I decided to start a new thread in my saga to have a reliable SV08 rather than continue the old one that was getting pretty long
I’ve had to replace the A/B belts. One belt was so short it could not be adjusted properly, the new belts solved that problem. I then had a failure of the 3 in one data cable from the motherboard to the toolhead. The power cable failed very close to the clamp that holds the cable to the toolhead.
Sovol has provided the parts under warranty for both these problems, but I was left to do the installation.
Now that I have the new cable installed the printer is running but is not reliable. Twice in two days the printer has failed with an error 8 "lost communication with the MCU extra-MCU. Turning the power off and restarting solves the problem for a while. I’ve checked all the connections at least twice, I have not done the troubleshooting for this error as stated in the SV08 WIKI, I will break out the Fluke if the problem persists.
I am beginning to truly believe that I was sold either a first day of the workweek or a last day of the workweek printer or the SV08 is just naturally unreliable. I’ve only had the damn thing since March!
I told you earlier that USB interference from noise on the DC cables is a known issue.
Leave the new cable in place but unplug both 24V connectors. Take your old cable and harvest the plugs for the 24V circuit.
Get some 18ga speaker wire. Start with a piece 24" too long, put one end in a low speed drill and twist the zip cord at 2 turns per inch. Wind it HARD at first as it will unwind some as you handle it. cut to length and splice on the connectors. Cut a bunch of 3/4" sections of semi rigid tubing (drinking straw will work if you have nothing better). Thread a small zip tie through one tube, around the existing (data) cable, back through the tube, around the new power wires an zip. Repeat every 3-4 inches until you’re happy with the effect.
Extra credit - use one of the (now) spare wires in the “data cable” from the ring terminal to the single wire ground pin at the main board.
I am going to build the power bypass you recommend but I don’t know why I should have to compensate for @Sovol3d crappy design. I have spent more time replacing poorly designed and installed components than I have in using the printer.
I have access to 22 awg twisted pair shielded wire, I may use it. It is aircraft quality wire and I can solder a drain onto the shield pretty easily.
This is going to make the connection between the motherboard and the tool head somewhat larger than even the original cable.
What kind of interference could the new cable power / ground line pick up while disconnected? My though is to just tie the connectors out of the way when I put the bypass in at both ends.
I did try a clip on ferrite ring core on the new cable about six inches from the motherboard, about where the sleeving starts. I managed to do the max flowrate Orca Slicer calibration and was about 15% through an actual print when error 8 occurred again.
One more thing, I will be doing all the tests recommended by @Sovol3d on the SV08 WIKI: How to troubleshoot "Lost communication with MCU ‘extra_mcu’ " | Sovol 3D Printer Wiki
You are 100% correct.
You should NOT have to do Sovol’s tests at that link
You should NOT have to modify toolhead power delivery.
Unfortunately the only alternative is box it up and return it. Use your refund to buy a better printer… Oh wait, every other similar printer is $200+ more expensive.
Begs the question though, will I ever have a printer reliable enough to complete a print. At the moment the answer is NO!
I don’t know how @Sovol3d works returns. I purchased the printer through Amazon and that return window is closed. I would do an exchange for a new production SV08 and just pray it’s more reliable than this one. Not everyone that has one of these has the kinds of problems I’ve had, at least I don’t think they have.
Hello,
The best thing to do is to call AMZ and tell them that the printer is not working.
You are under warranty. They will ask you to explain the issues and then to return the printer and buy a new one.
That’s what I did with a previous TENLOG printer.
I just completed the troubleshooting steps contained in the wiki.
The results are:
1-Overall hardware test (mainboard, adapter board, signal cable)
GND to UN expected .58M results = .6M
GND to UP expected 2-2.8K results = 2.749K
GND to 24V expected 1-30M results = 309K
2-Adapter board test
GND to UN expected Inf results = 20.9M
GND to UP expected 2-2.8K results = 2.76K
GND to 24V expected 1-30M results = 327K
3-Mainboard & Signal Cable Test
GND to UN expected .6M results = .6M
GND to UP expected .6M results = .6M
Sovol sent me a new 3 in one data cable and a toolhead adapter board set. Software/firmware was already installed on the adapter board. The data cable was similar to but slightly different than the original cable that had the tool head power lead break. This new one had a shield drain at the tool head end and the MCU end the data connector also had a small ground. I say similar because it seems to be smaller in diameter than the original cable. It includes all connectors and wiring that the original cable had.
I did all the calibrations as if the machine was new, even did a Benchy with some Polymaker PLA. I switched to American Filament PCTG and it has been printing now for over 5 hours with no issues.
I do have a question for the group. Does the firmware/software on the toolhead control the bed mesh? The reason I ask is now during a print when it does the mesh just before starting to print, that is shown in Orca slicer as the heightmap and the name of the mesh is adaptive followed by a long hexadecimal number. I don’t remember noticing that before it always showed the default mesh, at least I thought it did, I may be completely wrong.
The toolhead MCU does interface with the eddy sensor.
BUT the whole point of Klipper is to move all the computation to the “Host computer” and only use the MCU(s) to oversee time synchronization and compress/decompress raw sensor data and output states.
Quotes around “Host computer” because the SV08 mainboard has both the MCU and the Host computer on 1 circuit board.
Klipper interacts with the toolhead board using pin addresses. With no changes to printer.cfg (or an included .cfg) there should be no change in how anything works.