This whole issue is entirely cause by the power supply not offering "soft-start" functionality. It takes time to transition from constant voltage to constant current and I would bet that just about any power supply would behave in similar fashion - have an initial overshoot to some degree, before transitioning into CC mode and dropping the voltage to the necessary level to maintain that current.
I guess you're right about the soft-start stuff. That would probably alleviate the problem. However, the fact that we have a transition here from CV to CC mode in the first place is what irks me. If you have a look at the design of classic analog power supplies, you'll find an op-amp that does the CV regulation and an op-amp that does the CC regulation. The "minimum" of both op-amps' outputs is taken to drive the output transistors (there are several possible circuits there). There is no hard transition between CV and CC.
@@robertssmorgasbord In an analog circuit, comparators would be employed resulting in minimal latency of the transition and little to no overshoot as a result. Chances are that this power supply is using some sort of digital circuitry which is reading a current shunt further amplified by an op-amp and most likely converted from analog to digital via an ADC of some sort, to see whether the current setting is exceeded and then, after all of those propagation/calculation delays, finally does the switch from CV to CC. These significantly longer delays, compared to what would be present in an analog circuit with very low latency modern comparators, result in the huge if brief voltage overshoot that you measured.
@@SteelBlueVision Exactly, an analog CV/CC regulation would be much faster with little to no overshoot. And you're probably right about the design of that power supply. My questions/complaint is why they've done it this way in the first place. I know, it's a bit cheaper. But having a MCU with two ADCs (just for displaying actual voltage/current) and two DACs to set voltage/current in an analog regulation circuit is not really that much more expensive. You'll probably need a few more op-amps and passives, but that's it.
I wonder if this is a feature of the start routine settling, or a reaction to a change of load? Could you turn on the Owon's output, let it settle, and then add the load and see what it does?
Well, adding the load via a mechanical switch or plugging in a banana plug will definitely add a whole lot of bounce. If done right the load needs to be switched on electronically, e.g. via a MOSFET. Maybe I'll do that ... in the future. For now I've made enough videos about that OWON ;-) Anyway, I'll put it on my wall of video ideas ...
@@robertssmorgasbord Hi Robert, first of all thank you very much for the video. I've ordered the Owon and I'm going to do a few tests. Since I also have an electrical load, I could take up Steve's idea.
I'm probably very late to the party and you might have already resolved this problem but did you ever use a current probe to measure the actual current during turn-on? If the PSU is set to regulate current, surely you should be measuring current?
Well, if I remember right I used the scope to measure the voltage drop across a shunt resistor when doing current measurements. So I did measure the current.
First of all, thanks for the praise! And to answer your question: Not really. It was one of those glitches you couldn't consistently reproduce. Currently the display still works fine.
@@nutsinstuffprod Yes, I'm still using the Owon as my main powers supply (my other one is a LM317 based DIY without current regulation - so no contest).
The noise you see in this video is external EM induced into all the long unshielded cables I'm using. Have a look at part 1 ( ua-cam.com/video/IbGax79w-RQ/v-deo.html ) where I do some proper ripple and noise measurements using shielded cables.
Ah, do you mean I mistakenly cut the power supply some slack because of the low cold resistance of the halogen light bulbs. Maybe so, but it is easier to regulate a current through a constant resistance than through a changing resistance. So I can't really see how the low cold resistance would help the power supply do its job.
This whole issue is entirely cause by the power supply not offering "soft-start" functionality. It takes time to transition from constant voltage to constant current and I would bet that just about any power supply would behave in similar fashion - have an initial overshoot to some degree, before transitioning into CC mode and dropping the voltage to the necessary level to maintain that current.
I guess you're right about the soft-start stuff. That would probably alleviate the problem. However, the fact that we have a transition here from CV to CC mode in the first place is what irks me. If you have a look at the design of classic analog power supplies, you'll find an op-amp that does the CV regulation and an op-amp that does the CC regulation. The "minimum" of both op-amps' outputs is taken to drive the output transistors (there are several possible circuits there). There is no hard transition between CV and CC.
