Author Topic: Should A Good Power Supply Have Measurable Voltage/Current When Channel OFF?  (Read 5703 times)

0 Members and 1 Guest are viewing this topic.

Offline jstarrTopic starter

  • Regular Contributor
  • *
  • Posts: 67
 I bought a Rigol DP832.  I found that with all the channels OFF I can still measure (on a NIST Agilent 6 1/2 digit meter) 3.8mV @ 1.4mA out of Channel 1; and 6.9mV @ 35uA on Channel 2.   This supply reads to 1mV and 1mA display resolution and I find that the display readout on the supply appears to be off by this very amount of phantom voltage and current.  For example, if I set the supply voltage for 5.000 vdc the readout will show on Ch 2 .... 5.007 vdc and the measured voltage on my is 5.007v.  Would re-calibration effect the output when Channel is supposed to be switched off?

 Is it normal to have anything other than zero out with the Channel off?   Does anyone else have a measurable voltage/current when channels are OFF?

john
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Is the negative terminal floating when you make this measurements?  If not, you could be reading part of a ground loop voltage.

Recal should help, but the response is still wierd because it means the output is not truly switch off; it could be the effect of a soft switch since they have finite impedance over a mechanical switch.

When isolated, and channel is off, the voltage must be zero after x sec; how much depends.  The output capacitor must be bled by internal circuitry so it will drop by whatever RC time constant is in the design.  Any residual voltage between channels is generally not good, it means each is not fully switch off and worse, the channels are not fully isolated. 


 Is it normal to have anything other than zero out with the Channel off?   Does anyone else have a measurable voltage/current when channels are OFF?

john
Best Wishes,

 Saturation
 

Offline dadler

  • Supporter
  • ****
  • Posts: 851
  • Country: us
I'm seeing the same issue with my DP832. For example, with channel 1 off, I measure -0.0036v across channel 1. When I enable channel 1 (set to 9.000v), the DP832 display output reads 8.998V, and I measure 8.997V across the terminals. I see similar behaviour on the other channels, the measured non-zero voltage when channel OFF is always identical to the measured offset error with the channel ON.

The output voltage seems fine overall, but the non-zero voltage with channel OFF seems strange. Anyone with more experience than I know what is going on here? The non-zero voltage (with channel OFF) is only present when the DP832 is on, and it is very consistent. If I turn off the power supply, the voltage goes to zero.
« Last Edit: February 18, 2015, 07:58:20 am by dadler »
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
I don't own this unit but what I heard is that when switched off the unit is programmed to output zero volts rather than to physically switch the output off.  So if not adjusted properly, the output will read some residual output.
Best Wishes,

 Saturation
 

Offline dadler

  • Supporter
  • ****
  • Posts: 851
  • Country: us
That is exactly what it seems to do, no relays click when enabling output. I would think zero/off would be a special case warranting physical disconnection, but maybe that would fight the designer's desires for programmability/dynamic response.

I really wish it would sit at zero potential when off, as I'm not comfortable with -3mV against my circuit while I am mucking around on the breadboard. I find myself unplugging one of the banana jacks. Maybe I will put a physical switch in-line.

The sense output seems accurate-ish. I wonder if calibration would get it closer to zero? I don't have reference equipment to do that properly, though.
 

Offline Chris Wilson

  • Super Contributor
  • ***
  • Posts: 1081
  • Country: gb
  • Race car engineer, dog lover, hoarder.
I was thinking of buying one of these supplies, as you seem to have done some accurate testing I am wondering if there's any overshoot if you look at an output being turned on, when set to X Volts? My Farnell one has played up and one output has a destructive overshoot when it's switched on, I don't want to buy another over shooting supply! Thanks.
Best regards,

                 Chris Wilson.
 

Offline Richard Crowley

  • Super Contributor
  • ***
  • Posts: 4317
  • Country: us
  • KJ7YLK
1) Why is 0.003 v an issue for anything?
2) Were you measuring it open circuit (with a high-impedance meter)?  Or were you measuring it with a nominal load (like your circuit) connected?

IMHO/IME, this is much ado about nothing. (or 0.003 away from nothing)
 

Offline Kevin.D

  • Frequent Contributor
  • **
  • Posts: 290
  • Country: england
Hi there
Do you have a schematic for this rigol supply ? If so post it then we can see what the cause maybe.
This is a fairly common bug you'l find in many cheaper bench supplies (even a few expensive ones, but this doesnt  necessarily mean this is definitely the cause with this rigol supply there is prob an offset adj for it schematic = verify).

Usually when it's present on bench supplies I know it's caused by Aux bias current flowing backward through the load.
Below I have posted a standard simplified bench supply setup. See that aux/ref bias current (usually a few mA) flows through R1 then through the Vset pot then around the main loop returning via the pass transitor to aux gnd. BUT if the pass transistor is turned off  then Aux Bias currents only path is reversed through Rload to AUX gnd.
As you see this reverse current(reverse Voltage will be limited to -.7 V by the reverse protection diode across output) is only going to be a few mA and is only seen when the pass transistor is disabled or the current limit is set lower than the bias leakage current.(Try it on few different bench supplies, set current limit to 0 and you may find a few mA at upto -.7 V reverse voltage going through your load, its usually does not cause a prob in most cases).

More expensive and better designed supplies somtimes avoid this problem by say simply using an extra opamp to sense V out hence no large bias current from Vref is injected directly into main loop.
 
Even when the above is done though Any other source of aux bias current entering the main supply loop will cause the above and there can be other much smaller  bias current sources (< 1 mA) that can cause a tiny reverse current to flow when output pass disabled.

Regards   


« Last Edit: February 19, 2015, 07:54:36 pm by Kevin.D »
 

Offline dadler

  • Supporter
  • ****
  • Posts: 851
  • Country: us
1) Why is 0.003 v an issue for anything?
2) Were you measuring it open circuit (with a high-impedance meter)?  Or were you measuring it with a nominal load (like your circuit) connected?

IMHO/IME, this is much ado about nothing. (or 0.003 away from nothing)

It is much ado about nothing. I just expected hard zero with the outputs off.

I've tried with both my circuit disconnected and connected. Measured potential with my Brymen BM689 and Fluke 287, which measured very similar to the sense output on the Rigol display (with channel on. With channel off the Rigol must disable its ADC update loop and show zero, even though I consistenly measure the -0.003V with my meters.) The negative offset seems identical to the delta between the requested voltage and true sensed output voltage.

The circuit I had on the board at the time was just a power filter circuit and presented no appreciable load. But the non-zero potential was sufficient to partially charge one of my caps (with the channel output OFF). It was also reverse biasing one of my electrolytic caps but at only 0.003V I don't think that matters.

I suppose I would prefer a physical relay disconnection when the channel is OFF. Like my Rigol waveform generator behaves. I'll try to find the DP832 schematic.
« Last Edit: February 20, 2015, 12:35:01 am by dadler »
 

Offline electr_peter

  • Supporter
  • ****
  • Posts: 1302
  • Country: lt
If you want "true zero" with power off, use relays or switches (DPDT for a good measure). Lab power supplies are used in >0.3V or >0.001mA region anyway, minute voltage in off state won't do any damage.
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 28380
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
If you want "true zero" with power off, use relays or switches (DPDT for a good measure). Lab power supplies are used in >0.3V or >0.001mA region anyway, minute voltage in off state won't do any damage.
Or find a better PSU.  :-DD
Avid Rabid Hobbyist
Siglent Youtube channel: https://www.youtube.com/@SiglentVideo/videos
 

Offline TooOldForThis

  • Regular Contributor
  • *
  • Posts: 57
  • Country: us
  • H: 42.576MHz/Tesla
The small voltage offset might be an environmental problem - it was calibrated at one room temperature but it's being used at a significantly different temp.    Since this power supply doesn't have dedicated sense lines, the voltage drop through the external test leads can be far greater than any calibration error.
It would be an interesting project to bring the internal sense lines out to new external terminals so the power supply could correct for the test lead drop.   But things can go badly if the sense lines were ever left open. 
 

Offline mzzj

  • Super Contributor
  • ***
  • Posts: 1245
  • Country: fi

I really wish it would sit at zero potential when off, as I'm not comfortable with -3mV against my circuit while I am mucking around on the breadboard. I find myself unplugging one of the banana jacks. Maybe I will put a physical switch in-line.

Time to expand your comfort zone then  8)

3mV is nothing to worry.
30mV would be still nothing to worry.
300mV would  be kind of ugly but still mostly nothing to worry.
3V would be no good.
300V unconfortable  ;)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf