Comparing two power supplies: Extech 382260 and TTi EL302P

[IanB]

So having recently looked at the control response of my TTi EL302P power supply, I wondered how my Extech 382260 would compare. The answer: not very well.

The test I performed was to close the switch on a 12 V 12 W incandescent bulb. This bulb has a filament resistance of about 12 ohms at operating temperature, but only a fraction of that when cold. So when you apply power to the bulb there is a large surge current before the filament warms up. If you set a current limit on the power supply it should catch the surge and start the bulb up slowly.

For all of the tests I had the power supply set to regulate at 8.0 V and at 1.0 A. There was a 1 ohm resistor in series with the bulb to act as a current sense element. I measured the responses using my Rigol DS1102E oscilloscope, exported the data to CSV and plotted in Excel.

I ran two tests with each power supply. In the first test I closed an external switch on the bulb with the power supply already on. This measures how quickly the supply can catch the surge current and bring it under control. In the second test I used the output on/off switch on the power supply itself. This gives the power supply the best opportunity to regulate current and voltage within the set limits since it has full servo control with no external disturbances.

The first test is the EL302P with the external switch. There is a momentary current spike (as the output capacitor discharges?), but then the internal regulation circuit cuts in and controls the current within microseconds.



Next, the Extech with the same test under exactly the same conditions. We see here a very poor result. The switch is closed at marker 0 on the chart. There is a large surge in current (output capacitor discharging?), and for 50 ms or so the power supply does not react while a large excess current is dumped into the load. At marker 1 the power supply apparently takes notice and starts to reduce the output voltage to control the current. But the control algorithm is evidently quite slow and it takes another 50 ms until marker 2 before the output current is at the set point.

With the current regulated the voltage climbs until it reaches 8 V when the power supply switches back to CV mode. However there is an overshoot of about 0.3 V before the voltage comes back to the set value.



In test 3, the EL302P is used with its internal output switch instead of the external switch. We see how the voltage and current climb up rapidly from zero, but then the current is caught as soon as it reaches the set value and now the voltage climbs slowly as the bulb warms up. Everything looks good with little overshoot and no instability.



In the last test, the Extech is also used with its internal output switch, duplicating test 3. Here we see horrors. The current and voltage climb rapidly from zero, but then the current wildly overshoots the set value and keeps climbing. Not only that, but instability is observed in the regulation (the spikiness is from the power supply, it is not digital noise from the scope). It takes about 30 ms before the power supply catches the current and starts to bring it down again, and almost 50 ms from the start before the current is fully controlled. (When I ran the same test with a current limit at 1.5 A the supply didn't catch the overshoot at all and never  tried to limit the current.)



Conclusion: the Extech 382260 is really not a good performer if you need fast and accurate regulation. I imagine this is a limitation of the switch mode configuration. However, if you only need to power mechanical devices or you can live with response times measured in milliseconds instead of microseconds then the Extech is a small, light and convenient power supply.

[LaurenceW]

EEK! These overcurrent surges would be enough to fry many electronic circuits!

I have an earlier TTI PL330TP which performs similary to your results, but I reckon the VERY short current spike is more than the output cap discharging, There must be at least some delay (uS or less) while the analogue current feedback loop regains control.   

Is it possible that the Extech is trying to do current and/or voltage feedback in software? Such "apparently" hi-tech solutions are usually MUCH slower and so poorer than direct analogue feedback. But hey, software control must be better, right? WRONG

[codeboy2k]

You can make a switcher that responds fast.  More likely they didn't want to be bothered doing the proper loop response analysis.  The hardware can be expensive and often out of reach.  Too many power supplies are designed and thrown over the wall without doing a proper FRA.. The engineer often gets a poor load-step response  and has to intentionally slow down the loop by adding more capacitance, just willy-nilly until it doesn't oscillate.  Then it's shipped.