Voting resources, early voting, and poll worker information - VOTE. ... Adafruit is open and shipping.
0

Unexplained Equipment Failures
Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.

Unexplained Equipment Failures

by silverstar on Sat May 09, 2020 8:57 am

I have a project that controls the air conditioning in a railroad car. The AC system is quite large and requires 480 3-phase power to operate. I control the system by managing the 120VAC relays that control the compressor and condenser. My device is located within 3 feet of the 480 volt relays. The problem that I am having is that the temperature sensor will occasionally cease to operate correctly when one of the 480 volt relays operates. The only way to recover is to physically unplug my device and then plug it in again. The device must operate, unattended 24/7 and it is located 45 miles away from my home.

I am not an electrical engineer but it seems to me that there is enough electrical interference created when one of the 480V relays activates to cause my sensor to fail occasionally. I need some help getting this project to work reliably. I am currently using the AM2302 sensor. In the past I have used the MCP9808 sensor which seemed to fail more frequently than the AM2302. I have located the AM2302 about 5 feet further away from the AC equipment. Note that there is no obvious problem with the Adafruit Metro when the relays fire.

I am asking for help with this because its important that this device work reliably

silverstar
 
Posts: 314
Joined: Tue Feb 04, 2014 10:46 pm

Re: Unexplained Equipment Failures

by dastels on Sat May 09, 2020 11:55 am

Are control lines for the relays opto-isolated? Do the relays have shunt diodes on their drivers/coils?

Dave

dastels
 
Posts: 3532
Joined: Tue Oct 20, 2015 3:22 pm

Re: Unexplained Equipment Failures

by silverstar on Sun May 10, 2020 8:55 am

I do not know but I will find out. The contactors in question are probably25 years old. Standby.

silverstar
 
Posts: 314
Joined: Tue Feb 04, 2014 10:46 pm

Re: Unexplained Equipment Failures

by adafruit_support_bill on Sun May 10, 2020 9:43 am

Between the relay coils themselves and the compressor motor loads they are switching, it sounds like there is plenty of potential for interference there. Even with optical isolation, it could be coupling inductively. Twisted pair and/or shielded cabling can help there. And if you have any flexibility in the cable routing, try to keep your controller and sensor wiring separated from the high voltage/high power wiring wherever possible.

adafruit_support_bill
 
Posts: 78758
Joined: Sat Feb 07, 2009 10:11 am

Re: Unexplained Equipment Failures

by silverstar on Sun May 10, 2020 12:45 pm

Note: the basic 120V wiring in the car is 120 3-phase thanks to the type of transformers used to change the 480 to 120. Odd arrangement, I know.

I use a Adafruit #3191 (driven by a Metro) in line with one of wires leading to the coil. To keep the 120 volt line as short as possible, my controller is located very near the relay. In addition, this makes it possible to place the temperature sensor in the plenum.

From what you are advising, the controller and temperature sensor should be located further away from the 480 lines. Much further? My problem then is controlling the relay. I suppose a BLE device would also be negatively effected by virtue of its proximity to the 480 lines.

silverstar
 
Posts: 314
Joined: Tue Feb 04, 2014 10:46 pm

Re: Unexplained Equipment Failures

by adafruit_support_bill on Sun May 10, 2020 1:09 pm

EMI problem solving is as much a black art as it is a science. There are theories that explain everything, but even if you could model the whole system, the math would be pretty scary.

So we usually start working through a list of 'best practices'. Typically, some combination of (relatively) easy remedies will bring the problem under control.. But sometimes you need to put an oscilloscope on it to better understand the nature of the noise.

For inductively coupled noise, the first rule to remember is that everything is an antenna. The wires powering your compressor are broadcast antennas for EMI. And any wire connected to your sensor and/or controller is a receiving antenna that can bring that noise into the system.

There are several things you can do to minimize this inductive coupling. These can be use separately or in combination. Every application will have different constraints and tradeoffs.

