Author Topic: Testing AC power supplies with a DC electronic load  (Read 4361 times)

0 Members and 1 Guest are viewing this topic.

Offline calzap

  • Regular Contributor
  • *
  • Posts: 202
  • Country: us
Testing AC power supplies with a DC electronic load
« on: April 05, 2013, 09:27:46 pm »
Has anyone tried using a DC electronic load to test AC power supplies?  Seems to me the output of the AC supply could be run through a full wave rectifier and then to the load.  I'm not sure how well DC electronic loads handle pulsed DC and maybe a smoothing capacitor would be necessary.  The scope and meter leads would have to be attached to the output of the AC supply.  What I have in mind is testing relatively low voltage supplies (24 VAC or less) at currents of up to 6 A and am mostly interested in voltage fall-off as current draw is increased.

Comments?  Suggestions for capacitors?

Mike
 

Offline codeboy2k

  • Super Contributor
  • ***
  • Posts: 1838
  • Country: ca
Re: Testing AC power supplies with a DC electronic load
« Reply #1 on: April 07, 2013, 03:26:55 am »
Has anyone tried using a DC electronic load to test AC power supplies?  Seems to me the output of the AC supply could be run through a full wave rectifier and then to the load.  I'm not sure how well DC electronic loads handle pulsed DC and maybe a smoothing capacitor would be necessary.  The scope and meter leads would have to be attached to the output of the AC supply.  What I have in mind is testing relatively low voltage supplies (24 VAC or less) at currents of up to 6 A and am mostly interested in voltage fall-off as current draw is increased.

Comments?  Suggestions for capacitors?

Mike

I did exactly that when I wanted to test transformers to find the point when the voltage is reduced 20% down from it's unloaded voltage.  For me, I consider a transformer has reached it's maximum output voltage and current when the output voltage drops by 20% (from the unloaded voltage). That's its max VA rating (for me)

Other engineers may have 10% or 15% limits, and it may be a spec of your design what the permitted voltage drop is (from the no load voltage)

I used this method to test some unknown transformers and unknown windings on multi-winding transformers that I scavenged.

I put the transformer through a full wave bridge, no smoothing caps, and put that on the DC load.  The DC load has enough bandwidth in it's feedback loop to handle the pulsed DC at 120Hz, that's no problem for it.

In theory, it shouldn't matter.  With or without caps, I think the RMS current at the transformer will read the same when the voltage drops, but I wanted to keep it simple so I didn't have any caps.  Caps add reactance, then current starts to lead voltage, and too much capacitance means the current seen by the load is mostly from the caps, and not the transformer, and the transformer only sees the ESR of the caps for short instants in time... I wanted a purely resistive load.

Using this method, I measure the transformer's output voltage and current using 2 DMM's, then ramp up the current at the DC load until the transformer's output voltage has dropped 20% from its starting point. The AC RMS voltage x the AC RMS current is the VA rating for the transformer.

 

Offline calzap

  • Regular Contributor
  • *
  • Posts: 202
  • Country: us
Re: Testing AC power supplies with a DC electronic load
« Reply #2 on: April 07, 2013, 04:46:38 pm »
Thanks for the reply codeboy.  I think I'll try a little low current testing at different frequencies using a full wave rectifier and  a function generator as the power  source.  Manufacturers of DC electronic loads seem to be silent about pulsed DC in their spec sheets.  My load is a B&K, but I've also checked specs on Agilent and Gossen (aka Array) loads.  Spec sheets sometimes give ripple specs, but I assume these are measures of time variation in the load's ability to suck-up power, not its tolerance for time variation in the input voltage (or current).

If anyone else has additional knowledge or experience in using pulsed DC input to electronic loads, I'd like to hear it.

Mike
 

Offline codeboy2k

  • Super Contributor
  • ***
  • Posts: 1838
  • Country: ca
Re: Testing AC power supplies with a DC electronic load
« Reply #3 on: April 07, 2013, 11:12:11 pm »
Thanks for the reply codeboy.  I think I'll try a little low current testing at different frequencies using a full wave rectifier and  a function generator as the power  source.  Manufacturers of DC electronic loads seem to be silent about pulsed DC in their spec sheets.  My load is a B&K, but I've also checked specs on Agilent and Gossen (aka Array) loads.  Spec sheets sometimes give ripple specs, but I assume these are measures of time variation in the load's ability to suck-up power, not its tolerance for time variation in the input voltage (or current).

