Maestro issues (Firmware not updating and no ethernet)
-
When you boot up with the SD card removed, does the hot end still power on immediately?
-
@dc42 said in Maestro issues (Firmware not updating and no ethernet):
When you boot up with the SD card removed, does the hot end still power on immediately?
It no longer seems to be experiencing this fault with or without the SD card being inserted.
I powered the board from my printer's PSU and connected the USB to my laptop (running off battery to avoid ground loops) and the heater and thermocouple board seem to be working properly.
(Aside: When I don't have an SD card inserted, it still can read the proper temperature from the thermocouple board. This surprised me as I had assumed that the config.g file was not loaded into flash. I would have expected garbage values without the config.g file present on the SD card to tell the firmware to utilize the optional board)
-
Still no ethernet? I figured I'd try to build RRF3.2 for the Maestro with the changes indicated in this post: https://forum.duet3d.com/post/200505
to see if that would skip past the auto negotiation and bring the link up, but ran into some errors building it - will poke it some when i find the time
meanwhile if you have a managed switch, you could try setting a fixed link speed on the port the maestro is connected to.
-
@bearer said in Maestro issues (Firmware not updating and no ethernet):
meanwhile if you have a managed switch, you could try setting a fixed link speed on the port the maestro is connected to.
Unfortunately none of my switches are managed. I do have a spare wireless router that is compatible with DD-WRT which I could test with.
I guess one other thing I could test is to go back to 3.1.1 and see if it's able to get a lease now that I've reflashed the firmware via BOSSA, as it was able to previously.
-
I remembered that I had an old router that only had 10/100 ports on it in the basement. After dusting it off and setting it up, I found that there was no difference. Still no lights, and no lease. I even tried again with a static IP to no avail.
-
The maestro is only 10/100 as well, but a very small handful of users have experienced incompatibility with some network gear and auto negotiation.
Perhaps the easiest test is to try a direct cable connection to a computer, and change the drivers on the computer to have a fixed 100M full duplex connection instead of auto negotiation.
-
I'll give it a try, but I don't understand why it no longer works with the exact same hardware.
-
I connected my laptop and the board to each other through ethernet, set my laptop's IP to 192.168.0.2, subnet 255.255.255.0, gateway to 192.168.0.1, speed to 100baseT, and the printer's IP to 192.168.0.3, subnet 255.255.255.0, gateway to 192.168.0.1. No change. Also tried changing it between half and full duplex.
I also just downgraded the firmware to 3.1.1 through BOSSA as that was the original firmware version that it worked on, but the ethernet doesn't work on it either.
-
It seems like whatever failure/short happened to cause the hotend to power up uncontrolled also damaged the ethernet port. Can you disconnect everything and get some clear, well-lit photos of the board? Perhaps we can spot a damaged component.
-
Here you go: https://www.dropbox.com/sh/d23sm2xxk35xh3v/AABKB4PaqMS7xzck2IXK-XPKa?dl=0
(The little black circle on the bottom left of the Wiznet chip seemed to be a little bit of finger oil, which came off with some isopropyl. IMG_5878.jpg and IMG_5880.jpg were taken after cleaning)
I don't see any obvious evidence of a damaged component.
-
Nothing jumps out at me either, but failures aren't always spectacular.
When and where did you purchase the Maestro?
-
I purchased it from spool3d.ca on October 15th, 2019. It's taken quite a while to get everything set up - it has only done one print so far.
-
That is unfortunately well out of warranty.
Perhaps @bearer may have a keener eye and can spot something in your photos or has an idea on potential repair.
-
I see. I'm kind of bummed out here to be honest, as this is the second piece of Duet hardware that has failed on me. The first being the IR sensor, which had a bad solder joint on the IR receiver, and now this. This certainly isn't something I would have expected from a premium product.
-
If you wish you can contact roland@duet3d.com and inquire about doing an exchange for a refurbished board at reduced cost, subject to availability.
But hopefully bearer or DC42 has a better idea of what could have gone wrong and whether it's repairable. From the troubleshooting we've done it's safe to say that the network interface is no longer working.
-
(Pardon the delay)
I missed that it worked prior to the upgrade; to troubleshoot I'd probably monitor the SPI bus between the W5500 and CPU and replace whichever failed to initialize as can't see any obvious failures.
-
I assume I would need a logic analyzer or oscilloscope with decoding functionality in order to do that?
Are the pins exposed on some easily accessible test point? I don't see any traditional SPI pins on the back.
Would it be possible for back-emf to destroy one of the stepper drivers is such a way that it would be messing up the rest of the SPI bus? I had my cantilevered bed fall a short distance before I noticed the board went bad.
-
@j3d said in Maestro issues (Firmware not updating and no ethernet):
Are the pins exposed on some easily accessible test point?
no, but the kicad sources are available https://github.com/Duet3D/Duet-2-Hardware/tree/master/Duet2/Duet2Maestro_v1.0 your options are the vias on the bottom or the pins on the top. and yes a logic analyzer would help.
if lacking then starting with replacing the w5500 would be my bet, or you could try https://forum.duet3d.com/post/127089 - or as Phaedrux suggest maybe a refurbished exchange.
-
Do you happen to know what the SPI frequency is?
-
not sure, but i think <=12mhz (could be 1mhz but not too familiar with the code, which is also available on the duet3d github)
(also stepper drivers on the maestro use uart, not spi)