Need some advice on what to do with this.
Situation is this, I have two modbus loops with four generators on one and three on the other. We had some initial trouble with spotty timeouts on the generators and on the generators themselves we would get a amber caution light at the time we didn't know what that was but we worked on decreasing point refresh rates and increasing modbus timing and it all went away for about six months but now is back in only one generator. Mostly what I am asking for is a sanity check on my timing values below.
The problem today: On the loop that has four generators I have ONE genset that has some points that come back with timeouts and then will work fine on the next poll. On the generator itself it has the amber caution and according to the generator service tech the alarm is "Logic Scada comm not active." All the other generators are fine, all points coming back no errors at all. Anyone think this is a problem with the generator? I'm thinking yes.
Generators are Cat with logic EMCP 3.3, which is pretty standard. This is one of those rare situations where we are talking to the generators directly and not through some ATS switchboard.
Modbus 485, 9600, 8N1,
Response Time-Out: 600 ms -----go even further?
Intercharacter Delay: 40 ms -----decreased to default 20 no change
Interpacket Delay: 50 ms -----increased to 90 no change
Maximum Retries: 5
In the equipment file we are polling 22 analog points with refresh times of 1 minute to 3 minutes for the less important stuff. Also are polling 9 digital points with refresh times of 30 seconds. IMO we aren't asking for a lot of information here.