TIDA-010935: Questions about multi-node and device selection

Hi Kelly,

For you direct questions:

1. 5 nodes should be fine - each node is high impedance and in high voltage applications the system impedance is very low so the addition of extra nodes really isn't concerning until you probably get significantly higher node counts - I don't know of any specific node counts used in systems with this design - most of my knowledge is that multiple groups are testing this for their own systems but I don't think anything is in full production yet ( I could be wrong - it has been a few months since I have checked this application space in depth) 

2. Switching to the THVD8010 could result in lower transmission distances due to the fact that you need a much larger signal swing than the THVD8000 to register a logic 0. For the THVD8000 the signal magnitude would need to reach between 150mV and 225mV depending on frequency chosen (due to lower data rate most likely closer to the 225mV signal) - where the THVD8010 you would be looking at 1.2V for 20kbps to register a logic 0. How much that cuts the distance really depends on how much the THS6222 signal is attenuated across the bus - since the THS6222 provides the drive strength in the system you may still get a good distance but THVD8000 would be able to travel farther. The THVD8000 can work at lower speeds and if you don't need the large hysteresis in the system I'd stay with the THVD8000 - if the system is extremely noisy the THVD8010 can be a good choice but it will come at the trade off that you will have a shorter max distance - I can't give a straight number because the system variables really will impact max distance (cabling, environmental conditions,  and other circuitry that may exist) 

3. Transformer selection is generally pretty simple for this application - you want it to be designed for your signal range - so low attenuation due to leakage inductance essentially. That is the main concern - we want to minimize data signal attenuation over the transformer so leakage inductance should be very low relative to modulation frequency - but beyond that it is very flexible on what transformer you use. 

4. I am not sure why the nodes are in series - they should be in parallel like the block diagram; CA1 and CB1 are extremely necessary - you can not remove these - if you don't you don't create the proper blocking network to protect the communication node. 

5. The star network question is challenging for this application. In general to use Star configurations with THVD8000 we do suggest that you use a THS6222 as well as terminate every single node on bus with equivalent characteristic impedance of bus - for general RS-485 this is 120 ohms - but for power applications like this - most likely the cabling isn't impedance controlled. Ultimately to use a Star network every node needs to be impedance matched which is going to be very hard to do in these applications. Since your data rate is lower and you can use a lower modulation frequency - the reflections could be relatively small depending on bus length - but my biggest worry is data corruption due to reflections that are harder to control in these systems. You could try terminating between differential I/O before the transformer to potentially dampen some reflections - but I am skeptical as how much that would actually help. 

Please let me know if you have any other questions and I will see what I can do!

Best,

Parker Dodson