Chassis CAN Logging To ASCII Text Plus Graphing

[lolachampcar]

race cars are long gone Kind of a been there done that type of thing. The planes are much cheaper to play with (I do all my own work) and they hold their value.

Lots'o pics here
Cars

The car I was running at the end-
99Shell Atlanta 1 - YouTube

__________________________________________________________________________________________________________

Logger firmware updating

The logger uses a drag and drop firmware update.
The current firmware is C0000X8 where X denotes TPS Threshold for logging
A is 75%
B is 50%
C is 25%
These are arbitrary and easily changed if needed.


The update files are 126 kb in size and are named UPDATE_C0000X8.EBL


Follow these steps to update the logger-
1)
drag and drop or copy the desired .EBL file onto the logger. It will appear as another file on the “disk” when copied over.
2)
Cycle power to perform the update. Upon re-powering, the LED will go Red while the logger checks the file integrity then Blue while the actual update occurs. If there is a file integrity error (does not pass RipeMD160 checksum) you will get a two blinking Blue error code. The Blue light extinguishes when the update is done. This occurs within a few seconds.


There are two ways to cycle power and they must be handled differently.
The first is to simply disconnect and reconnect the USB plug. The update will occur and the FAT file system will be erased BUT the data file storage section will NOT be erased. If you use USB power cycling to do the update, you will need to erase the log data file so that the logger will erase the flash storage space when it is plugged into the car (like you would normally do while using the logger).


The second way to cycle power is to disconnect the USB after copying the .EBL update file onto the logger and plugging it into the car. When powered, you will see the Red then Blue update LED sequence. When the Blue light goes out, the logger will perform its normal boot sequence as if you had erased the log data file. It will do the RGB post LED sequence followed by steady Red while it erases the flash data storage area. This is how I normally do it as the firmware update part is fast enough that there is no imposition.


Dropbox - UPDATE_C0000A8.EBL
Dropbox - UPDATE_C0000B8.EBL
Dropbox - UPDATE_C0000C8.EBL

 

[Bill D]

...Sure would love to lure some X90D, DL and P90D, DL folks to this thread. I've got a few more loggers and would love to see some data for X to S comparison.

Bill - Consider me lured. I've been busy and just now catching up on this thread. I'd love to borrow a LolaLogger and contribute data from my SP90DL.

 

[AWDtsla]

I'd be happy to log, but no quarter miles until April.

 

[MikeBur]

Ok then. More coming, though first results are in for pre/post Ludicrous upgrade.

Ludicrous installed and fun fun fun. I tried to get everything as similar as possible - temp, road conditions, Bat Temp, same piece of (private) road. Graph images, axis, scales the same. Check, check and double-check. SoC is "close", so potential caveat there.

First: 0-60+

Before, aka Insane:

After, aka Ludicrous:

Some numbers (note: best numbers with Throttle position at 100% - seems fairest comparison)

NO rollout	Insane	Ludicrous
0-60	3.31	3.07
0-86.3	6.38	5.81

 

[lolachampcar]

It's overlay time

 

[MikeBur]

It's overlay time

This work? (Dotted == Insane, Solid (translucent) == Ludicrous)

- - - Updated - - -

And zoomed in to what would appear to be the best comparison- moving when TPS==100

- - - Updated - - -

Note: I also moved from BatV and BatI to just using Pwr(kW), as well as expanded mp/h and switched axis for data... to simplify (at least that's what I told myself ;-))

 

[apacheguy]

I'm no performance expert, but once you know the conversion efficiency of kW from the pack to torque at the wheels, can't all these data be derived simply from REST readings? As for the 0-60 s, I'm highly skeptical of getting quality data by relying on the cars speedometer (even if it's at 100 Hz or whatever) since you really need high precision GPS (VBOX).

 

[MikeBur]

I'm no performance expert, but once you know the conversion efficiency of kW from the pack to torque at the wheels, can't all these data be derived simply from REST readings? As for the 0-60 s, I'm highly skeptical of getting quality data by relying on the cars speedometer (even if it's at 100 Hz or whatever) since you really need high precision GPS (VBOX).

