DC voltage reading on a single leg?

Thread Starter

bootsnjeans

Joined Dec 5, 2018
6
We recently installed some new equipment and found an issue with one of the relays in the 24VDC control circuits. The relay was sticking open, so we checked the power to the relay. Oddly enough when the digital meter common (black) probe was placed on the 0VDC terminal to the relay, the meter showed voltage... the other probe (red) from the meter wasn't touched to anything. Thought the meter was bad, so I tried another meter and got the same result. Has anyone out there ever seen a digital meter back feed like this to give a reading when only touching one side of a circuit? If so, what is causing this?
 

Thread Starter

bootsnjeans

Joined Dec 5, 2018
6
A little over 60 VDC on just the black probe. When we tested properly across the circuit, about 35VDC and variable; tested in reverse, -24VDC.
 

Thread Starter

bootsnjeans

Joined Dec 5, 2018
6
Thanks nsaspook, I read the article. The strange thing in our situation is that the circuit was never completed through the meter; it has a reading on DC voltage with on just common side alone. I tried touching the positive side alone and get nothing, as you would expect. However with just the common probe of the meter, you get a reading.
 

ebp

Joined Feb 8, 2018
2,332
With a meter with unusually high input impedance this might be less unexpected, but most hand-held DMMs used in industrial environments have 10 megohm input resistance, which is not what I'd regard as particularly high.

My suspicion is that there is high-amplitude high-frequency unipolar (or apparently unpolar when averaged) noise on the black wire. With noise, the meter itself and the other lead, particularly if held onto by a person, would for a capacitive path or "antenna" allowing high-frequency current flow. There is effectively some capacitance across the meter's input, so narrow noise "spikes" could charge that capacitance sufficiently to produce an appreciable DC voltage - but 60 V is huge! 60 VAC wouldn't be too surprising. There will be substantial inductance in a long connecting wire. The noise might not exist at the other end, but instead be being coupled in form some other source via capacitance.
 

Thread Starter

bootsnjeans

Joined Dec 5, 2018
6
In this situation no one was touching the red lead, so it would have to be acting as the antenna itself if your theory holds ebp. I agree that there is some type of capacitive coupling happening here that is adversely effecting the 24VDC system... the voltage readings are all over the place when the circuit is properly completed (35-45VDC and the circuit should be steady at 24VDC), so obviously there is some sort of bleed over interference. I have just never seen a volt meter read off of one probe like this with a seemingly uncompleted circuit, so thought I would share and see just how rare this might be.
 

Tonyr1084

Joined Sep 24, 2015
7,852
Been a long time since I saw a meter do something similar to that. My first thought was a high frequency bleeding through.

A very long time ago I installed burglar alarm systems. One home had a habit of triggering the panic circuit every time the owner set the alarm. She was getting furious with us because when we'd come out we couldn't make it trigger. So we replaced the base unit. Same issue still existed. So we thought there could be a bare wire touching inside the multi-lead cable. We replaced that. No change. I set the original base unit on a test bench and set up a mechanical switching device to simulate turning the alarm on and off repeatedly. Ran it for weeks and it never triggered the panic circuit.

So we isolated the panic circuit into its own two twisted pair with shield and that solved the problem. The close proximity of the main circuit wiring to the panic wiring was bleeding over and the base unit was seeing that as a panic event. We encountered that problem a second time at a different home. Setup was similar in that there was a long run of wire between the control pad and the base unit. After that we just ran the panic circuit in "Long Run" setups as a separate cable.

Does your circuit have a long run of wire? Are there multiple wires in a single sheath? Maybe within a single shielded cable? If you're not able to run separate wiring, I'm wondering (since I don't know) if maybe putting a capacitor across the input to absorb those bumps in the signal. Or maybe a choke. Like I said, I don't know the answer. Hopefully this may trigger a different line of thinking that may lead to a solution.

Peace all. And happy holidays if I don't respond to this thread again.
 

ebp

Joined Feb 8, 2018
2,332
I should have clarified - I wasn't suggesting galvanic connection to the free test lead end, just holding onto the probe body so there would be additional capacitance.

I've certainly never seen DC readings anywhere near that high on a meter with one lead floating. What kind of environment are we talking about? Is there machinery with things like variable frequency drives or perhaps electric smelting furnaces, induction heating equipment or other "unusual" equipment nearby?

I would begin by going back to what is supposed to be the 24 volt source and measuring it. If it looks OK at DC, measure the AC voltage present across its output. Some DMMs (most Flukes) have a DC blocking capacitor for the AC ranges, but with some you would need to use an external capacitor to block DC. If you measure more than a fraction of a volt of AC, I would suspect something wrong with the power supply unless the DC voltage isn't regulated and then you might see AC due to line frequency ripple (usually twice line freq due to full wave rectification) on the output. Lack of regulation might get the voltage up by 20% or so, but any more than that seems unlikely (and relays get too hot if you run them much above their rated coil voltage). An unregulated supply using a small transformer with full wave rectification and capacitive filtering could be 30% or more above nominal with very light loading.

You may need to obtain the use of a suitable oscilloscope. Ordinary scopes with ordinary probes can be a problem in industrial systems because of the connection between the probe ground and AC mains ground. Some battery operated scopes are good for industrial work. You can get "differential" probes you can use with any scope that allow reading between two points, neither at mains ground.
 

Thread Starter

bootsnjeans

Joined Dec 5, 2018
6
I should have clarified - I wasn't suggesting galvanic connection to the free test lead end, just holding onto the probe body so there would be additional capacitance.

I've certainly never seen DC readings anywhere near that high on a meter with one lead floating. What kind of environment are we talking about? Is there machinery with things like variable frequency drives or perhaps electric smelting furnaces, induction heating equipment or other "unusual" equipment nearby?

I would begin by going back to what is supposed to be the 24 volt source and measuring it. If it looks OK at DC, measure the AC voltage present across its output. Some DMMs (most Flukes) have a DC blocking capacitor for the AC ranges, but with some you would need to use an external capacitor to block DC. If you measure more than a fraction of a volt of AC, I would suspect something wrong with the power supply unless the DC voltage isn't regulated and then you might see AC due to line frequency ripple (usually twice line freq due to full wave rectification) on the output. Lack of regulation might get the voltage up by 20% or so, but any more than that seems unlikely (and relays get too hot if you run them much above their rated coil voltage). An unregulated supply using a small transformer with full wave rectification and capacitive filtering could be 30% or more above nominal with very light loading.

You may need to obtain the use of a suitable oscilloscope. Ordinary scopes with ordinary probes can be a problem in industrial systems because of the connection between the probe ground and AC mains ground. Some battery operated scopes are good for industrial work. You can get "differential" probes you can use with any scope that allow reading between two points, neither at mains ground.
This is on a spinning lathe machine. The power supply reads a good 24VDC output. The power source for the equipment is 480V 3PH and it is conditioned before going to the machine. There are really no long runs on this like you would typically expect to see in a bleed over scenario. And, none of the motors on the equipment were running when we tested, so it's not signal noise type issue due to rotating equipment. I'm honestly at a loss, but luckily this is a brand new machine so perhaps the manufacturer will know the fix.
 

danadak

Joined Mar 10, 2018
4,057
But the 480 supplied to the motor switch is driving some
current thru stray C...............

Just a guess. Disconnect 480 at mains breaker box and see what happens.

Regards, Dana.
 

Thread Starter

bootsnjeans

Joined Dec 5, 2018
6
But the 480 supplied to the motor switch is driving some
current thru stray C...............

Just a guess. Disconnect 480 at mains breaker box and see what happens.

Regards, Dana.
Unfortunately can't do that... the 480VAC supplies all the stepdown transformers, including the 24VDC power supplies. When testing, we have all of the motors shut off, so there should not be any load on the 480VAC feed other than stepdown to the control system circuits.
 
Top