Using TM-Spy to see Model S data.

[David99]

So what are other's thoughts on what threshold should be considered "out of balance"? I haven't exceeded 90% charge in probably about 2 months and have watched my stable pack mV spread increase to about 16 mV. Personally, I'm sure this is by no means considered out of balance, but just looking for input. I'm curious to see how this spread tightens after forcing a pack balance.

I haven't range charged in many months and the difference is 4-5 mV. 16 mV seems high. Don't get me wrong, it's still very little drift and nothing to worry about, but it's 3-4 times of what I and others see. The app seems to take a while to scan and read all modules and it doesn't read all at the same time. It scans one after the other, so any change in power draw creates a false representation. Try turning the car off and wait a minute and see if the difference is less.

 

[apacheguy]

Yeah, the largest delta I see across my bricks is 5-6 mV and I regularly charge to 80%. The pack is very well balanced.

 

[garygid]

I see an interesting pattern in the brick voltages from each module while Supercharging. In each module, two bricks (a pair) report low, two report high, and two report in the middle ... like the 3 bears story ... with a very similar pattern reported from each module.

I suspect that the high charging currents somehow affect the measurement of the brick voltages in each module, and that the actual brick voltages are not really that different while supercharging. Perhaps it is due to some N millivolts lost in some longer internal brick interconnections, and those N mv are getting included in the measurement of the middle pair of bricks, and 2 x N mv are getting included in the measurement of the highest pair of bricks, and then being reported as if they were part of the brick voltage.

A smaller M mv lost in perhaps shorter connections between the two bricks in each pair might, in the same way, account for the smaller differences reported between the two bricks of each pair.

 

[earcandy]

73.9kWh highest readout. 3.25 years old. 44.4k miles. Original B version battery.

I have similar car info to Benjamin: 3 years old, 38.7k miles, Original B battery. Pack 71.5. (voltage cell spread of 5mV)

Interestingly today, when I plugged in, the car's rated range jumped pretty quickly from 189, to 193, as we had a cool night last night, however, the pack reading was 71.5 as well a couple days ago when temps were in the 90's.

I'll report back if anything changes this weekend, after our temps drop further, and I have a range charge/road trip planned.

 

[Andyw2100]

When the Tesla console is installed (not easily user removable like some other non-Tesla
consoles), it is reported that it becomes difficult (or impossible) to lower the open "mouth"
of the cubby enough to access the TDC in its "classic" position behind the cubby.

So, it was reported (or guessed) that the installation of the Tesla console required
routing the TDC's cable to the right (rather than to the left in the "classic") and
the TDC connector itself was poked out to the right, to be accessed behind the
"hockey stick" trim. I have not yet seen this myself, but some pictures would help.

Question is, when your console was installed, was the TDC moved, or not?
In any case, you, or someone generally more flexible than I, will need to locate
the TDC in YOUR car.

I am also learning ... Does this help you any?

Yes, this definitely helps.

For starters I definitely won't go exploring just for fun, but will wait until I have all the parts, if not longer. I may wind up waiting until I have a ranger out again who needs access to the TDC, and watch and/or ask about it.

Well, I wound up waiting until I had all the parts, and then some. (I also ordered a flat ribbon cable, in case everything worked, and there was room, and I could fish the ribbon cable through the back of the console and leave the reader there permanently. I would have been connecting the generic flat cable to Chris' customized cable.)

In any case, it appears my TDC was not moved, as I could not find it on either side after removing the hockey stick trim. I can't lower the front end of the cubby to look for it, as the installed Tesla center console is in the way. Oh well.

 

[garygid]

Andy...
Has your car been in for a service, annual or otherwise?
When it goes in, presumably the service guys will need to access the TDC, possibly removing the console and moving the TDC over to behind the hocky stick trim piece, and re-installing the console. Just a thought.

 

[Andyw2100]

Andy...
Has your car been in for a service, annual or otherwise?
When it goes in, presumably the service guys will need to access the TDC, possibly removing the console and moving the TDC over to behind the hocky stick trim piece, and re-installing the console. Just a thought.

The car has been in for one annual service, and is going in for another (and a bunch of other stuff) on Monday. One of the other things it is going in for is the Ludicrous upgrade, so I had been hoping to get some stats on the battery before it left, just on the outside chance things somehow get reset through the upgrade process.

Ironically, they actually are going to have to remove the console anyway, as one of the other things they'll be doing is swapping consoles for me. I managed to find an obeche matte console from someone who was selling one, and it will be replacing my piano black console, to better match my obeche matte interior.

