Issues with the latest beta (2.03 b3)? "bad command"

norbs12

I've been printing for almost two weeks straight without issue, however right after updating I've seen this on my last 2 of 3 prints. Not saying for sure that it's the firmware at fault since I have not downgraded yet just trying to figure out what is going on.

This is the error:
alt text

After downloading the GCODE back from the printer, I see that it does have the bad GCODE but the copy i still had on my computer doesn't have the same error.
alt text

"Diff" comparison on the same gcode on my computer and the one I downloaded from the printer.

< G1 X78.204 Y146.464 E0.240
---
> G1 X78.204 Y146.464 E0.0240
342073c342073
< 0G1 X83.701 Y145.109 E0.0055
---
> G1 X83.701 Y145.109 E0.0055

bot

Did you ever run a simulation to generate print time on this file? I believe the firmware will write to the file in this instance -- something which I feel should be forbidden. Print Gcode files IMO should be ready-only by the firmware.

I have no idea if this is the case, or if this is just an odd corruption that happened when transferring the file, but it's worth bringing up. I often deal with Gcode files in excess of 1 GB, so I do not want the firmware trying to write to that file -- it will likely result in crashes or corrupt files.

norbs12

@bot said in Any having issues with the latest beta (2.03 b3)? "bad command":

Did you ever run a simulation to generate print time on this file? I believe the firmware will write to the file in this instance -- something which I feel should be forbidden. Print Gcode files IMO should be ready-only by the firmware.

I have no idea if this is the case, or if this is just an odd corruption that happened when transferring the file, but it's worth bringing up. I often deal with Gcode files in excess of 1 GB, so I do not want the firmware trying to write to that file -- it will likely result in crashes or corrupt files.

Naw I don't usually run simulation before printing and I'm sure I didn't in this case, though a simulation would have likely caught the problem.

Luckily in this time the print didn't get ruined by the bad line, had another one where it threw everything way off.

norbs12

Ran a few additional tests. Uploaded a 100MB gcode file to the printer then renamed it and downloaded it. Diff showed quite a bit of differences...

383344c383344
< G1 X159.187 Y158.973 0.0064
---
> G1 X159.187 Y158.973 E0.0064
383378c383378
<  G1 X142.123 Y162.035 E0.0064
---
> G1 X142.123 Y162.035 E0.0064
654509c654509
< G1 662.429 Y157.011 E0.0041
---
> G1 X62.429 Y157.011 E0.0041
1052118c1052118
< G1 X27.83 Y248.237 E0.0060
---
> G1 X27.883 Y248.237 E0.0060
1052158c1052158
< G1 X28.993 Y237.054 E0 .0054
---
> G1 X28.993 Y237.054 E0.0054
1218337c1218337,1218338
< G1 X121.249 Y126.937 E0.0060G1 X121.480 Y127.361 E0.0066
---
> G1 X121.249 Y126.937 E0.0060
> G1 X121.480 Y127.361 E0.0066
1218372c1218373
< G1 X130.009 Y118.655 E0.01094
---
> G1 X130.009 Y118.655 E0.0109
1424671c1424672
< G1 X151.592 Y110.16  E0.0031
---
> G1 X151.592 Y110.164 E0.0031
1479890c1479891
< G1 X95.562 Y146.417F4025
---
> G1 X95.562 Y146.417 F4025
1479938c1479939
< G1 X86.659 Y129.117 E0.005Y1
---
> G1 X86.659 Y129.117 E0.0051
1811283c1811284
< G1 X223.60  Y172.224 E0.0024
---
> G1 X223.608 Y172.224 E0.0024
1908715c1908716
< G1 X196.599 Y177.845 E0.0083
---
> G1 X196.599 Y157.845 E0.0083
1909045c1909046
< G1 X180.891 Y10.748
---
> G1 X180.891 Y160.748
1909077c1909078
< G1 X1778.748 Y169.827 E0.0030
---
> G1 X177.748 Y169.827 E0.0030
2153494c2153495
< G1 X123.799 Y177.45 E0.0066
---
> G1 X123.799 Y177.455 E0.0066
2153496c2153497
< G1 X123.279 Y176.563 E0.075
---
> G1 X123.279 Y176.563 E0.0075
2153503c2153504
< G1 X1.722.086 Y173.193 E0.0017
---
> G1 X122.086 Y173.193 E0.0017
2250709c2250710
< ; tool H0.10 W0.083
---
> ; tool H0.100 W0.083
2250711c2250712
< G1 X183.407 Y170.883 E0.0005 F450
---
> G1 X183.407 Y170.883 E0.0005 F3450
2250730c2250731
< G1 X187.241 Y1Y171.521 E0.0039
---
> G1 X187.241 Y171.521 E0.0039
2328424c2328425
< G1 X230.623 Y243.489 E0.004
---
> G1 X230.623 Y243.489 E0.0046
2328466c2328467
< G1 X233.370 Y244.366 E0.60727
---
> G1 X233.370 Y244.366 E0.0727
2505308c2505309
< G1 X235129 Y238.909 E0.0022
---
> G1 X235.129 Y238.909 E0.0022
2505361c2505362
< G1 X231.4252 Y239.061
---
> G1 X231.422 Y239.061
2981519c2981520
< G1 X162.74  Y153.592 E0.0030
---
> G1 X162.741 Y153.592 E0.0030
2981521c2981522
< G1 X162.338 Y153.98 E0.0029
---
> G1 X162.338 Y153.498 E0.0029
2981561c2981562
< G1 X154.133G Y152.263 E0.0029
---
> G1 X154.133 Y152.263 E0.0029

