Inrush Current Simulation

[DPG4]

I have a box with ~50 loads that it drives, of which consist of lights, solenoids and some pretty hefty hydraulic motors (ranging from 5W to 4.5kW).
To test this box and put it through its paces (nominal and over current conditions), I am charged with developing a load box that will simulate these loads.

With that in mind, I am curious if there is an elegant solution for this, I have taken a look at electronic loads, but was not really impressed with their response time to produce the desired inrush, and their cost was more then expected as well.

Any suggestions?

 

[itsmoked]

For starters, I'd make sure all the various items work during your prototyping mode. Then to test production units I would only worry about the full loads.
For that, noting beats light bulbs. They instantly show what's happening. They also have an inductive component that is pretty severe and will 'test' your unit. For the larger outputs put a bulb in parallel with some sort of heater element which can be anything up to cheap space heaters, that thoughtfully, provide their own cooling fans.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[DPG4]

Thanks for the input!

The light bulbs are a great idea as long as I take into consideration the temperature of the light bulb. As the inductive kick from the light bulbs is a artifact of its temperature, right? So once the bulb is warm it shouldn't produce much of an inductive spike, and I would have to wait for the bulb to cool before I could perform another test.

And as far the heater, great idea. I may have to modify for my application, a tid bit, but its a great start.

Your suggestions lead me to what I was toying around with, simply putting an inductor and a capacitor network in parallel with some power resistors (basically what you had suggested, but with discrete components). But before I go ahead and attempt to design the capacitor/inductor network, does anybody out there have a circuit, suggestions or device that would provide this kind of a inrush current.

 

[waross]

Not much inductive kick from an incandescent bulb. The inrush is due to the low resistance when it is cold. For the purposes of inrush, no light = cold.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[bogeyman]

The incandescent lamp will draw up to 10 times its normal running current at switch on, measure its cold resistance and compare that to its calculated running value to get a more accurate number. The construction of the lamp will affect the time it needs to warm up and gain resistance.

 

[DPG4]

Yea, bogeyman, that is my point.

If I am to construct this test set up it should be repeatable, producing similar inrush currents each time (if possible).

So I will leave it open again for anyone, is there a common circuit configuration that would be able to reproduce this type of inrush current and be paired with a power resistor (for example) to dissipate a desired load. (not just for an incandescent bulb either, but rather a general circuit that could produce this current, and varying current magnitudes).

I was thinking something like what is shown here (

http://files.engineering.com/getfil...0-4600-b280-0e15792daaf1&file=InrushModel.pdf

) from power-one (basically a resistor in series with an inductor and then in series with a capacitor which would be in parallel with the load, see diagram if there is confusion).

Please comment, I think I am on the right track, but still potentially missing some critical details.

Thanks

 

[itsmoked]

That Power One thing is for DC, AC if you want some really huge components.

I really think you need to reassess this whole test scheme. From your description this tester could easily end up costing $100k. Is that what your company really needs?

Generally detailed results are only needed during design proving. You do those individually at the engineering level. Once the product is correctly designed you need a production tester that is physically tough, useful, simple, relatively self explanatory, qualitative, and affordable.

The tester you appear to be seeking is looking like a quantitative monster that will not meet many the goals above.

If you are testing something like an iPod, you don't! The customer does when they pull it out of the box. Doesn't work? They'll replace it.

A heart pacemaker? You do a full quantitative test process that could cost more the the unit itself.

Everything else, is somewhere in between.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[DPG4]

Thanks for the input, I think I am lying some where in the middle.

The original quote I got for the system using electronic loads was well over that, and hence why I am looking into other options.

Another problem I am facing that is not of a huge technical issue but rather one of economy is that I have 49 of these loads to drive (potentially at the same time). So the module that is simulating each load has to be relatively simple to operate (an electronic load would be nice, as they have a simple lab view type interface, but are extremely expensive; and a simpler rheostat/power resistor (+inrush current circuitry) combination could be cheaper but possibly more labor intensive as there is not a good computer interface for it).

So right now, I am trying to get a good idea of the options and I appreciate the input I have gotten so far.

 

[waross]

Step 1>
Forget the computer. Use a PLC. It has real world inputs and outputs built in. Many have communication capability and may be linked to a computer.
Thanks to Keith;

For the larger outputs put a bulb in parallel with some sort of heater element which can be anything up to cheap space heaters, that thoughtfully, provide their own cooling fans.

You can add a variac ahead of the space heater(s) to vary the load. You can calculate the inrush and preset the unit for production setting. If it must be set on the fly, let your computer do the calculations and send the switching instructions to the PLC.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Lakey]

This may be a bit late in the day, but you could try using a series capacitor to simulate the instantaneous current of the pump-motor (DC series wound). The amplitude of the inrush can be controlled with a resistor, i.e. pre-charged capacitor. This may save a few bucks.

Regards,

 

[DPG4]

Lakey,

Thanks, I had a similar idea, and a good thing about that configuration is that the leakage path would be way low so that upon removing power from the circuit, the charge from the capacitor would be drained off quite rapidly, which would providing a "refreshed" circuit quickly after power had been applied. The one thing about this is that I am cautious when using capacitors bigger then the size of my fist I am a little hesitant(10000 uf or around there to get a adequate spike, with a hefty safety factor for the voltage puts me at quite the large cap).

I want to thank everybody for their thoughts.

Cheers

DPG