I'm not sure you can do this directly, as wouldn't the efficiency vary based on number of other factors, incl. drivetrain? - eg. if you look at Bill's RPM data this is not consistently the same ratio of motor RPM to (albeit car-reported) speed through the dataset. I rely on others for interpretation and validity of data though, so caveat emptor here too

Assuming we want to make this a model for comparison between cars, then we would have to calibrate with reality as measured by high resolution GPS by applying any calibration offset, yep. At the moment, everything is relative - ie. I'm comparing my car with my car - I don't mind what absolute 0-n time I'm getting. Though saying that, Bill did get very similar 0-60 times, so I'd expect "similar" benefit for his upgrade (caveat battery health, tires, wheel size, etc... )

If anyone has got a vbox locally? I don't mind doing a high-res GPS to car-reported speed/distance if that helps?

Cheers, Mike

 

[Andyw2100]

This work? (Dotted == Insane, Solid (translucent) == Ludicrous)

So it looks like the max power output was increased by about 45 kW, give or take, yes?

 

[MikeBur]

So it looks like the max power output was increased by about 45 kW, give or take, yes?

Yes. If you notice on original chart this is ~1300 to 1500amps

 

[Andyw2100]

Yes. If you notice on original chart this is ~1300 to 1500amps

Thanks for pointing that out. I had failed to notice it before.

So...understanding that I have only the most basic understanding of this stuff...

I recall from the Ludicrous announcement Elon Musk talking about the then current fuse limiting the power to 1300 amps and how the new Ludicrous fuse would increase that limit to 1500.

Looking at those graphs it is clear that both those limits were exceeded by at least some amount in Insane mode and Ludicrous mode respectively. What does that mean?

Does it mean that when Musk spoke of the 1300 amp limit fuse in Insane mode it was really more of a 1350 or 1375 or even 1400 limit, and that in general Tesla would shoot to limit the power draw to 1300 amps, but the fuse wouldn't really blow (or whatever these fuses do) until something closer to 1400?

I also recall there being talk of how the Ludicrous fuse, being battery powered, could be more accurate.

So does that mean that, for example, if the Insane Mode fuse really blows at 1400, Tesla may be attempting to limit to 1300, to maintain a 100 amp safety margin, because the fuse could really blow anywhere between 1375 and 1425, but with the 1500 amp electronically controlled fuse (or whatever it is) in Ludicrous Tesla may have that fuse set at, say, 1550, and be shooting for 1500, using a smaller safety margin because the fuse is more accurate?

So in essence the new fuse allowed Tesla to kick up the amps by 200, but part of that was due to needing to maintain less of a margin for error?

Is any of the above close?

 

[apacheguy]

So P85DL gets 435 kW at 70% SOC. P90DL gets circa 460 kW at 90+%. I wouldn't be surprised if the P85DL is equivalent to the P90DL performance wise.

 

[MarcG]

I recall from the Ludicrous announcement Elon Musk talking about the then current fuse limiting the power to 1300 amps and how the new Ludicrous fuse would increase that limit to 1500.

Looking at those graphs it is clear that both those limits were exceeded by at least some amount in Insane mode and Ludicrous mode respectively.

I don't think the limits are exceeded by much at all. Looking carefully at Mike's graphs, at least to me it's spot on 1300A for Insane and 1500A for Ludicrous, give or take a few amps.

Insane:

Ludicrous:

 

[MikeBur]

So P85DL gets 435 kW at 70% SOC. P90DL gets circa 460 kW at 90+%. I wouldn't be surprised if the P85DL is equivalent to the P90DL performance wise.

Can get 445kW @78% SoC @46.81C on p85dL, yes. I wouldn't be surprised if p90DL is capable of more and would explain the MotorTrend numbers.

- - - Updated - - -

I don't think the limits are exceeded by much at all. Looking carefully at Mike's graphs, at least to me it's spot on 1300A for Insane and 1500A for Ludicrous, give or take a few amps.

Insane:

View attachment 112485


Ludicrous:

View attachment 112484

Peak is 1320 Amps and 1544 respectively, though also note that the sustained current is less (look at oscillation on boundary) and declines abruptly on insane packs. The decrease is significantly apparent at ~6 seconds of 1300A draw. I pulled some data in this post in an attempt to understand this more and Kenny was reinforcing the likely need for this based on the Bussmann fuse spec in the insane battery packs.

@Andy, I believe it was discussed in the HP thread about the reasoning for new fuse requirement. Suffice it to say that it is believed that although TM likely thought that 1500A was possible using existing fuses for 0-60 runs when announcing the insane software perf increase, this level of draw for more than a second or two would probably blow these fuses and was not reliable. The new pyro fuses supposedly are much more predictable at the draw limits. I say Supposedly as I don't believe we know the exact fuse type and specs.

