Been messing around here for hours on and off in my free time. Drawing out wiring, searching schematics, tracing out cabling, working out the system etc....
Guy who runs the place just gave me a Museum badge and a free coffee voucher (just one!) for the cafe as a thanks.
Great 
One thing not clear (to me) - did this problem just appear, or has it been there since installation, and it's finally bugging them (or they had a handy guy available suddenly, and figured he might tumble to something, so why not ask).
I believe its been discovered since they took over the building. Various people have had a look and given up.
I have recommended I rip out the shoddy ribbon cables and re-run the 24v AC lines in a more suitable cable and away from the mains as much as possible.
I have said I can't promise anything, but I am also not doing it for free. They have to at least cover the materials and a day of my time.
See what they say
Perhaps the choice of cables was not the best, however I have a question; Was this device working and you are now trying to solve a problem that popped up recently? Or is this a new installation.
If the first, I would assume something had degraded, or perhaps positioning of the cable was changed for some reason (likely unintendedly).
My first inclination would be to inspect / clean the connectors, look for corrosion in anything that is required for a good connection. And measure and post the difference in "0 not being 0"
What exactly caused them to look for help inn the first place? What were the problems/symptoms they were experiencing?
That would be the very worst thing to do. Begin to replace stuff in a rational manner. You have indicated grounding problems. ADD! Grounding cables to parallel the existing grounds. Test!
Then, perhaps parallel the 24 volt wires in twisted pair/shielded wires and test. Remove the old wire connections, test. Do the same with signal type wires using twisted pairs, one at a time. Test.
Don't just rip out and then try to put in new wires all in one big operation.
That could be your issue. You need to look at having a star point, and ideally seperate grounds for power circuits, analog and digital (yes at least three, maybe more "grounds". And a screen.
And decoupling of supplies. Details here.
This is existing and they recently had new heating valves installed. These valves work on a 0-10v position signal.
The valves are significantly 'twitching' when they are supposed to be off. The heating company that replaced the valves have walked away saying they only replaced the valves and are not responsible for the wiring issue.
If I turn off the heating (so there should be 0v on the signal), the heating valves still twitch.
Now I think this is due to two reasons. The 0-10v signals for all these various valves run from a controller using a few 20 core ribbon cables to breakout boards. This then is where the cable is changed to 2 core screened Belden cable out to the valve.
The second issue is the temp sensors are also wired in the same manner. They come from the valves in Belden (and various other types of cable) to a breakin board, where they become a fairly lengthy piece of ribbon cable that connects to the controller board.
If you watch the software that controls all this tat (it's SO over-complicated for what it needs to achieve), you can see the values of these sensors jumping all over the shop.
I have managed to mitigate that somewhat by tinkering with the software settings.
You cannot add any kind of over-sampling or smoothing, but I can add triggering delays.
Ideally it would average out the readings but it doesn't appear to have that function.
Ripping out isn't ideal, but It's not as destructive as it sounds. I am unplugging ribbon cables, not snipping cable runs.
My plan is to make some short, tidy breakout PCBS for the ribbon cable headers (these are the only place to interface into the controller board). From these headers I can then split out the
24v AC power and the 0-10v signal lines to the valves.
I will run those signal lines in screened Belden from the controller to the breakout boards where they head off to the valves.
I can then also run some screened cable for the 24v AC (which I don't think is much of the issue).
See what that achieves. See if that stops the twitching with the heating turned off (so that the twitching sensor readings are not influencing the valves).
If it fails, then I can simply put the ribbon cable back.
If it succeeds, then I can try the same approach with the sensor ribbon cable.
That is my thinking anyway
Hi,
What are the specs of the new valves?
What were the specs of the old valves?
How much current do they draw?
Have you inspected the quality of the new work done when installing the new valves?
Tom... 
Blimey Tom... free job and all that!
No idea what valves were here before. Spec for the new valves? Does that matter?
They require 230v supply which is all fine. Apart from that, they are industry standard 0-10v 23mA control signal.
If you have incoming sensor noise, improving output signals would be my second focus... I would start by trying to filter the sensor signals, as all else is downstream. How you do that, well, it depends on the sensor signal type, but convincing yourself that your grounding isn't responsible would be wise.
As I said before, good luck.
These are the signals coming to the valves? If so then put 10uF capacitors as close as possible to the valve.
Debugging is always hard, I feel your pain 
Think about strategy to apply some kind of system of elimination. For example, can you disconnect the wires for 24V AC, on the other end of ribbon cable and try it like that, to see is it a noise or interference problem, maybe it`s not and you are starting from the wrong assumption.
Okay so they are 230Vac powered and 0 to 10V control.
Tom..
OK. If you disconnect the outgoing ribbon cable from the main control board that feeds the valves, the valves still twitch. Not so much, but they are not static like they should be.
If I then unplug the ribbon cable at the other end at the breakout board..... BOOM, they stop moving.
Its the ribbon cable.
So, I'll fix that little bodge first, and then move on to the incoming sensors.
The fix there I have no doubt is the same issue (or at least some of it)
This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.