Odd charging behavior (seen in logs)

[gregd]

I guess I'm m just confused how the charger can initiate the charge. Does it send some sort of signal to the car to engage the contractors?

Yes, that's the "pilot" signal that was mentioned in the thread, which is sent by the car to the wall charger to signal its desires. My thought here is that "its desires" are either being mis-interpreted or garbled somehow by the charger, causing the charger to cycle power to the car, making the car think "Hey, I just got plugged in; time to charge!" Something like that...

 

[hcsharp]

Actually the car does not send the pilot signal to the wall charger. It's the other way around. The EVSE generates the pilot signal and the car manipulates it, along with the EVSE, to establish a form of 2-way communication. @supersnoop is correct that the wall charger can't really initiate the charge. But it can tell the car to stop charging, and a few seconds later tell it it's OK to start again. Wash, rinse, repeat...

 

[hjr]

8:12:09 PM Charged full in standard mode, 32A 240V up to SOC 97% and 181 miles
Now on its 27th cycle.
8:35:53 PM SOC 99% 185 miles

Looks like the replacement charger does not work either.

On any other charger I have used the Battery goes to 178 miles, and stops...It connects and disconnects only 1 time. No "balancing" of the bricks by re-connecting to the charger. the balancing is an internal PEM/Battery function.

I also tried to stop the charge using my OVMS. The charge stops, waits a minuteand then starts to cycle all over again.

I will report to Clipper Creek.

I called Tesla in California today. There is nobody in "Tech Support" who can answer these questions, nor were they willing/able to connect me to an engineer who knew anything about Roadster charging and handshaking. They said I had to go through my local service center (who also do not know much about this). He said only tech support and the service center are "front facing" - I wanted to tell him to kiss my "rear facing" ... The guy on the phone sounded very "large company" and not from the Tesla I fell in love with in the early days when you could just about get Elon on the phone.

They said that since a non-authorized charger was being used they could not help and Clipper Creek needed to call their contacts at Tesla. Harrumphh....

 

[gregd]

Actually the car does not send the pilot signal to the wall charger. It's the other way around. The EVSE generates the pilot signal and the car manipulates it, along with the EVSE, to establish a form of 2-way communication. @supersnoop is correct that the wall charger can't really initiate the charge. But it can tell the car to stop charging, and a few seconds later tell it it's OK to start again. Wash, rinse, repeat...

Ah, ok. Thanks for clarifying! I guess that makes sense. But it's still that manipulation that is controlled by the car (which is why I was thinking backwards...)

So the key words in what you wrote are "OK to start". If the car is done charging, it shouldn't take the bait and actually start, right? Somebody flinched here, and @hjr's observation is that the Clipper Creek charger cycles its main relay, so I'm guessing it's sensing something that isn't (or isn't intended to) be there.


Question to @hjr: when the CC charger cycles the main relay, is it a quick click-thunk, or is there a significant, perhaps variable delay (many seconds) between the off- and on- transitions?

Something probably unrelated, but while I'm speculating, I thought I'd toss this out there too... I took a trip in the car today, and returned just before dinner. A little warm out, but not too bad. Plugged the car in to the home system (Mobile adapter to dryer plug), and it started charging as normal. Then the main house A/C came on, which seems to be ok, load-wise, but I have a dual system and thought if all three get going it might be too much. I think they're all on the same 50A breaker... Anyway, I hit the Stop button on the car VDS, but left the car plugged in. A little while later, I hear the car charging. I didn't start it, and it's set to charge on plug-in. I'm guessing there was a glitch on the 240v line when the house A/C cycled, tricking the Mobile connector / car to start charging as if it just got plugged in. So, something really seems to have a hair trigger... Perhaps these CCreek HCS-40 installations have something glitchy on their power or pilot signal lines?

Edit: Or, perhaps the daily SOC check just kicked in and started the charge... Will check the logs when it's done.

Er, double Edit: Can't be the daily SOC check - that occurs 24hrs after the last charge ends. That won't be until tomorrow... Back to the glitch theory?

 

[hjr]

Henry: I know your are a Roadster specialist... But do Model S and X handshake differently from Roadster when they are plugge into an EVSE ? I suspect that the CC charger works fine with my X because the pilot is either more stable at teh 6V and 9V levels of the square wave, or they are set at a more precisely controlled voltage. I suspect the Roadtser voltages might be higher than 6V and lower than 9V.. And the CC charger is looking for a more precise voltage cue.