- - - Updated - - -

Question: Anyone know how to estimate from 0-60 without 1ft rollout to what it would be with? I remember it being mentioned in a thread though search terms I'm using are hitting too many possibilities

thanks, Mike

 

[P85DEE]

You'll see rollout estimated as a 0.3 second difference according to at least one source.

So subtract approximately 0.3 for rollout.

How We Test Cars and Trucks

 

[lolachampcar]

Mike,
I love the overlays..... One graph says everything I need to know about the L upgrade. On average with an XYZ mile battery, these are typical results. Adding Bill's P90DL data to the overlay list will answer the 90 v. 85 question just as quickly.

As for roll out, I did an integration of wheel speed earlier in this thread to estimate when my MS had covered one foot of ground. The time to do so is in that post. I'll see if I can find it (along with the supporting logic for generating the number).
update
Link to roll out comment #2563
Looks like I found my car used 0.21 seconds to traverse the first one foot.

apache,
I look at things as having several ways to collect data. The wireless/hardwareless approach has its appeal. I was a little more interested in high resolution data as close to the source as possible which is why I looked into the chassis CAN logging.

We are on MS number 5 and I've been surprised to see that all of them have been either spot on or dithering on GPS speed at 60 mph. My experience with OEM speedometers in the past has had most all of them being over optimistic in their reporting (thus presumably over optimistic in the car's efficiency which is what I was investigating when I checked my spedos). When compared to the VBox, CAN wheel speed has the benefit of an order of magnitude higher report rate and will be more repeatable even if there is an absolute inaccuracy. The repeatability allows for very accurate back to back tests even if there is an accuracy issue that would prevent perfect 0-60 (and like) measurements.

Basically, the logger is but one more option for those looking to understand their cars.

On a different note, I've heard the argument on the HP thread about the larger spec motors being able to generate more low end torque even though overall power is battery limited. By the looks of all the graphs, Tesla is torque limiting both motors (those flat initial lines) prior to battery limiting kicking in (the torque knees). There may be bigger motors in MS but it would not appear as though Tesla is making using of the extra torque available at lower rpm from over spec'd motors. This can also explain why times are not improved on sticky prepped drag strips. The launch is not traction control limited but simply torque capped by the electronics.

 

[Sad]

Tesla is torque limiting both motors (those flat initial lines)

How much more torque could those motors produce at higher RPMs? Sorry for the possibly dumb question, but I thought the motors had a physical limit on torque (as well as power), as displayed in the Tesla specs, isn't it what the flat torque curve at lower RPMs is about? Or are you referring to the fact that the sum of the two torque figures is ~5% below the 713 lb-ft official figure?

 

[lolachampcar]

Actually, my comment was directed at the low RPM portion in the first second where the plateau shows Tesla is holding back low RPM torque. The battery can source more current but Tesla chooses not to provide the motors with that current. Once the battery reaches power limit, the torque for front and rear decline as rpm increases basically consuming all that the battery can supply.

 

[Sad]

Actually, my comment was directed at the low RPM portion in the first second where the plateau shows Tesla is holding back low RPM torque. The battery can source more current but Tesla chooses not to provide the motors with that current. Once the battery reaches power limit, the torque for front and rear decline as rpm increases basically consuming all that the battery can supply.

That's what I was referring to as well, but isn't it the inherent torque limit of the motors that is producing this plateau, and not necessarily the traction which would be the logical reason for Tesla to limit the torque.

 

[arg]

So in essence the new fuse allowed Tesla to kick up the amps by 200, but part of that was due to needing to maintain less of a margin for error?

My understanding is that this is all about making sure the fuse actually blows when you want it to - if all you cared about was drawing more power, you could replace the fuse by a copper bar, but it would no longer offer any protection against actual faults.

Looking at these graphs, 1300A produces a voltage drop of ~75V and 1500A produces a voltage drop of 85V, which imply that a hard short-circuit at the battery terminals would produce a maximum current of about 6500A at that state of charge. Close to empty, the maximum fault current would be lower - perhaps 4500A. A real short circuit, caused perhaps by crushing cables in an accident, would have more resistance than that and hence a lower current still.

With a conventional fuse, you would struggle to find one that can reliably deliver 1500A for several seconds (and keep going at significant current after that) without getting close enough to melting that it wears out, while at the same time being guaranteed to blow in a second or so at 4000A that it needs to do for safety.