I don't think the re-routing of the TDC is really something I should be asking them about, though. If the car comes back and it is accessible, then great. If not, I guess they'll just have to futz with it any time they need to get at it.

 

[Roadrunner13]

I have the same discrepancy David99. The app shows only about ~1/3 of the kWh used from my lifetime trip meter (i.e. battery discharge estimate is ~5,000kWh and the total lifetime trip meter shows ~13,000kWh used).
There are only 2 things I can think of that might explain this lower number from the app, since I'm still on my original battery pack.
1) The battery discharge kWh counter was reset when I last did a battery swap station visit at Harris Ranch in California
or
2) The battery discharge kWh counter was reset when my battery's contactors failed and were replaced.
Since those two events were pretty close in time last year, and reasonably about ~5000kWh of energy usage ago, I can't be sure if one or the other was root cause.

Now this is odd. My battery discharge estimate in the apps is also around 5 000kWh like yours as seen in the pics below, but my millage is way below yours at 12600 miles! (20300 km).
(My Trip B Meter was reset, so I don't have that number)
This is an E-type battery and my other surprise was to see my battery millage being identical to the car's millage.
As I bought the car as a CPO, I don't have the full car's history and the VIN of the car (65xxx) vs VIN of the battery (86xxx) are quite a distance appart, so I was looking forward to see if both millages matched or if my battery had been replaced...and it looks like it's the original!
Other numbers like a 77.4 kWh pack and a 6 mV spread seems par for the millage.

 

[garygid]

Andy...
Perhaps take a dozen or two doughnuts or other goodies, cashews, chocolate chip cookies, etc. to your service advisor and then say something like ... Please share these bribe-goodies with at least those working on my car. Oh, I just rembered something that I wasted to ask you about. I have been reading on the forum about some guys listening to a CAN bus on the special Tesla Diagnostics Connector. Could you show me where that connector is on my car? ... I did not want to break anything by trying to look behind the cubby. Or, maybe just take a picture of it while the car is "opened up"?

 

[Marcboy]

I updated my car to 2.16.17, and TM-SPY is not connecting to the BT module anymore.
Someone else has TN-SPY working with 2.16.17 ?

 

[garygid]

Which BT dongle are you using?
When you plug in the dongle, does its power LED light come on?

 

[Andyw2100]

Andy...
Perhaps take a dozen or two doughnuts or other goodies, cashews, chocolate chip cookies, etc. to your service advisor and then say something like ... Please share these bribe-goodies with at least those working on my car. Oh, I just rembered something that I wasted to ask you about. I have been reading on the forum about some guys listening to a CAN bus on the special Tesla Diagnostics Connector. Could you show me where that connector is on my car? ... I did not want to break anything by trying to look behind the cubby. Or, maybe just take a picture of it while the car is "opened up"?

I'm over 200 miles from my service center. They come for my car and bring it there on a trailer, and drop me off a loaner car. So while that idea could work if I were going to be at the service center, I think I'll just have to keep my fingers crossed that when it comes back, the TDC is more accessible.

 

[Marcboy]

I updated my car to 2.16.17, and TM-SPY is not connecting to the BT module anymore.
Someone else has TN-SPY working with 2.16.17 ?

All fixed now. Stupid mistake I won't disclose

 

[garygid]

Oh, come on, confess ... we can all use a good laugh ...
perhaps accidently using a regular ELM-type BT dongle with iOS?

How do I know about that, you might ask ...
In any case, glad that you found the problem.

The most commom problem with people calling a computer manufacturer's help hotline with the complaint that "My computer/printer/monitor/router/etc. is not working ..." ... Yep, the power cord is not plugged in.

Sometimes that sharing helps us get past some of our own challenges ... especially of growing older.

After the 2.16.17 revision install, I tried the LELink BT 4.0 dongle and TM-Spy connected and gathered data just fine.

 

[garygid]

Turbo3 has made great progress on the iOS version of TM-Spy Beta version, currently at v0.1.30, but still only available to Testers. He needs to work on other projects for a while, so ... Jim, we wish you well in your other projects.

Meanwhile, I will try to get some real, all-msgid Logs accumulated with CAN-Do, while charging, while Supercharging, while driving, and while parked.

I intend to add a new function to CAN-Do: After I accumulate a reasonably good Log file, clean it up a bit, and save it, I will be able to select some (or many) msgIDs, and write a coordinated set of over 300 X-files for others to look at. Each X-file, as Turbo3 has designed them, is an ASCII file of the observed CAN messages, all with just one single message ID.