If anyone else has additional knowledge or experience in using pulsed DC input to electronic loads, I'd like to hear it.

Mike
You're welcome.

It can handle it. What's the point of a DC load if it can't handle a varying DC input voltage?  The frequency after full wave rectification is only double the line frequency, and that's really really slowwwww for it to respond to.  The bandwidth of the load is probably up in the high kHz.  It will faithfully vary load the load resistance to maintain the set load current.

They don't usually spec their servo loop bandwidth, but I did find one app note that said the BK8510 was measured at a slew rate of 1A/us. So for your 6A CC load, that would need 6us to slew up and 6us to slew down (assuming it's the same) and thus for a 6A load you theoretically can pulse at a 12us cycle time, or 83.333 kHz. So really, pulsed DC at 120Hz is a walk in the park for it.

Jim Williams has an app note where he builds a DC load and feeds it a constant voltage, but the control voltage is 1V p-p varying at 100 kHz (on the control signal, 1V = 100A). He shows a varying 0-100A sine wave load at 100 kHz. with a DC input of 1.5V.

AN133 - A Closed-Loop, Wideband, 100A Active Load

His load is spec'd at
BANDWIDTH = 540kHz @ 100A (tRISE = 650ns)
BANDWIDTH = 435kHz @ 10A   (tRISE = 800ns)

You could just as easily leave the control signal at 1V (setting it to 1A constant current) and vary the DC voltage, it would still respond just as good, because the control signal is on one input of the servo opamp, and the feed back is on the other.  It doesn't matter which one is varied.

 

Offline calzap

  • Regular Contributor
  • *
  • Posts: 202
  • Country: us
Re: Testing AC power supplies with a DC electronic load
« Reply #4 on: April 08, 2013, 03:49:56 am »
Again, thanks for the info; it's very useful.  Makes me wonder why AC loads have a market at all.  Add a full wave rectifier to the inputs of a DC load, and you've got an AC load.  There seem to be a lot fewer makes and models of AC loads than DC loads, and I'm starting to see why.  They also tend to be more expensive.  Is it the limited market or something else?

Mike
 

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: Testing AC power supplies with a DC electronic load
« Reply #5 on: April 08, 2013, 03:08:02 pm »
Hmm, while I have no doubt that you could hook up a DC electronic load via a bridge rectifier (and smoothing cap if required) to an AC source, and that the resulting test rig would do something, I don't think you would be able to tell very much from the results!

First of all, the circuit will not present a uniform load to the power supply, throughout the AC cycle. If you have a capacitor across the Electronic load, then the circuit will only draw current in two sharp "spikes", when the capacitor voiltage drops below the peaks of the AC supply (this always happens in simple AC>DC bridge/Cap power supplies).

Secondly, the Electronic load is software regulated to try and maintain a constant current, voltage, power or even resistance. It's going to struggle to do this if the voltage it is staring at is going up and down a lot, as rectified AC.

Be aware that the active element in a DC electronic load is just a whopping great MosFet (or several MosFets in parallel). These will present themsleves as a forward-biased PN junction if you attempt to feed reverse polarity DC  (or the negative half of an AC waveform) into your load, with perhaps unfortunate results...

A 'scope is going to help you understand what such a circuit is doing. Look at both amps and Volts. What is the "real" load that you are trying to simulate? Any simulated load needs to have the same or similar behaviour throughout the AC cycle as the real load. If your real load is resistive, try the DC electronic load with a bridge rectifier, but no smoothing Cap. It might get you close, but as I say, your Electronic Load may struggle to track the rectified AC.
If you don't measure, you don't get.
 

