsdev
>
>
More information about the OvmsDev
mailing list