Gubi,
Good success last night! Like you, I have had issues with WiFi when the bandwidth gets crowded. I have also had problems with the Pi4 when two busy USB3 cameras were both slamming the system. These problems are not limited to USB, but also impact networking.
When I asked about the Skywatcher HOTSPOT, my use of the word "router" wasn't particularly appropriate. While most HOTSPOTS will allow multiple connections, some do not allow data to pass from one client, through the HOTSPOT to another client also connected to the HOTSPOT. Others do. I know the general RFC for HOTSPOTS is to pass data between clients, but I have run into a couple of phones and other devices that dead-end at the HOTSPOT or route directy outbound to the Internet and directly back to the original client. I was thinking about taking the telescope off my local WiFi network and using the HOTSPOT as a stand-alone network if it would communicate between clients. I'll test that.
My setup, as of last night is a Pi4 named pi4-scope (with an attached Waveshare HAT board) running indiserver as my primary connection point (over WiFi) using:
indiserver indi_wmh_focuser indi_skywatcherAltAzMount indi_asi_ccd "SVBONY SV305 0"@pi3-scope
and a Pi3 named pi3-scope running:
indiserver indi_sv305_ccd
all controlled by a Linux laptop running Ekos, also on the local WiFi network.
Where the focuser is the pi4-scope onboard Waveshare HAT, and the skywatcherAltAzMount connected via my local WiFi network over UDP:11880, the ZWO camera is on a USB3 port of pi4-scope and the SV305 is connected to a USB2 port on pi3-scope. I have connected pi3-scope and pi4-scope via a short null ethernet cable configured as a 2-client simple network. My goal in doing this was both to reduce network traffic and USB load, as I had it all plugged into the Pi4 originally. As it happens, there were no bandwidth or USB issues last evening with this configuration. All worked without connectivity issues.
I used strickly Ekos last evening. Plate solving was great. Both tracking and slew worked well. I was able to go from Saturn, to Jupiter, to M2 and finally M15. Each slew was a bit off, but corrected automatically via solve and slew. The automation used about 3 adjustments for each target, but they all ended up right dead center on the screen.
The only minor problem I had all evening was on the trip from Jupiter to M2. It went about halfway, did a little dance, then gave up. Similar, but not as crazy as the synscan driver in about the same area of the sky. All I had to do to recover was to click slew to target again, and it finished the trip without further issue.
I didn't do any guiding last evening, partly because I tested PHD2 with the new configuration and it connected. When I can get PHD2 to connect, it is very straight-forward and effective. Tonight, I'll prove that guiding does in fact work by doing some actual imaging. I have high hopes. The skywatcherAltAzMount really did work a LOT better than the synscan_telescope driver, and I didn't have to add the tablet into the mix. I really appreciate all the work you did on that driver to get it goiing (especially the UDP bit).