Maybe the CC needs to be set for a touch hggher than 6V and lower than 9V?

 

[hjr]

Question to @hjr: when the CC charger cycles the main relay, is it a quick click-thunk, or is there a significant, perhaps variable delay (many seconds) between the off- and on- transitions?

Car stops charging and relay of EVSE and car click off. After 40 seconds - 1 minute the EVSE and car relays click back on.

As Henry said Wash - Spin - repeat.

 

[gregd]

If anyone besides @hjr, @hcsharp, and I are watching this, and for posterity purposes, a quick update...

I brought my Roadster over to Clipper Creek, and we put it on an instrumented HCS-40 charger via my CAN-JR. All looked pretty normal; charging at 32 A / 240V. Based on my earlier experience with the "Stop" button (see above), I thought we could try a quick experiment while waiting for the car to get to a full Standard charge. Immediately after pushing Stop the car stopped as expected, but then seconds later, went to "Preparing to Charge" and then right back to Charging again. Hit it again. Same result. Problem reproduced! Tried it on Jason's car (using his CAN-JR, so different car and adapter), and got the same result. Yea!

Once the car fully charged in Standard mode, we again experienced the unprovoked repeated short cycling, as before. Tried doing a full disconnect of the charging cable, and closing the charge port cover on the car, then plugged everything back in. The cycling continued. We were hoping that the reset might give us only the expected single cycle. Nope.

They've got scope traces of the whole test, and can now reproduce it at will (via the Stop button), so we're hopeful that the engineering team can get to the root cause of the behavior.

At this point, I am suspecting that the issue is local to the HCS-40 charger, and the 2.x Roadsters, but perhaps not all. If anyone has seen this sort of behavior, either the Stop button's effect being only temporary, or the car doing quick cycling on its own, please let one of us know. Clipper Creek is truly interested in getting to the bottom of this (thanks CC!).

 

[hcsharp]

Henry: I know your are a Roadster specialist... But do Model S and X handshake differently from Roadster when they are plugge into an EVSE ? I suspect that the CC charger works fine with my X because the pilot is either more stable at teh 6V and 9V levels of the square wave, or they are set at a more precisely controlled voltage. I suspect the Roadtser voltages might be higher than 6V and lower than 9V.. And the CC charger is looking for a more precise voltage cue.

Maybe the CC needs to be set for a touch hggher than 6V and lower than 9V?

That was essentially my suspected theory but was mostly proven not to be the problem when they got the scope on it. Indeed the Roadster pilot voltages for 6 and 9v are more like 7 and 10v but the CC charger tests for 7.5v so that's not what's causing this. I have a new theory now, although it's a long shot.

 

[gregd]

That was essentially my suspected theory but was mostly proven not to be the problem when they got the scope on it. Indeed the Roadster pilot voltages for 6 and 9v are more like 7 and 10v but the CC charger tests for 7.5v so that's not what's causing this. I have a new theory now, although it's a long shot.

And...?

Don't leave us hanging! Anything I can do to verify it with them?

 

[CSPHD]

If anyone besides @hjr, @hcsharp, and I are watching this, and for posterity purposes, a quick update...

At this point, I am suspecting that the issue is local to the HCS-40 charger, and the 2.x Roadsters, but perhaps not all. If anyone has seen this sort of behavior, either the Stop button's effect being only temporary, or the car doing quick cycling on its own, please let one of us know. Clipper Creek is truly interested in getting to the bottom of this (thanks CC!).

I can confirm something similar happens to me with my HCS-40 and 2.5 sport. I bought the HCS-40 because my UMC doesn't work when the ambient temperature is above about 80 (it shuts down due to excessive heat build-up in the fully potted electronics).

When it is cool enough for the UMC, I always get a green light around the inlet in the morning. With either my OpenEVSE or HCS-40 it is always blue with a "preparing to charge" on the VDS. I've stopped standard charges a couple of times before completion on either the HCS-40 or OpenEVSE and then come back later to find the car charging.

I do have DMC and over/undervoltage errors sometimes, but I've always assumed it is because the local utility thinks plus/minus 12V from 120VAC is good enough.

 

[gregd]

Ooh, tell me more about the OpenEVSE. Which model, what options, etc.? I was just about to order their 50A Deluxe kit, so that I wouldn't have to rely on (or wear out) my MC240. If it doesn't work with the 2.x Roadsters, I need to know!

The Mobile Connector works just fine for me right now - not that hot in the garage, I suppose - but it's way more expensive to replace than the OpenEVSE, and being portable, I might need it for road trips.

 

[CSPHD]

I've got the 50A version with the V3 board. I've had it since May of '15. It charges the car in the expected time. I usually charge overnight to get the off-peak rate. I've always thought it was curious that the HCS-40 and the OpenEVSE didn't end with a green light in the inlet like the UMC. But since everything appeared normal, I didn't think much more about it.

However, now that a couple of other folks have noticed odd things with the HCS-40, it makes me wonder. I've pulled logs from the car today and am upgrading the latest OpenEVSE firmware to see if I can learn more about what is happening.

Regarding the OpenEVSE, it was easy to build (disclosure - I am a computer scientist with a serious electronics habit). I have not had problems with it failing in the heat or humidity. Summers around here can get into the low 100s with the heat index June-September. I use hcsharp's can to connect the OpenEVSE J1772 to the car.

 

[gregd]

I've got the 50A version with the V2 board. I've had it since May of '15. It charges the car in the expected time. I usually charge overnight to get the off-peak rate. I've always thought it was curious that the HCS-40 and the OpenEVSE didn't end with a green light in the inlet like the UMC. But since everything appeared normal, I didn't think much more about it.

However, now that a couple of other folks have noticed odd things with the HCS-40, it makes me wonder. I've pulled logs from the car today and am upgrading the latest OpenEVSE firmware to see if I can learn more about what is happening.

Regarding the OpenEVSE, it was easy to build (disclosure - I am a computer scientist with a serious electronics habit). I have not had problems with it failing in the heat or humidity. Summers around here can get into the low 100s with the heat index June-September. I use hcsharp's can to connect the OpenEVSE J1772 to the car.

Thanks for the info. I guess the color of the charging ring isn't too critical, in itself, but you are right that there's probably something going on behind the scenes that we should understand. I'm intending to do another charging experiment at the HCS-40 in town on Friday evening, and will take a special look to see what the final color is. Didn't bother to notice it before. The car did say "Done Charging" on the VDS, however. What color did you get?

Let us know what you find in the logs. Specifically worrying would be if you see a charge session end with a bunch of little mini-charges (30 seconds to a minute in length each) before settling down.

Long-term owner of a soldering iron here, too. Should be fun to build.

 

[CSPHD]

Let us know what you find in the logs. Specifically worrying would be if you see a charge session end with a bunch of little mini-charges (30 seconds to a minute in length each) before settling down.

After examining the logs from the car I can now confirm that I see the same exact behavior on the Clipper Creek HCS-40 as @gregd. The car completes a standard charge and then begins a series of "mini-charges". This might explain why I don't see a green light around the inlet after using the HCS-40. In addition to the lack of a green light, the VDS says "Preparing to charge" instead of "Done".

I also looked at OpenEVSE charging sessions. It does *not* have the "mini-sessions" after a complete standard charge. I also do not get a green light around the inlet with the OpenEVSE. I can't recall at the moment if the VDS says "Done" or "Preparing to charge" after the OpenEVSE charge. I've updated the firmware on my OpenEVSE today and will watch the next charge cycle.

A shout out to @tomsax for the awesome VMSParser!

 

[hcsharp]

Thanks for the info. I guess the color of the charging ring isn't too critical, in itself, but you are right that there's probably something going on behind the scenes that we should understand.

After examining the logs from the car I can now confirm that I see the same exact behavior on the Clipper Creek HCS-40 as @gregd. The car completes a standard charge and then begins a series of "mini-charges".

The green ring goes away after a few minutes so if you aren't there shortly after it finishes you might not see it. The "Preparing to charge" message vs "Done..." might be significant.

What you should do is post a question to the OpenEVSE development/support team, or just read the code, to determine what it does to the pilot signal at the end of a charge (EOC). Consider EOC to be after the EV allows the pilot signal to go up to 9v from 6v on the positive side of the square wave. If the EVSE briefly stops the square wave so you have a solid +9v, or drops it to 0v before re-starting the square wave, then you will see the cycling problem that some of you are experiencing.