* Keep wiring as short as possible.
* Keep the wiring separated as much as is practical.
* If wires must cross, have them cross at as close to a 90 degree angle as possible.
* Shielding the wiring (this can be via shielded cable, or by running the wires in separate conduits).
* Use twisted pair wiring.

For conductively coupled noise:
* Make sure you have snubber diodes on all coils and inductive loads.
* Use power-line filters for AC circuits
* Add bypass capacitors to DC supplies
* Add ferrite chokes on wiring coming into the processor.

https://www.prelectronics.com/support/p ... n-signals/

adafruit_support_bill
 
Posts: 78758
Joined: Sat Feb 07, 2009 10:11 am

Re: Unexplained Equipment Failures

by silverstar on Sun May 10, 2020 7:45 pm

Thank you for your help. I am meeting with our electrical people tomorrow or Tuesday and I will discuss your messages.

The only part of the project that is effected in this manner is the temperature sensor. I could power down the sensor when not needed then power it up for a reading (then shutting it down) before activating the relay.

If you wish, I will update you on our project. Thank you, again.

silverstar
 
Posts: 314
Joined: Tue Feb 04, 2014 10:46 pm

Re: Unexplained Equipment Failures

by adafruit_support_bill on Mon May 11, 2020 5:21 am

Yes. Let us know how things work out.

adafruit_support_bill
 
Posts: 78758
Joined: Sat Feb 07, 2009 10:11 am

Re: Unexplained Equipment Failures

by silverstar on Wed May 20, 2020 7:13 pm

I finally had an opportunity to speak with our electrician. He said that there is not much we can do about the heavy electrical components on the rail car but he agreed that using shielded twisted pair would most helpful. I will replace the wire as soon as possible. I intend to modify my controller and software to control the power to these devices so that unless they are needed, they have no power and once needed, they will have power only long enough to do their job. More on that later as the project continues.

silverstar
 
Posts: 314
Joined: Tue Feb 04, 2014 10:46 pm

Re: Unexplained Equipment Failures

by lefs on Thu May 21, 2020 4:07 pm

A 100nF-1uF capacitor between ground and supply voltage near the sensor sometimes helps. If the problem is in the sensor. If there are supply voltage spikes, some kind over voltage protector could help too. And a second capacitor between sensor output and ground could help, if it is the sensor output which has problem. This capacitor slows the output but I don't think it matters with relays.

Relays sometimes gives nasty voltage spikes but because they are slow, there can be filters and checks everywhere, there is no rush.

lefs
 
Posts: 23
Joined: Mon Jul 30, 2018 10:00 am

Re: Unexplained Equipment Failures

by ductsoup on Sun May 24, 2020 9:40 am

Stuff like this can be very difficult to track down.

It sounds like you've basically got two paths for interference. It'll be either radiated EMI, in which case the shielded twisted pair should help. Or, it'll be inrush from the relays (either side). Filter capacitors should help on the low voltage DC side you're using for the control. If it's on the AC side you might consider adding some varistors (MOV) to mitigate the voltage spikes.

You might also want to consider a PLC rather than a micro controller for something like this. They're more suited to environments such as this and are very inexpensive these days. I would typically use the Automation Direct CLICK for something along these lines. One similar application I did a while back was a two story, gas fired 1M BTU backup heating system that I needed to bring under building automation control. Primary power was 480/3, control power was 120/1. It interfaces with the SCADA system via Modbus-TCP.

ductsoup
 
Posts: 29
Joined: Thu Nov 29, 2018 6:49 pm

Re: Unexplained Equipment Failures

by silverstar on Sun May 24, 2020 6:43 pm

we, initially, had some trouble with what we believed to be inrush so we moved the Arduino and all its parts to a different circuit that was powered by different 480 legs and that seemed to have fixed that problem. Sad to report that the decision has been made to fully replace the old temperature control system with a new system featuring a direct replacement for my project. So the problem is solved but by vastly different means.

I want to thank you very much for your interest and support.

silverstar
 
Posts: 314
Joined: Tue Feb 04, 2014 10:46 pm

Please be positive and constructive with your questions and comments.