OCP behavior of KORAD 3005P

[17_29bis]

Hello,

Recently after reading some favorable reviews I decided to get a linear power supply and since this is for hobby, I have chosen KORAD 3005P, paid 200 CAD (i.e not as cheap as it can be found in US),  shipped from Amazon.ca

The PSU is not ideal, includes incorrect  documentation that has not been updated for quite some time (I mean setting of M1-M5 modes), OCP, OVP are turned off by default, occasional strange behavior requiring power cycling etc - the list goes on and those issues are mostly known to KORAD existing users.

But there is one thing that really baffles me - the way OCP works in this PSU.

I must say I am a big fan of OCP mode (and don't care about OVP) . And since I mostly work with low current devices I always keep the OCP limit low and close  the to it realistically consumed/drawn value, i.e. average  + 50%-70%  which gives quite workable numbers. I.e. if a MCU board consumes 50 mA (max) I would set OCP limit to 130 mA and start working on it being sure that in XX cases out of 100 OCP will save the board if I short something. The time has proved this approach to be correct and saved me and the hardware many times.

Having said that I expected the same behavior from KORAD 3005P and what I see is totally different. I connected a small MCU board normally consuming about 40 mA to the KORAD with OCP limit set to 200 mA and when I turned the output on the PSU immediately turned it off  due to the triggered OCP. I increased the OCP limit to 0.4A - no dice, 1A - nope, 1.5A - no bueno, 2.6A - success.  If I turn the OCP off the power supply powers the board without any problem, the drawn current is about 40 mA.

What is going on? Turns out a capacitor (4700 uF)  installed on the + 5V rail of MCU board  starts charging and trips OCP but this is not right.

Eventually, I decided to test PSU using the following simplified test setup (using capacitive load to simulate the small MCU based board):



0.1 ohm is a current shunt resistor allowing us to see the drawn current during output voltage raise time using blue trace (the channel 3 of SDS1204X-E), while the yellow trace (the channel 1) will allow us to monitor the output voltage.


How do I find the OCP limit to be able to power the load with enabled OCP? I will turn the OCP off and measure the current on the shunt. Please pay attention to the initial current spike and max output voltage (10.24 while PSU output is set to 10V, but btw it was not the worst case, I will show worse cases later)



If  the OCP limit is set to less that 2.5A  and OCP is on then the PSU will constantly  trip overcurrent protection and I find it very strange.

My conclusion regarding the work of OCP in this particular PSU is this: if during the output voltage raise time the drawn current exceeds the OCP  limit then OCP will be engaged. It seems to be wrong due to multiple reasons and one of them is because some loads (not necessarily capacitive ones) may drawn more current then underpowered (i.e. the output voltage did not reach the nominal value yet).

I have a different PSU (DPS 5005) with OCP and OVP support  and the OCP is implemented differently - during the output voltage raise it will LIMIT the current to the OCP set limit and when the output voltage reaches its programmed value then the OCP will start monitoring the drawn current and turn the output off in case if the limit is exceeded.



This is how it works when OCP limit is set to something like 50 mA - it has no problem with starting up with OCP enabled.




*************************************************************************************************************
At this point I have two questions:

a)  Could someone who has experience with PSU from known brands  tell me how OCP behaves in their PSU during voltage raise time?

b) Could someone who has KORAD 3005 make a test by simply connecting some sort of capacitive load (a 1000 uF capacitor  would be enough) and check the behavior of their unit?  May be I simply have a defected PSU and should send it back for replacement.

[17_29bis]

And 2 more screenshots showing the voltage spikes while the output voltage is set to 10V (OCP and OVP are off)






**********************************

If this is the normal/typical behavior of KORAD 3005 then I simply don't understand how people call it "inexpensive and decent linear lab. power supply". Inexpensive - yes (although 200$ is 200$), linear - yes, but decent - I don't know.

[SenseofScale]

My ka3005p won't even turn the unloaded output on reliably, when set below 155 mA, with OCP turned on.

[17_29bis]

My ka3005p won't even turn the unloaded output on reliably, when set below 155 mA, with OCP turned on.

Yes, I saw this exact problem being mentioned in the internet. This behavior makes OCP pretty useless for low currents.  Rumor has it that it was fixed in the latest versions of KORAD PSU and as a matter of fact my PSU turns on with the output current limited to even smaller values with OCP turned on just fine but only for resistive/resistance loads.

[SenseofScale]

So, to follow up with your other question, I did the test you asked for about fifteen times and didn't see any of the spikes you found. They're pretty brief, so I'm not sure if your scope is possibly picking up something that mine missed.

Attached are a couple images that I captured.

Used a 4700 uF and a .5 ohm resistance, in the same configuration as yours, minus the bleeder resistor.

I want to add that it is indeed a junk supply. I had to repair mine on the first day. When mine was assembled a heat sink was piercing one of the mains wires, so I was measuring something like 60 V to ground from the outputs. The colors were miss-matched on the output cables (red plug terminated to a black alligator clip) so I blew the anti-parallel diode attempting to charge a battery with it. The latter is partly my fault as I should have noticed the cables but it should give you an indication of the quality you're dealing with.

On the other hand, I haven't had trouble with it since that day about 4 years ago.

[SenseofScale]

I just took several measurements with a single channel in order to up the sample rate. I also took the .5 ohm resistance out of the mix and went directly across the cap and brought the time scale down far enough to achieve 1 Gs/s and set the trigger level 1 V above the output setpoint and it never caught anything. I also tried various voltages.

I guess there is something wrong with your supply?

Edit: I put the resistor back in and did a bunch of measurements again. I wasn't really thinking about how going directly across the cap was going to suppress the spikes. I still didn't measure any spikes. I only recently dragged all my toys out again and my understanding never got to be very good but I'm also rusty.

[17_29bis]

Hello SenseofScale,

First of all thank you for your tests.

I guess there is something wrong with your supply?

I gave away the PSU as a gift so I don't have it anymore. I thought  about sending it back for replacement  but I guess it would be difficult to explain what the problem is and the total time I spend to get the unit replaced (if ever) would not worth it.


Those current spikes could be seen only in case of a capacitive load and you would not get them in case of a pure resistive load.
I think it can be explained by the current charging the output capacitor while ignoring (for a very short time) the current limit set on the PSU.   If you cannot reproduce/see  those spikes then most likely our PSUs are slightly different - my PSU does not have a problem (like yours) with turning on when the OCP is on and the current limit is set below 150 mA while yours does not have those current spikes in case of a capacitive load and OCP works a bit differently.

I had to repair mine on the first day. When mine was assembled a heat sink was piercing one of the mains wires, so I was measuring something like 60 V to ground from the outputs.

That speaks volumes.

One more time - thanks for the time you spent for tests,  all that  just  confirmed my suspicions about this particular PSU.