Offline codeboy2k

  • Super Contributor
  • ***
  • Posts: 1838
  • Country: ca
Re: Testing AC power supplies with a DC electronic load
« Reply #6 on: April 08, 2013, 11:56:54 pm »
Hmm, while I have no doubt that you could hook up a DC electronic load via a bridge rectifier (and smoothing cap if required) to an AC source, and that the resulting test rig would do something, I don't think you would be able to tell very much from the results!
You just described every linear power supply in existence, and linear power supplies are always connected to DC loads. I think what you are saying is that the results won't tell you much about the AC SOURCE, but I think it will, as I described; the trick is to try to keep the load appearing like a resistive load, without filter caps, and measure the voltage drop at the AC source.
Quote
Secondly, the Electronic load is software regulated to try and maintain a constant current, voltage, power or even resistance. It's going to struggle to do this if the voltage it is staring at is going up and down a lot, as rectified AC.
It won't struggle. In any load that I've ever seen, the feedback loop is a hardware loop.  The software just sets the load current for the loop to maintain. The loop will see changes in voltage drops across its sense resistor and servo to change the gate voltage to compensate, and as I've said in my previous posts, the feedback loop has plenty of bandwidth to do this at AC line frequencies.
Quote
Be aware that the active element in a DC electronic load is just a whopping great MosFet (or several MosFets in parallel). These will present themsleves as a forward-biased PN junction if you attempt to feed reverse polarity DC  (or the negative half of an AC waveform) into your load, with perhaps unfortunate results...
I don't think anyone here is advocating applying AC to the DC load. Yes, through a bridge rectifier, which will supply pulsed DC to the DC load.

Although I don't suggest you try it, most DC loads are protected against reverse voltage. In the worst case, probably a fuse blows, in the best case, it's half-wave rectified and appears as DC.  But I don't suggest anyone try this here.

Quote
A 'scope is going to help you understand what such a circuit is doing. Look at both amps and Volts. What is the "real" load that you are trying to simulate? Any simulated load needs to have the same or similar behaviour throughout the AC cycle as the real load. If your real load is resistive, try the DC electronic load with a bridge rectifier, but no smoothing Cap. It might get you close, but as I say, your Electronic Load may struggle to track the rectified AC.

Good point about trying to simulate the actual load that the AC supply is really intended for. Ultimately , that's the goal. If the AC supply is intended for resistive loads, then the electronic DC load is a good substitute. if it is intended to supply reactive loads, then maybe it's not and won't give meaningful results, as you've said.

 

Offline Harvs

  • Super Contributor
  • ***
  • Posts: 1161
  • Country: au
Re: Testing AC power supplies with a DC electronic load
« Reply #7 on: April 09, 2013, 03:25:10 am »
I've been wondering about this, what's the bandwidth in constant resistance mode of these loads? Constant current mode is relatively easy because once the mosfet is in regulation, that is there's enough input voltage for the feedback loop to achieve the set current, the voltage to current gain of the mosfet will take care of a lot of the regulation. Where as I constant resistance mode you're relying on whatever feedback mechanism you've got.
 

Offline calzap

  • Regular Contributor
  • *
  • Posts: 202
  • Country: us
Re: Testing AC power supplies with a DC electronic load
« Reply #8 on: April 10, 2013, 02:26:40 am »
Quote

Quote

    A 'scope is going to help you understand what such a circuit is doing. Look at both amps and Volts. What is the "real" load that you are trying to simulate? Any simulated load needs to have the same or similar behaviour throughout the AC cycle as the real load. If your real load is resistive, try the DC electronic load with a bridge rectifier, but no smoothing Cap. It might get you close, but as I say, your Electronic Load may struggle to track the rectified AC.

End Quote

Good point about trying to simulate the actual load that the AC supply is really intended for. Ultimately , that's the goal. If the AC supply is intended for resistive loads, then the electronic DC load is a good substitute. if it is intended to supply reactive loads, then maybe it's not and won't give meaningful results, as you've said.

The majority of power supplies that I want to test are DC.  There are a few AC units though.  Some are just transformers, like industrial control transformers.  These units are 120 VAC in, 24 VAC out and are in the 50-150 W power category.  They are used to power motor contactors and valve solenoids ... nasty inductive loads with big inrush currents.  I've had a few of these that did not handle the inrush, particularly in summer heat.  If a DC electronic load (sitting behind a full wave rectifier, of course) were to be used to simulate this situation, it would need to go from little or no current draw very quickly to maximum or greater rating of the transformer and then settle down to half-maximum or less.  My B&K can be programmed to do this.
 
There are few AC wall warts that feed older electronic equipment; their output in most cases gets rectified in the equipment.  Then I have a breadboard power supply with DC and AC outputs, all 15 V or less.  Can 't remember last time I used the AC outputs, but would like to test them anyway.

Mike
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf