Problem with DC power supplies...

[dx593]

An interesting problem came up at work the other day. We were installing a 3-wire pressure sensor (wires are COM, DC+ 24v, and SIGNAL (0-5vDC).
The transfo feeding the controller is 120v-24v AC, one side grounded on the secondary. The DC power supply for the sensor is fed from the same transfo; it has as outputs + and - stamped on the circuit card. It is set up as a half-wave rectifier.

The COM and DC + of the sensor were wired into the DC outputs of the power supply; the SIGNAL went to an input on the controller (exactly as per the instructions).

However, we were getting very erratic readings on the input. Checked the voltage from COM to DC+, 24v. However there was a difference of 9v DC between the controller common and the DC power supply common. Changing
the AC wire positions feeding the power supply made no difference.

We swapped the power supply out for a different one; this one had as outputs DC+ and COM stamped on the circuit card. Now there was no difference between the DC COM output and the common on the controller. Figured that was one problem solved.

However we now measured 24v AC on the DC COM! We swapped the AC inputs on the power supply and bingo, no more problems. Sensor is now reading fine.


Now my questions are dealing with the internal circuitry of the power supplies. First, why would the first PS have as outputs POS and NEG, and the second POS and COM? Is there an internal difference between the two? I thought all of these linear power supplies were basically the
same.

And how did we have 24v AC on the common? And why did swapping the AC inputs on the power supply correct it? (I have noted that some power supplies have the AC waveform symbol next to the inputs, and others say AC and COM, or HOT and COM. Is there a difference in the internal
wiring?).

If anyone can shed some light on this I would appreciate it…


Ross McGregor
dx593@freenet.carleton.ca

 

[peebee]

Sounds like your first DC PS had a floating output, whereas the second one had a solidly-bonded common negative side. The output of the first PS had no reference to the AC input, the second one did.

When you replaced the PS with the new one, you apparently referenced the DC output to the AC hot. Once you swapped the input leads, you instead referenced it to the AC neutral/ground.

 

[dx593]

BTW, neither DC PS had a ground wire. So internally, how did the second PS have a reference to the AC input? They both had a 4-diode rectifier, and were set for half-wave via a jumper. Mmmm. Maybe I'm gonna have to break out the old breadboard and make up some PS circuits and try to duplicate this. It must be something to do with how the different regulators are wired up or something. It actually is a bit of a problem on a few of our sites; some stuff got wired for floating AC; other stuff has a grounded secondary. Then something else gets added and problems pop up. So it might be worthwhile to find out exactly what is causing it.

R McGregor

 

[IRstuff]

a> You should a qualified electrician to check your power supplies before you try to do anything else.

b> You should get an electrical engineer to do the electrical design and checkout. You think you know something about electricity, but you've made a number of incorrect assumptions, one of which could have proven fatal to you or someone else.

c> You've given no information about the brand of power supply and you expect a conclusive answer.

There are literally at least a dozen ways to implement an AC to DC converter. At least one of them will have a direct path between the AC neutral and the DC common. Reversal of these leads could prove fatal to the user, as you could potentially connect the AC hot side to the DC circuit. Not to mention the possibility of damaging whatever it was that you connect to it.

TTFN

 

[automatic2]

Yes, be careful. On your half wave, the AC Nuetral is referenced ground, as Hot provides the potential for rectification. On a full wave, your DC com output is floating. Doen't be putting your AC Hot on the Neutral connector in a half wave. That could really hurt.

 

[Heydave]

It sounds as if the the power outputs you are using are floating but there is nothing that totally "floats". There ia always some coupling. In a typical transformer, the primary line windings are layered. The winding layer on one end will be commected to the power neutral while the other connects to hot. If the secondary windings are also layers, then there is capacitive coupling to the learest power layer. If that layer is hot, then there will be a capacitive coupled AC voltage into the secondary. Reversing the power wires will reduce this. Even transformers with split windings can couple through the core. Some transformers solve this problem with a grounded shield between the primary and secondary but the added cost and required ground reduce the use. Switching supplies can have a similar problem related to RF filtering circuits.

If you truly bonded to primary line hot with common, you would have primary line voltage, not 24 volts. The value of 24 volts makes it suspicious that there are other pathways involving the secondaries and regulators.

 

[peebee]

One of us is missing something, Heydave. . .

The DC source was fed from a 24VAC transformer, so the primary voltage to the DC supply is 24V. That seems to negate your second paragraph above unless I'm lost somewhere . . . .

 

[dx593]

Thanks, Peebee, that was what I was looking for.

I examined the second PS, and as you mentioned, one of the AC inputs is connected to the DC common.

What is strange is the installation sheet that came with this PS. It shows the connections between the transfo and the PS, and has the *fused* side of the transfo going to the AC input that is, you guessed it, the one that is connected to the DC com. That is why when we first hooked it up we had 24 v AC on the common.


Heydave makes an interesting point. I wonder if that *coupling* between the primary and secondary would explain why floating 24v outputs (on a 120v/24v class 2 transfo) often have one side higher than the other (with reference to ground). For example, one side might be 12v AC to ground and the other only 1 or 2v AC to ground...