 

[CSPHD]

The green ring goes away after a few minutes so if you aren't there shortly after it finishes you might not see it. The "Preparing to charge" message vs "Done..." might be significant.

I guess that means I'm just catching the end of the UMC charge sessions in time. That would also explain the one or two times I've seen the green inlet color with my OpenEVSE when I just happened to be in the garage as the cycle ended.

What you should do is post a question to the OpenEVSE development/support team, or just read the code, to determine what it does to the pilot signal at the end of a charge (EOC). Consider EOC to be after the EV allows the pilot signal to go up to 9v from 6v on the positive side of the square wave. If the EVSE briefly stops the square wave so you have a solid +9v, or drops it to 0v before re-starting the square wave, then you will see the cycling problem that some of you are experiencing.

My Clipper Creek HCS-40 exhibits the "mini-charges" issue that @gregd is describing. My OpenEVSE does not do this.

 

[CSPHD]

The car's logs show that when charging with the Clipper Creek HCS-40, my scheduled charge begins, ramps up the current to 30 amps, and finally tapers off at the end of the standard charge as expected. Then seconds after the charge cycle is complete, another charge will start. Current will ramp up again and then the charge will terminate after a few minutes.

The most recent charge on the HCS-40 charged as expected from 26% SOC to 84% SOC then ended. Seconds later a new charge cycle started and ran for 5.5 minutes then ended. Seconds later a new charge cycle started and ran for 2 minutes then stopped. Seconds later a new charge started and ran for 2 minutes. This pattern repeated for several hours where the car would charge for 2 minutes, stop, and then start charging again. It does eventually stop looping.

 

[spaceballs]

My Clipper Creek HCS-40P also has these "mini-charges" on my 2.0. These additional new charge cycles ever ~2-10 minutes are adding unnecessary wear on the charging relays inside the PEM (in my view) and looks like I will be using a different charger until this is fixed.

I can also reproduce it with the issue if I hit the Stop button on the display, it will automatically restart the charge. My guess is when the car stop charging, the pilot signal somehow gets temporary interrupted (when main relay in the clipper creek unit disconnects?) thus the car thinks the plug has been unplugged and then re-plugged back in.

 

[gregd]

My Clipper Creek HCS-40P also has these "mini-charges" on my 2.0. These additional new charge cycles ever ~2-10 minutes are adding unnecessary wear on the charging relays inside the PEM (in my view) and looks like I will be using a different charger until this is fixed.

I can also reproduce it with the issue if I hit the Stop button on the display, it will automatically restart the charge. My guess is when the car stop charging, the pilot signal somehow gets temporary interrupted (when main relay in the clipper creek unit disconnects?) thus the car thinks the plug has been unplugged and then re-plugged back in.

Exactly.

It seems there are a number of folks out there that are using the Clipper Creek HCS-40 charger, and have this problem. But Clipper Creek thinks that there are only three (@hjr, Jason at Clipper Creek, and me), and have refused so far to look for a fix. Their belief is that the car is to blame, which may be true, but unless we can convince Tesla to implement a fix (if that's even possible), it will be up to Clipper Creek to put in a work-around. That, or lose potential customers. I, for one, have decided not to go with (literally) the home town supplier due to this.

So, the several others of you who are seeing this behavior with the HCS-40, and have a spare moment, could you contact Clipper Creek and add your vote for a fix / work-around? I agree that the multiple extraneous charging cycles are a stress on the EVSE, car, and worst of all, the battery.

 

[CSPHD]

So, the several others of you who are seeing this behavior with the HCS-40, and have a spare moment, could you contact Clipper Creek and add your vote for a fix / work-around? I agree that the multiple extraneous charging cycles are a stress on the EVSE, car, and worst of all, the battery.

I contacted Clipper Creek this past week and am in the process of gathering some additional information to share with them about the issue. Based on what has been reported to them, the issue is restricted to the Roadster. The technician (correctly) pointed out that the Roadster EVSE charging protocol pre-dates the J1772 standard and is therefore not 100% compliant with the 2010 standard.

They have been responsive and genuinely interested in assisting me with my questions. If the problem is as suspected, the CCID test at the end of the cycle causes the Roadster to start a new session, fixing it could cause issues for other folks with J1772 compliant vehicles. Charging the Model S with HCS-40 works fine for me.