Diagnosing a failing ESP-12



  • Is there a good way to do this without a frequency analyzer/o-scope?

    I got one of my printers back up and (mechanically) running but something seriously funky is happening with the wifi.

    First of all, yes, this is out of warranty. My FIlastruder ship date for this particular board was May 21, 2018.

    Hoping to confirm it's just the ESP taking a dump since my fat fingers have the best chance of desoldering that compared to the SAM8 chip.

    Board boots, all LEDs light up, and the wifi connects. M122 from Pronterface doesn't show any errors, and the networking section reports an IP, etc.

    Running RRF3.1.1 and WIfi server 1.23 the same as my 4 other Duet Wifi boards. I can see the IP assigned in Windows Server DHCP, I can run an nslookup on it, sometimes it returns pings, and I even got into the FTP once. However; DWC never loads, it's dropping packets like mad, I've been through three different SD cards and run an M997 S1 about 20 times today. I'd really prefer not to order another board with the mini 3 seeming so close.

    Do the details below suggest that I might just be able to swap the ESP, reflash the wifi server, and get another couple years from the board? The latency on the pings is all over the map. Log from the duet attached. Three other duets live in the same room with a wyze cam, and none of the other three are exhibiting the same issue.
    duetstat.png

    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Request timed out.[duetLog.txt](/assets/uploads/files/1599877358830-duetlog.txt) 
    Request timed out.
    Request timed out.
    Request timed out.
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Request timed out.
    Request timed out.
    Request timed out.
    Request timed out.
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Reply from 10.10.30.23: bytes=32 time=61ms TTL=254
    Reply from 10.10.30.23: bytes=32 time=213ms TTL=254
    Request timed out.
    Request timed out.
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 2, Lost = 2 (50% loss),
    Approximate round trip times in milli-seconds:
        Minimum = 61ms, Maximum = 213ms, Average = 137ms
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Reply from 10.10.30.23: bytes=32 time=689ms TTL=254
    Request timed out.
    Reply from 10.10.30.23: bytes=32 time=978ms TTL=254
    Reply from 10.10.30.23: bytes=32 time=46ms TTL=254
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 3, Lost = 1 (25% loss),
    Approximate round trip times in milli-seconds:
        Minimum = 46ms, Maximum = 978ms, Average = 571ms
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Request timed out.
    Request timed out.
    Reply from 10.10.20.1: Destination host unreachable.
    Reply from 10.10.20.1: Destination host unreachable.
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 2, Lost = 2 (50% loss),
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Request timed out.
    Reply from 10.10.20.1: Destination host unreachable.
    Reply from 10.10.30.23: bytes=32 time=824ms TTL=254
    Request timed out.
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 2, Lost = 2 (50% loss),
    Approximate round trip times in milli-seconds:
        Minimum = 824ms, Maximum = 824ms, Average = 824ms
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Reply from 10.10.30.23: bytes=32 time=59ms TTL=254
    Reply from 10.10.30.23: bytes=32 time=62ms TTL=254
    Reply from 10.10.30.23: bytes=32 time=2ms TTL=254
    Reply from 10.10.30.23: bytes=32 time=18ms TTL=254
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
    Approximate round trip times in milli-seconds:
        Minimum = 2ms, Maximum = 62ms, Average = 35ms
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Request timed out.
    Request timed out.
    Request timed out.
    Request timed out.
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Request timed out.
    Request timed out.
    Request timed out.
    Request timed out.
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Request timed out.
    Request timed out.
    Request timed out.
    Request timed out.
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Reply from 10.10.30.23: bytes=32 time=249ms TTL=254
    Reply from 10.10.30.23: bytes=32 time=60ms TTL=254
    Reply from 10.10.30.23: bytes=32 time=12ms TTL=254
    Reply from 10.10.30.23: bytes=32 time=83ms TTL=254
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
    Approximate round trip times in milli-seconds:
        Minimum = 12ms, Maximum = 249ms, Average = 101ms
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Reply from 10.10.30.23: bytes=32 time=328ms TTL=254
    Request timed out.
    Request timed out.
    Reply from 10.10.20.1: Destination host unreachable.
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 2, Lost = 2 (50% loss),
    Approximate round trip times in milli-seconds:
        Minimum = 328ms, Maximum = 328ms, Average = 328ms
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Request timed out.
    Reply from 10.10.30.23: bytes=32 time=25ms TTL=254
    Reply from 10.10.30.23: bytes=32 time=227ms TTL=254
    Reply from 10.10.30.23: bytes=32 time=23ms TTL=254
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 3, Lost = 1 (25% loss),
    Approximate round trip times in milli-seconds:
        Minimum = 23ms, Maximum = 227ms, Average = 91ms
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Request timed out.
    Request timed out.
    Reply from 10.10.20.1: Destination host unreachable.
    Reply from 10.10.20.1: Destination host unreachable.
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 2, Lost = 2 (50% loss),
    
    C:\Users\Phil>ping 10.10.30.23
    
    Pinging 10.10.30.23 with 32 bytes of data:
    Reply from 10.10.20.1: Destination host unreachable.
    Reply from 10.10.20.1: Destination host unreachable.
    Reply from 10.10.20.1: Destination host unreachable.
    Reply from 10.10.20.1: Destination host unreachable.
    
    Ping statistics for 10.10.30.23:
        Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
    

    Also, if replacing the ESP, is desoldering the only method, or can I use some flush cutters to remove majority of the ESP pcb and desolder the pads separately so I can work one side at a time?duetLog.txt





  • Well. I figured it would be best to know sooner rather than later if I would need a new board.
    KSGER 2.1s station with medium chisel and some genuine dry-aged RadioShack 2mm solder wick from pre-2011.
    20200911_221408.jpg

    All the pads are there! I think all the SMD components as well. ESP-07s is on order. I'll clean the flux and prep for the transplant.

    I was tempted to stick an in-wall HD ubiquiti AP in the printer room, but they're still almost 200USD. Wifi has worked for years.

    I wonder what jlcpcb would charge to make some daughterboards to solder to the bottom/ethernet pins similarly to the ESP. I saw that SBC support for duet wifi was on the 3.2 roadmap....

    Swapping a pi is a lot easier than desoldering ESPs.



  • sbc worked a month or two ago when i last tried 3.2 and i saw dc42 posting preview beta links to troubleshoot issues in another thread so 3.2 shouldn't be too far off methinks.

    edit: btw you don't need a pcb to make the ESP pluggable; won't be super sturdy though

    buy some through hole 2.0mm 8-way headers (or 40 way to break apart). get the right angle female version and straight males. solder males into esp. fit female header onto already soldered male headers and trim leads to stay out if trouble when aligned onto the duet. tack down female headers while attached to esp and aligned to duet, remove esp and solder remaining leads. et viola! (a little fiddly, but doable)



  • @deadwood83 said in Diagnosing a failing ESP-12:

    I was tempted to stick an in-wall HD ubiquiti AP in the printer room, but they're still almost 200USD. Wifi has worked for years.

    a tp-link access point flashed to openwrt only costs a few bucks.

    or a GL.iNet GL-MT300N-V2 very cheap and openwrt preinstalled



  • @bearer said in Diagnosing a failing ESP-12:

    sbc worked a month or two ago when i last tried 3.2 and i saw dc42 posting preview beta links to troubleshoot issues in another thread so 3.2 shouldn't be too far off methinks.

    edit: btw you don't need a pcb to make the ESP pluggable; won't be super sturdy though

    buy some through hole 2.0mm 8-way headers (or 40 way to break apart). get the right angle female version and straight males. solder males into esp. fit female header onto already soldered male headers and trim leads to stay out if trouble when aligned onto the duet. tack down female headers while attached to esp and aligned to duet, remove esp and solder remaining leads. et viola! (a little fiddly, but doable)

    Right, I get that. I was just thinking a way that could come pre-assembled (all smd done by JLC) and then with ~6 touches to a soldering iron you're done. Then you have your wifi and headers with inline resistors underneath running parallel to the board for connecting to an SBC if and when so desired. a bit of ribbon cable and presto. Keeps the clutter down. Then run parallel mp1584 bucks with LC filter to power the pi (if it's a 4B, 3b+ is OK with one). I did a similar setup in my duet powered resin 3d printer. That project looked simple in the brainstorming phase but got complex pretty quickly lol. Eventually I;d like to run that all on an MX8m board (putting motion control on the Cortex M4 core) with hat, but I have been writing the SPI control on the firmware and working a custom kernel for aaaaages. I'm awful at it.

    @Veti said in Diagnosing a failing ESP-12:

    @deadwood83 said in Diagnosing a failing ESP-12:

    I was tempted to stick an in-wall HD ubiquiti AP in the printer room, but they're still almost 200USD. Wifi has worked for years.

    a tp-link access point flashed to openwrt only costs a few bucks.

    or a GL.iNet GL-MT300N-V2 very cheap and openwrt preinstalled

    Had that exact setup with a TP-Link N-band router in that room running OpenWRT. Compared to Ubiquiti's system, it is fiddly at best. The signal advantage was ignored by most clients since the main AP had fewer 'hops'.

    UBNT HD line with a controller (which I run in a VM on my server serving my wireless VLAN) has not only band steering but AP steering. The real draw is not just the wifi but the low-profile with PoE from my main PoE switch (600W of 802.11at) and a 4 port VLAN-capable integrated switch with PoE pass-through. Oh and all the whole remote monitoring and automated config backups, centralized stats into my analytics VM, etc. Would also make cameras a lot more stable and reduce load on that section of my breaker box.

    Heck, after typing that paragraph I bought one. I guess for the price of a new duet I can now convert at least 3 of my machines to ethernet if I so desire.... I think I will get better reception on my toilet as well. Lots of piping, high amperage wiring, and tiling between my throne and current AP.



  • @deadwood83 said in Diagnosing a failing ESP-12:

    (all smd done by JLC)

    sadly the selection of headers is practically non-existent, even as smd headers. oterhwise their serivce is fan-******-tastic:)



  • It's ALIIIIIVE!

    It was the ESP all along. One of the diodes (too small to see?) along the edge with GND and CS had fractured solder.

    Guess I could have baked it, but honestly not worth the effort when an ESP-07s is available.

    20200916_200820.jpg

    I think I now have the confidence to try replacing the E1 driver on my 1.02 board if I ever take it out of the DLP printer.



  • @deadwood83 said in Diagnosing a failing ESP-12:

    I think I now have the confidence to try replacing the E1 driver on my 1.02 board if I ever take it out of the DLP printer.

    there is no thermal pad under the tmc2660 but there are a lot more ground planes over there. preheating the board makes life a lot easier.


Log in to reply