@@robertssmorgasbord In an analog circuit, comparators would be employed resulting in minimal latency of the transition and little to no overshoot as a result.
Chances are that this power supply is using some sort of digital circuitry which is reading a current shunt further amplified by an op-amp and most likely converted from analog to digital via an ADC of some sort, to see whether the current setting is exceeded and then, after all of those propagation/calculation delays, finally does the switch from CV to CC.
These significantly longer delays, compared to what would be present in an analog circuit with very low latency modern comparators, result in the huge if brief voltage overshoot that you measured.
@@SteelBlueVision Exactly, an analog CV/CC regulation would be much faster with little to no overshoot. And you're probably right about the design of that power supply. My questions/complaint is why they've done it this way in the first place. I know, it's a bit cheaper. But having a MCU with two ADCs (just for displaying actual voltage/current) and two DACs to set voltage/current in an analog regulation circuit is not really that much more expensive. You'll probably need a few more op-amps and passives, but that's it.
Hello, would you recommend this device as a result?
Keeping the large switch-on transients under high loads in mind, assuming that the problem with my unit was an exception, and the price: Yes.
I wonder if this is a feature of the start routine settling, or a reaction to a change of load?
Could you turn on the Owon's output, let it settle, and then add the load and see what it does?
Well, adding the load via a mechanical switch or plugging in a banana plug will definitely add a whole lot of bounce. If done right the load needs to be switched on electronically, e.g. via a MOSFET. Maybe I'll do that ... in the future. For now I've made enough videos about that OWON ;-) Anyway, I'll put it on my wall of video ideas ...
@@robertssmorgasbord
Hi Robert, first of all thank you very much for the video.
I've ordered the Owon and I'm going to do a few tests. Since I also have an electrical load, I could take up Steve's idea.
@@harrystopic2112 Hi Harry, first you're welcome of course. It would be great if you could do these test. Please let me know how they went.
I'm probably very late to the party and you might have already resolved this problem but did you ever use a current probe to measure the actual current during turn-on? If the PSU is set to regulate current, surely you should be measuring current?
Well, if I remember right I used the scope to measure the voltage drop across a shunt resistor when doing current measurements. So I did measure the current.
@@robertssmorgasbord and it spiked in the same way the voltage did at turn-on?
@@Chris-hy6jy Yup, see "OWON P4305 Lab Power Supply Review (1): First Tests, Failure & Teardown" 16'36'' ua-cam.com/video/IbGax79w-RQ/v-deo.html .
Just wondering if you ever figured out what was wrong with the display. I’m thinking of buying one. Really good review by the way!
First of all, thanks for the praise! And to answer your question: Not really. It was one of those glitches you couldn't consistently reproduce. Currently the display still works fine.
@@robertssmorgasbord Great! Do you still use the Owon as your main power supply, or have you replaced it with something else?
@@nutsinstuffprod Yes, I'm still using the Owon as my main powers supply (my other one is a LM317 based DIY without current regulation - so no contest).
Thanks! Ordered one today.
@@nutsinstuffprod You're welcome! Have fun with it! And better luck with your one than I got with mine initially 🙂
That PSU looks to be very noisy going by your scope trace. Even at 10v the output is super noisy.
The noise you see in this video is external EM induced into all the long unshielded cables I'm using. Have a look at part 1 ( ua-cam.com/video/IbGax79w-RQ/v-deo.html ) where I do some proper ripple and noise measurements using shielded cables.
Your critic is mistaken, the low start R of globes will actually help the power supply!
Ah, do you mean I mistakenly cut the power supply some slack because of the low cold resistance of the halogen light bulbs. Maybe so, but it is easier to regulate a current through a constant resistance than through a changing resistance. So I can't really see how the low cold resistance would help the power supply do its job.
yay
:-)
ALL power supplies go CV & CC at the SAME time, ...this like vast majority of supplies has CRAP CC control loop.
You're right. I guess you've got to spend a lot more dough on a power supply with a decent CC mode.
Yep, still not great but interesting stuff nevertheless....cheers.
Interesting? Yes, very! But in a very bad way. That's not how you implement a constant current mode :-(