TCP Client & Server - DLS-W Troubleshooting and Config changes

[alienatedsec]

Following the latest commit c8a107e, I have made changes to the configuration of the TCP Client to ensure its stability for my use case. Unfortunately, my datalogger DLS-W is bombarded with requests (more here), and sometimes it produces spikes on the Solis Cloud (an example below). The spike before 10:00 made me experiment with more aggressive settings which caused spikes at around 20:00 proving the point for further amendments.

Therefore, I had to tune up some settings to ensure those spiking issues are mitigated. The change seems to stabilise datalogger connectivity and prevents producing spikes on the Solis Cloud portal in certain scenarios (definitely in mine), and you may not have experienced the same.

Regardless, those spikes are not contributing to any daily production figures but would make the graph unreadable.

This change is still being tested, and I will keep updating this discussion with any progress (or regress).

[alienatedsec]

So, the latest configuration used to work for about a month and started playing around with the above spikes again. I looked into it a bit closer and amended some Framing Rule settings, which are currently under review - those spikes are likely because of my unusual setup and overloading the datalogger, but I am testing if those can be mitigated via Waveshare settings. Looking into the documentation, the frame settings are unusual and can be amended to anything outside of those numbers.

I set both TCP Server and TCP Client to match the following settings. It looks like Framing Rule could be equivalent to Ethernet MTU. As much as I am not expecting support for jumbo frames, I expect 1500 to be the norm (as the rest of my network). This is working and will be updated if stable for at least a week.

[alienatedsec]

Further amendments (part 4) to the TCP Client in More Advanced Settings. The below change is based on documentation Section 7.1
https://www.waveshare.com/w/upload/4/4d/RS485-to-eth-b-user-manual-EN-v1.33.pdf

[alienatedsec]

73cb1bc submitted for further testing

[alienatedsec]

The latest changes have not improved the situation with my DLS-W datalogger where I had occasional Cloud spikes; weird, as some previous settings used to be working for several weeks. The recent changes to TCP Client still have not improved anything and there was no real pattern to observe. That made me question something else - my in-home infrastructure.

Therefore, I reconnected my DLS-W Datalogger to another WiFi network - again, no problems since that change - just wondering for how long.

[alienatedsec]

Unfortunately, changing Wi-Fi has not helped.

In the end, it turns out that my DLS-W datalogger could be impacting the Solis Cloud readings. There are several reasons to believe it is at fault:

• Solis Cloud was the only one affected by a wrong data with my current setup.
• My local SolisMon3 queries and its respective database are correct - there are no spikes or false readings like in the cloud.
• The situation is not getting improved regardless of Waveshare settings, and it is intermittent.
• I hear no other people are affected by the same issue.
• I already had some historical problems when I used my DLS-W with the Solarman integration back in November 2022 - see Invalid modbus packets accepted and give corrupted data StephanJoubert/home_assistant_solarman#120

As the outcome, I replaced my DLS-W with S2-WL-ST - enough is enough.

I did reset both Waveshare devices to defaults, and I configured the following:

• DHCP on both Waveshare devices, with reservations, so the same IP is with both devices at all times
• TCP Client port is set to 0 by default
• Restart for no data unticked
• Framing rule is also default - 1300 Bytes and 3 ms