I could make such a set of X-files available as one large zip file to sufficiently serious and interested private parties who have access to TM-Spy (currently just the IOS version). I will try to accomplish this "soon".

Then, even those without an Adapter cable or an appropriate OBD dongle, or reluctant to dig into their car to access the Tesla Diagnostic Connector, could have real, parallel msgid data files to look into, investigate, examine, or just "play with", even while totally off-line, not even in the car, ... or even those without a Tesla, like Turbo3.

Would any TM-Spy Tester like to try this, keeping the data for their own personal, non-commercial use only, at least for now?

 

[garygid]

Anatomy of an X-file, from observation, apparently is:

1. A Header Line: The Start-Recording DateTime Value (a unix-type datetime value, the number of miliseconds past 12:00:00 on Jan 1, 1970, local time), one space, and the number (N) of message frames in the file.

2. The Message Frame Lines: One CAN message per line, in Hex characters (1 ... 9, A ... F), with a 3-char MsgID, eight 2-char message data bytes, and two 2-char iOS-device acceleration bytes.

3. The Ending Line: Just the End-Recording DateTime Value, in the same format as the DateTime in the Header Line.

The file name for the X-file: X_MID_YYMMDDHHmm.txt where MID is the recorded MsgID in Hex. The Year Month Day and Hour minute is probably when the file was written (when the gathered data was Saved) but I am not sure ... yet.

 

[Turbo3]

Anatomy of an X-file, from observation, apparently is:

1. A Header Line: The Start-Recording DateTime Value (a unix-type datetime value, the number of miliseconds past 12:00:00 on Jan 1, 1970, local time), one space, and the number (N) of message frames in the file.

2. The Message Frame Lines: One CAN message per line, in Hex characters (1 ... 9, A ... F), with a 3-char MsgID, eight 2-char message data bytes, and two 2-char iOS-device acceleration bytes.

3. The Ending Line: Just the End-Recording DateTime Value, in the same format as the DateTime in the Header Line.

The file name for the X-file: X_MID_YYMMDDHHmm.txt where MID is the recorded MsgID in Hex. The Year Month Day and Hour minute is probably when the file was written (when the gathered data was Saved) but I am not sure ... yet.

You can use a start time of 0 which is what I use when making up a file. Then the end time is just the number of milliseconds it took to capture the data. If you want to default to 100 frames per second just multiply the number of frames in the file by 10 and use that as the end time.

 

[garygid]

Turbo3, Good suggestion, thanks.

I suppose you could just have just put the duration in ms on the first line, then the N message frames following, and not need to write the last line datetime value at all?

Is the date and time in the X-file file name the time that the data was Saved, or the time the data gathering terminated?

 

[garygid]

Using two new features in v0.1.30 of the iOS TM-Spy app, I recently captured the following 0x00E data, which appeares to be related to the position of the steering mechanism.

The first new feature is the use of the pseudo-msgID 0x800 in a Recipe to match any selected real msgID for doing the frame capture. In this case, I selected the first on the app's list, 00E, for this frame capture. This feature allows one Recipe to be used to capture any listed msgID. Many msgIDs have been observed on CAN3, and are listed, but only a few are "known".

The second feature is using the Byte C "filter" settings to actually capture just 1 out of each N observed frames, cutting down the amount of data in the X-file "log". Here, I had N set to 10, to discard all but one of each 10 observed frames. The Byte C# is set to 9 to enable this feature, and Equals is set to N, in this case 10, but it could be anything from 2 to at least 250, probably more. This feature enables making very long recordings, like 24 hours monitoring the DC to DC converter and the 12v bus voltage.

Both appeared to work well.

The data below goes high for a right turn, and low for turning left. The little bump up at the end was me turning slowly right into my friend's driveway. Not ground-shaking data, but vaguely interesting to nerds, perhaps. Here, I am using a data-specific Recipe in an attempt to scale the raw data to something like degrees of turn of the steering wheel, perhaps.

Yes, about an 11 minute drive.
Cheers, Gary

 

[garygid]

Looking a bit more at the data in the 00E frames, I observe a second trace in bytes D3 and D4, as shown below. To see the detail better, I have zoomed in (to 4x) on the leftmost data ... What do you see?

To me, the upper (red) data looks like a derivative of (a mathamatic term for the slope of) the lower data. So, it might possibly represent the power or current used to move the steering mechanism by the power steering. Yes, it would be nice to know what it really represents.

Data bytes D5 and D6 seem to be essentially constant. Bytes D7 and D8 seem to be some blocky square shaped waveform, with some unknown (to me) meaning.

Cheers.