Announcement
Collapse
No announcement yet.
MSS54 DS2 Tool: FREE - DME tune, program, flash counter, codes, adaptions, etc
Collapse
X
-
Hi zivagolee,Originally posted by zivagolee View Posthey karter16, thanks for the awesome tool! I love that it gives so much more information than other tools without having to click a dozen buttons or just suppressing all the log information.
I recently uploaded a tune file, however, for some reason I encountered 'unexpected k-line echo' and then it broke the connection. I then did a bunch of troubleshooting trying random things, but I eventually got it connected again and reflashed the tune file but was successful (was sweating a bit
)
I had Claude go thru the logs and here is what it said:
===
21:56, the original brick, and the actual culprit. The first session shows a healthy DME: identify clean, flash counter marker 0xFFFF available, error memory readable. Then Write Tune starts, the tool boosts the line from 9600 to 125000 baud, and 4.6 seconds into writing master data it aborts with "Unexpected K-line echo." Look at the expected-versus-got bytes: identical for dozens of bytes, then a few corrupted mid-stream (FF CE 04 00 00 01 became FF E6 30 00 01). K-line is a single wire, so the tool hears its own transmission back and verifies it; those bytes got mangled on the wire. Immediately after, keepalives returned garbage (BE 7A 01 6B). Your physical link corrupts bytes at 125000 baud. That's the root cause of the entire evening: not the laptop battery, not the tool, not the DME. Signal integrity at fast baud, mid-write, leaving a data-programming session open. I'll own the correction: I nagged about AC power all night, and that's still correct practice, but the logs name a different killer.
Why connecting seemed random afterward. Every "DME did not respond" had a mechanical cause: after an abnormal session end, the DME stays listening at 125000 while the tool reconnects at 9600. The 22:49 log shows it explicitly: 9600 identify failed, then the 125000 recovery probe succeeded and the tool stayed there. Ignition off/on resets the DME to 9600. So your connects weren't flaky; they were a baud mismatch that KL15 cycling or the recovery probe resolved each time.
Forcing 9600 only:
Your own log's settings line is the map here. It shows two fast-mode keys the tool loads: fastReadEntryDataTune=True and fastReadEntryFullBinary=True. Those are the toggles for boosting to 125000 on reads, and they should appear in the tool's Settings page as fast-read options. Turn both off and your reads and backups will stay at 9600 (slower, but 9600 was flawless all night in your logs).
The honest limit: for writes, both logged write flows show "PROGRAM baud: requesting 125000" with no corresponding setting in the loaded-settings line, which suggests the write-path boost is hardcoded rather than configurable. Two moves there:- Look in the tool's Settings page anyway for anything named fast mode, program baud, or write speed; a UI option may exist that simply wasn't echoed in that log line.
- If nothing's there, check the config directly: close the tool, open C:\Users\z\AppData\Local\MSS54 DS2 Tool\appsettings.json in Notepad, and look for any key resembling fast, baud, or 125000 beyond the two read keys. Change only something that plainly matches; don't invent keys.
- If writes can't be slowed, that's a feature request for the tool's author, and it's a good one given your logs: "echo corruption at 125000 bricked my DME; please add a 9600-only write option." Attach the 21:56 log. A 384KB write at 9600 would take on the order of ten-plus minutes, which is a perfectly good trade for a recovery scenario.
- Either way, treat the latency and USB power fixes as the real solution, because they target the actual failure. Your logs show 9600 was perfect and 125000 corrupted bytes; with latency at 1, suspend off, and a direct port, 125000 very likely becomes clean too, and if you want certainty, a better K+DCAN cable is the last piece.
I am curious on why it seems to be negotiating at 125000 baud (or maybe it should?). I am running this on an older lenovo laptop and it does have usb3 so I guess it could transfer it much faster but I do have an older OBD cable. Per Claude, I set it the FTDI latency timer to 1 and, I guess, it's supposed to help with this -- I haven't tried tuning again after this.
Question:
1. Do I have a bad cable maybe? Should I get another dongle? I have had this for quite some time.
2. Maybe we need a setting so it stays at 9600?
3. Am I totally misunderstanding the error message?
I didn't see source code so I couldn't get Claude to check it out but posting this here for feedback. Again, thanks for this!
When you connected was it via COM port (e.g. COM1, etc.) or with the FTDI D2XX device? If you'd be able to send me (via PM if you don't want to share here) your log file that will make it a lot easier for me - thanks :-)
- Likes 1
Leave a comment:
-
hey karter16, thanks for the awesome tool! I love that it gives so much more information than other tools without having to click a dozen buttons or just suppressing all the log information.
I recently uploaded a tune file, however, for some reason I encountered 'unexpected k-line echo' and then it broke the connection. I then did a bunch of troubleshooting trying random things, but I eventually got it connected again and reflashed the tune file but was successful (was sweating a bit
)
I had Claude go thru the logs and here is what it said:
===
21:56, the original brick, and the actual culprit. The first session shows a healthy DME: identify clean, flash counter marker 0xFFFF available, error memory readable. Then Write Tune starts, the tool boosts the line from 9600 to 125000 baud, and 4.6 seconds into writing master data it aborts with "Unexpected K-line echo." Look at the expected-versus-got bytes: identical for dozens of bytes, then a few corrupted mid-stream (FF CE 04 00 00 01 became FF E6 30 00 01). K-line is a single wire, so the tool hears its own transmission back and verifies it; those bytes got mangled on the wire. Immediately after, keepalives returned garbage (BE 7A 01 6B). Your physical link corrupts bytes at 125000 baud. That's the root cause of the entire evening: not the laptop battery, not the tool, not the DME. Signal integrity at fast baud, mid-write, leaving a data-programming session open. I'll own the correction: I nagged about AC power all night, and that's still correct practice, but the logs name a different killer.
Why connecting seemed random afterward. Every "DME did not respond" had a mechanical cause: after an abnormal session end, the DME stays listening at 125000 while the tool reconnects at 9600. The 22:49 log shows it explicitly: 9600 identify failed, then the 125000 recovery probe succeeded and the tool stayed there. Ignition off/on resets the DME to 9600. So your connects weren't flaky; they were a baud mismatch that KL15 cycling or the recovery probe resolved each time.
Forcing 9600 only:
Your own log's settings line is the map here. It shows two fast-mode keys the tool loads: fastReadEntryDataTune=True and fastReadEntryFullBinary=True. Those are the toggles for boosting to 125000 on reads, and they should appear in the tool's Settings page as fast-read options. Turn both off and your reads and backups will stay at 9600 (slower, but 9600 was flawless all night in your logs).
The honest limit: for writes, both logged write flows show "PROGRAM baud: requesting 125000" with no corresponding setting in the loaded-settings line, which suggests the write-path boost is hardcoded rather than configurable. Two moves there:- Look in the tool's Settings page anyway for anything named fast mode, program baud, or write speed; a UI option may exist that simply wasn't echoed in that log line.
- If nothing's there, check the config directly: close the tool, open C:\Users\z\AppData\Local\MSS54 DS2 Tool\appsettings.json in Notepad, and look for any key resembling fast, baud, or 125000 beyond the two read keys. Change only something that plainly matches; don't invent keys.
- If writes can't be slowed, that's a feature request for the tool's author, and it's a good one given your logs: "echo corruption at 125000 bricked my DME; please add a 9600-only write option." Attach the 21:56 log. A 384KB write at 9600 would take on the order of ten-plus minutes, which is a perfectly good trade for a recovery scenario.
- Either way, treat the latency and USB power fixes as the real solution, because they target the actual failure. Your logs show 9600 was perfect and 125000 corrupted bytes; with latency at 1, suspend off, and a direct port, 125000 very likely becomes clean too, and if you want certainty, a better K+DCAN cable is the last piece.
I am curious on why it seems to be negotiating at 125000 baud (or maybe it should?). I am running this on an older lenovo laptop and it does have usb3 so I guess it could transfer it much faster but I do have an older OBD cable. Per Claude, I set it the FTDI latency timer to 1 and, I guess, it's supposed to help with this -- I haven't tried tuning again after this.
Question:
1. Do I have a bad cable maybe? Should I get another dongle? I have had this for quite some time.
2. Maybe we need a setting so it stays at 9600?
3. Am I totally misunderstanding the error message?
I didn't see source code so I couldn't get Claude to check it out but posting this here for feedback. Again, thanks for this!Last edited by zivagolee; 10-23-2026, 02:08 PM.
Leave a comment:
-
The mismatch is to do with the fact that I’ve implemented writing the program and tune as separate actions. At the end of each action the software “finalizes” the write which performs a series of validations in the DME. At the point the program is the new version and the tune is the old version the backup ZIF entries don’t match. Interesting that it cleared after a drive cycle rather than after writing the tune (and the finalization that goes with it) I’ll have a look and see if I can see why that might be.Originally posted by Pklauser View PostWhat's interesting is that after a test drive, the ZIF warning went away 🤔
Sent from my iPhone using Tapatalk
- Likes 2
Leave a comment:
-
What's interesting is that after a test drive, the ZIF warning went away 🤔
- Likes 1
Leave a comment:
-
Yep correct - essentially nothing to worry about, have it on my list to make an update to the app to handle this more gracefully :-)Originally posted by oceansize View Post
There is a ZIF entry that tracks hardware and software variants on your car. There is a back up zif entry that is what it says. I believe the zif copies mismatch is referencing a mismatch between the main entry present in UIF versus the backup entry.
However, its inconsequential as to how the car performs. I noticed that in this post karter16 referenced a "misunderstanding" concerning the backup zifs persistence. I'm sure he can shed some more light.
- Likes 3
Leave a comment:
-
Can you flick me the log file from the session in question and I'll have a lookOriginally posted by Cubieman View PostWhat might be reason(s) that the program will not read a partial of the DME yet is able to read error codes etc.?
I should have taken a photo, but I *think* after it didn't read the partial it said something similar to "5 attempts at reading, timed out".
Leave a comment:
-
What might be reason(s) that the program will not read a partial of the DME yet is able to read error codes etc.?
I should have taken a photo, but I *think* after it didn't read the partial it said something similar to "5 attempts at reading, timed out".
Leave a comment:
-
There is a ZIF entry that tracks hardware and software variants on your car. There is a back up zif entry that is what it says. I believe the zif copies mismatch is referencing a mismatch between the main entry present in UIF versus the backup entry.Originally posted by Pklauser View PostFirst things first, amazing tool, very cool. Encountered something today that I'm not sure if is a bug, or me doing something wrong.
In an attempt to flash the Euro tune like mentioned here, I went and grabbed the 211322001501JD34 program, and the 211322001501J424 tune. My understanding being that the crucial "1501" program being the same between the two was what mattered.
First I did a program and tune read through the DS2 tool to have a backup. When prompted to do a backup as part of the flash I thought "what the hell" and did that backup as well.
Program and tune flashed just fine, but then I had a "Warnings: ZIF Copies Mismatch" on the DME Info tab.
Thinking this was related to the different JD34 vs J424 configurations, I went ahead and flashed the original program and tune that I had previously pulled from the DME Read tab, however the "ZIF Copies Mismatch" message remained.
It wasn't until I went ahead and used the "Write this Program" and "Write this Tune" buttons from the backup made before the flash, that I was able to get rid of the "ZIF Copies Mismatch" warning.
Was I doing something wrong in trying to restore the program and tune that I had previously read? What does a "ZIF Copies Mismatch" actually mean?
However, its inconsequential as to how the car performs. I noticed that in this post karter16 referenced a "misunderstanding" concerning the backup zifs persistence. I'm sure he can shed some more light.
- Likes 2
Leave a comment:
-
First things first, amazing tool, very cool. Encountered something today that I'm not sure if is a bug, or me doing something wrong.
In an attempt to flash the Euro tune like mentioned here, I went and grabbed the 211322001501JD34 program, and the 211322001501J424 tune. My understanding being that the crucial "1501" program being the same between the two was what mattered.
First I did a program and tune read through the DS2 tool to have a backup. When prompted to do a backup as part of the flash I thought "what the hell" and did that backup as well.
Program and tune flashed just fine, but then I had a "Warnings: ZIF Copies Mismatch" on the DME Info tab.
Thinking this was related to the different JD34 vs J424 configurations, I went ahead and flashed the original program and tune that I had previously pulled from the DME Read tab, however the "ZIF Copies Mismatch" message remained.
It wasn't until I went ahead and used the "Write this Program" and "Write this Tune" buttons from the backup made before the flash, that I was able to get rid of the "ZIF Copies Mismatch" warning.
Was I doing something wrong in trying to restore the program and tune that I had previously read? What does a "ZIF Copies Mismatch" actually mean?
Leave a comment:
-
I don't login here often, but I had to today to tell you, you are an f'in legend.
Thanks for creating a great program and sharing it with the community!
Edit:🤣 just noticed this is my first ever post even though I am a member since 2022....
- Likes 2
Leave a comment:
-
Thanks mate and We all pray that God Blesses your family with Health and Happiness.Originally posted by karter16 View PostQuick update - the last couple of months for me have been family with back to back sicknesses, etc. plus starting a new contract at work so I haven't had enormous amounts of spare time. I did publish v1.2.2 today which is a very small change. Both George Hill and ac427 have shared logs with me that contain occasional failures directly after programming mode entry and baud rate elevation. It seems that the root cause is possibly not allowing enough time for the channel to settle after the rate change, so I've extended that waiting period to a more conservative value in the hopes of solving this. The failure occurs at a non-critical point in the various workflows, so no risk to DME integrity, just annoying to have to restart the DME and app and try again.
- Likes 1
Leave a comment:
-
Essentially non-zero values are compensations to the injection amount (as a result of the injectors not all behaving equally). The caution I'd give is I *think* that if you happened to have a vacuum leak that affected one or more TBs it could also result in compensation, even though the injector isn't the root cause. So I think this can be useful, but is not an entirely causative diagnostic test.Originally posted by E46 M3 View Post
Is there a guide on how to interpret the data? Been thinking of getting my injectors cleaned for a while now.
- Likes 2
Leave a comment:
-
Quick update - the last couple of months for me have been family with back to back sicknesses, etc. plus starting a new contract at work so I haven't had enormous amounts of spare time. I did publish v1.2.2 today which is a very small change. Both George Hill and ac427 have shared logs with me that contain occasional failures directly after programming mode entry and baud rate elevation. It seems that the root cause is possibly not allowing enough time for the channel to settle after the rate change, so I've extended that waiting period to a more conservative value in the hopes of solving this. The failure occurs at a non-critical point in the various workflows, so no risk to DME integrity, just annoying to have to restart the DME and app and try again.
- Likes 7
Leave a comment:
-
I had forgotten about the rough running data.Originally posted by karter16 View Post
It's already there :-) Data can be interpreted in the same was as you would in DIS/INPA/ISTA
There is also an injector activation section in DIS where you can activate individual injectors. I'm not sure how useful this though. Since all you could do would be to listen for the clicking of the injectors solenoid.
Would there be any other use for the injector activation part of DIS?
- Likes 1
Leave a comment:

Leave a comment: