Sorry for the direct follow-up, but I just had to do some more testing of the load and unfortunately - though not unexpected - things get even worse:
I had a closer look at the dynamic behavior of CV, CP and CR modes. This time I also did some more measurements on my Maynuo M9712 load for comparison. My setup is basically the same that I used in my previous contribution, I only wired the voltage probe so it displays the voltage across the load this time.
So let´s start with the Maynuo which more or less performs as expected:
Here's the CV mode at 50Hz external switching frequency, 18R series resistance with the power supply:
And the same thing at a different timebase setting:
The voltage settles at somewhere between 2.5ms and 4ms. Not particularly fast but that´s probably what one could expect from an "el-cheapo" electronic load. Usable at least...
I then changed the load to constant power mode (CP) at 10W, 50Hz, once again two different timebase settings:
The behavior is not perfect, it takes the load some 350µs to recognize that there´s a voltage present and to cut down the current/raise the resistance of the passbank but then it settles pretty quickly at the expected figures. Once again, no real reason to complain considering the price.
Finally the last test with the Maynuo in CR mode, 20Ohms selected and once again externally cycled at 50Hz:
What else can I wish for - the proper current is flowing 600µs after the external voltage has been applied. Virtually no droop or whatever can be observed. Well done, Maynuo!
After these tests, the DL3021 was due, and what should I say -- in some cases, I cannot even interpret my findings. In general it seems that Rigol ATTEMPTED some "adaptive" parameter matching but I would say there´s still a considerable distance of road for Rigol to travel in order to come up with a useful result. I tried to duplicate the tests I did with the Maynuo M9712 load before but it turned out that in most cases, the frequency setting of the externally applied power that I used for the Maynuo was just too much for the DL3021. So here we go, first test CV mode @ 10V, 18Ohms in series, 1Hz at two timebase settings:
The load needs some 200ms to settle within 5% of the selected voltage. It´s quite curious that the slope doesn´t seem to be an expenential decay but it rather seems to take much longer in the end to actually reach the preset voltage level. The 1Hz period of the externally applied signal doesn´t suffice to achieve that.
During the next test, things turned out really awkward -- in CP mode (10W), I found the load not to reach any steady condition if I toggled the external supply with 50Hz:
The load kept on modulating the power drawn at a slow pace. It seems the internal "optimization" routine was completely overtaxed. But things turned out even worse in the consequent test: I changed the external toggle frequency to 1Hz in order to permit more time to for the load to catch up -- which may not have been a good idea: After several seconds of operation in a rather unstable manner, the load shut itself down with an error message similar to "Overcurrent detected, input shut down" or the like (don´t remeber the phrase exactly). At least here´s a screen shot of the things going on:
Hint: Don´t take the current displayed before shutdown to be correct -- it´s well possible that my current probe was satuated by the current drawn by the load that was much higher than that. This may also correlate with the warning message on the screen of the load.
Finally, here´s an enlarged sceen shot at the load´s input. Why did it change the current to 1.6Amps after it adjusted to the proper current (0.5A) initially?
And here we are with Rigol´s CR mode, crap as well... (sorry, tried to stay neutral but getting disappointed more an more...
). Load adjusted to 20 Ohms and externally toggled at 1Hz. Funny thing is the current "approximates" to (almost) the final value of 1A over several periods (something like 25...50) and then suddenly settles at the proper value and stays there (as far as I can tell) as long as the load isn´t switched off:
Funny thing, at 50Hz toggle frequency this won´t work -- the modulation continues until I was tired to wait any more:
And the final test @ 10Hz showed the settling again, yet with considerable overshoot in between. It also took the load a considerable time to reach that "steady-state" that´s visible at the right of the screenshot:
If you now ask me about my opinion regarding this instrument, I'ld put it like that:
If Rigol was a car manufacturer that´s only half-way respecting themselves, they should call back ALL of their DL3021 (probably all of the DL3000 series) loads sold so far and refund their customers in full, regardless of their distributor´s offerings. This product isn´t up to meet the market yet. If Rigol doesn´t act like that, flush them down the drain. They have got quite some acceptable equipment but this load is not ripe to be supplied to the public. It´s a shame to have their customers doing the job their "alpha-testing department" should have done. I sincerely hope someone who´s got some authority at Rigol reads these lines and provokes corrective actions. I´m almost certain that the shortcomings of the load cannot be all corrected by a software update but rather need a revised hardware.
I defintely recommend not to spend the money on that load - the internal features don´t differ too much from the aesthetics of the front panel -- unfortunately!
Cheers,
Thomas