Ended up downgrading to 2.02 and after two tests there were no differences between the uploaded file and downloaded file. Considering going back to 2.03 B3 but I've got a few things I need to print.

Don't know if I had a bad firmware flash or something else but it's working correctly for now.

dc42

Which Duet do you have?

If you download the file back to your PC again, does the diff show exactly the same errors?

norbs12

@dc42 said in Issues with the latest beta (2.03 b3)? "bad command":

Which Duet do you have?

If you download the file back to your PC again, does the diff show exactly the same errors?

I have a duet 2 wifi (converted to ethernet) (v1.03)

It seems to be different each time. Also I just tried it again while the printer was going for the past ~8 hours and it seems like I'm having the same issue on 2.02 firmware. So I guess that eliminates it being a 2.03 B3 issue, and it's just related to either how busy (or hot) the board is.

dc42

@norbs12 said in Issues with the latest beta (2.03 b3)? "bad command":

It seems to be different each time.

Is that different that every time you download the file from the SD card to the PC, meaning that at least some of the errors are occurring during the download process? Your initial report suggests that an error occurred during the upload process.

norbs12

@dc42 said in Issues with the latest beta (2.03 b3)? "bad command":

@norbs12 said in Issues with the latest beta (2.03 b3)? "bad command":

It seems to be different each time.

Is that different that every time you download the file from the SD card to the PC, meaning that at least some of the errors are occurring during the download process? Your initial report suggests that an error occurred during the upload process.

Was actually just about to clarify because I realized I misunderstood your question. I have not yet downloaded twice. So far I've used the same original gcode to upload multiple times, and then download after. Each time the download had different errors. I did not try downloading it twice but I can tomorrow. However since I saw an error during printing that was never in the original GCODE, I don't think it has anything to do with download transmit errors.

dc42

My guess is that the SPI communication between the Ethernet module and the Duet main processor is not working reliable. Is your Ethernet conversion a standard one, using our Ethernet daughterboard plugged into low-profile socket strips? The SPI runs at 30MHz clock rate, so signal integrity is critical.

norbs12

@dc42 Yeah I bought the ethernet module from filastruder.

It seems to work fine as long as I'm not printing when I upload the gcode. Is there anything simple that I can check?

dragonn

My Duet is original with WiFi and on 2.03 I once experienced that uploaded file was cut, a few thousand lines were missing and the print didn't finish.
I never experienced it on previous firmwares, but this was so rare. It only happen once

FBG

2.03 (no beta) happen me the same.
Duet Ethernet v1.04b and say: 4G Bad Command

Some idea?

norbs12

@fbg My issue turned out to be some sort of hardware problem. Thing is I've had odd issues since I bought the board but always though it was attributed to Wifi. When I moved to Ethernet the issues persisted and I'm already way out of my warranty period. Either way I ended up buying a Fystec Duet from Ali because in no way am I going to spend another $200+ for the real thing.

Either way my fystec has been working perfectly, and I'm saving the authentic Duet (with the hardware issue) for a klipper-based build since klipper doesn't rely on a network connection to the duet board.