RS485 network hardware questions / details

[lesthegrngo]

Hi all, this is a thread that follows on from the posts in the 'DCS BIOS over RS485' thread, but is more to discuss hardware setup rather than the generic issue. I didn't feel it was right to hijack that thread for this.

To sum up the story so far, my A10C cockpit currently uses about 40 slave Arduino Nano's. At the suggestion of the very helpful people on this site I have gone for three Megas to be used as masters, one of which has three MAX487 chips so has effectively three RS485 busses, the other two have one Max487 chip each. This was originally to allow one RS485 bus for each 'leg' of the cockpit, so left console, main dash and right console.

After trying out the networks, I found that I had some unwanted behaviour of the stepper motors if I connected any LCD or OLED connected nanos to the same bus as stepper connected ones, and so the two single Mega Masters will be used solely to drive those devices with steppers, and the three way bus Mega will be used to drive the rest. This works well as the steppers are all in the right console and dash, so they can be treated as individual legs.

As an attempt to fix the strange flickering of the steppers I put in a dedicated 5A 5v power supply, however one of things that I didn't do was run a discrete 5v bus to each device, using the RS485 wiring to run the power. I am going to retry the setup I have with a dedicated 5v bus to see if that removes or reduces the stepper twitching. I would appreciate any input on this

For the Mega masters, I believe that using a USB2.0 connection is fine and will be using a powered USB hub for this purpose. Again, if anyone has feedback I would like to hear it.

Depending of how successful I am in eradicating the unwanted behaviour, I may revisit some of my PCB's to decouple the devices still further. For example the nano driving the OLED that is used to display the fuel quantity is on the same PCB as the nanos driving the Hydraulic gauge steppers and fuel gauge stepper, and so the RS485 connection for those are common. I have a dual axis stepper in the post for the fuel gauge to properly implement it so will take that opportunity to run two independent RS485 busses on that PCB. 

Cheers

 

Les

[No1sonuk]

You could try a largish capacitor where the 5V goes into each arduino. 100uF more or less.  Be careful, though: capacitors that big are usually electrolytic and have to be fitted the right way round.

That should smooth out the dips.

[lesthegrngo]

Thanks - I'll try it with a pinned module so I can install or remove it to see if there is any difference

Les

[Vinc_Vega]

Hi Les,

what power input do you use for the Nanos? Following has to be considered:

If using the VIN power input pin, up to 12V can be applied to the board and internal regulation will be used to limit 5V to the chips and outputs. In case you supply less than 7V at VIN, the regulator may consume more than 2V. Than probably less than 5V will reach the chips and the board may be instabile.

If using the 5V in pin, the Nano's power regulator is bridged and the full voltage goes into the chips and the output pins. But than the voltage supply has to be stabilized by the external capacitors.

 

Regards, Vinc

[lesthegrngo]

I use the 5v pin, so it sounds like I'll have to get busy with the capacitors. However I have a question,  if I have multiple nanos hardwired to one PCB, do I just have one capacitor for the entire PCB?

Interestingly, today I was trying to get the new monitor lua set up, and connected up the RS485 devices just to mess around, but without any attention to what was connected to what. As it turned out, everything ended up connected to just one of the busses on the master with 3 MAX chips on it, and there was no hint of the stepper twitching. I've left it how it is to look at tomorrow to see if there is something. Apart from where the connecting wires go, nothing else has changed, so I have to confess to being confused 

Cheers 

Les

[Vinc_Vega]

Theoretically, the capacitors should be as close as possible to the affected chipset. Practically, I would start to smooth out the voltage of the whole PCB by one capacitor only. If that doesn't work, each Nano may have it's own capacitor.

Regards, Vinc