Thermotron 8800 RS232

[J.New]

Hello.  I am interested in comments anyone has regarding inconsistent behavior of different versions of Thermotron 8800 controllers over RS232.  Here's my story...

I have software built on LV2014 that has been in use on a variety of test stands for quite some time.  It works fine at 9600 baud rate with standard settings on any 8800 controller with software version 2.xx on it.  The newer 8800 controllers that have been coming in are in the 4.xx version range.  The newer ones have been having issues with VISA sessions disconnecting at 9600.  We have been able to get around this by turning the baud rate up to 57600.  But on the controllers that needed 57600, the VISA session still occasionally disconnects.  It was rare enough that I gave up trying to figure it out.

Recently we had a 2.xx controller fail and replaced it with a new 4.xxx version.  This one is the worst yet.  It doesn't like any baud rate.  I can get a thermal started at 57600, but it won't last long.  We tried other known good PCs, cables etc.  This one just doesn't like RS232.

When I look at the Thermotron log of commands it receives, there are some garbled characters in addition to what the stand sent out.

So I decided to try switching to TCP/IP.  It works great.  I've had it cycling for two days now with no issues at all.  So that's the answer I'm going with. 

But people are asking why after all these years of RS232 working fine would we need to switch to Ethernet for newer controllers. 

Anyone have similar experience with these?

[J.New]

Well.  I have to add to this that Ethernet wasn't the perfect answer either.  I ran a longer thermal with more frequent ramps than our usual ones.  It ramped 36 times and let go on 37.  I used a standard network cable.  I found one with better shielding and will try again to see if that helps.

[cweeks12]

To me, it looks like there's one common denominator throughout all of these tests, and that's the controllers, especially the new ones. Is there anything in the spec sheet that looks like it could be causing any issue with this?

[Minions]

Both RS-232 and Ethernet are expecting cables with 100 ohm impedance and a low capacitance.  Using an impedance controlled cable creates the most ideal interface.  The distance and baud rate should be somewhat inversely proportional.  What are you using for the connection medium?

[J.New · Accepted solution]

So here's what I found...

It seems that the newer versions of the controllers take longer to process a command.  The original coders only put a 10ms delay at the end of their command loop and do not wait for an acknowledgment before sending another command.  That was fine for the older ovens.   I had to increase the delay all the way to 2000 for this particular controller.  I'm not experienced enough to know if I should have to wait that long or not... but it seems to have resolved my issues.