Lab Power Supply Turn ON and OFF Characteristics

[mawyatt]

Since there's been some discussions on various Lab Power Supply Turn ON and OFF misbehavior, thought a thread dedicated to such might be in order.

https://www.eevblog.com/forum/testgear/power-supply-for-home-lab-do-i-really-need-a-rs/

First off, think creating a setup and parameters that can be utilized for comparisons is necessary, so we'll start off with a proposed setup for evaluating the Turn ON and OFF characteristics.

1 ) Use 50Ω Load Resistor (usually available to everyone as a discrete resistor, or BNC termination type, or directly within the DSO).

2 ) Use General Purpose 10X DSO probe.

3 ) Use 20MHz DSO Bandwidth limit.

4 ) Set Power Supply Voltage to 8V.

5 ) Set Power Supply Current Limit to 40mA.

6 ) Set DSO to 0.5V/Div for full scale with 4V.

7 ) Trigger DSO on Rising Edge Single Shot for Turn ON Behavior (Falling Edge for Turn OFF Behavior)

8 ) Use Power Supply Enable ON/OFF to activate the output.

9 ) Capture DSO waveforms at various Time Scales starting with 50ms/Div, add additional Time Scales and identify such to show more details such as fast initial glitch (100us/Div) and slow settling (200ms/Div). 

10 ) Annotate waveform displays with details.

11 ) Main purpose is to show Voltage/Current Overshoot (if any) at Lab Supply Turn ON and Turn OFF.

Voltage/Currrent overshoot at Turn ON and/or OFF is our main concern with a Lab Supply, others may have additional concerns, but for now let's stay focused on the Turn ON/OFF Supply behavior.

Hopefully this will yield a means to compare various Power Supply characteristics with a more uniform test procedure/parameter setup.

Anyway, comments, recommendations, additions welcome with goal to yield a set of expert agreed upon setup and test parameters to allow various Lab PS evaluations.

Best

[Kleinstein]

With a 8 V supply  and possibly higher it is not a good idea to use the scope internal 50 ohms. Depending on the scope this may already be damaging from overload, especially over longer time.

The transients from a lab supply are usually slow enough that the exact probing ( 1:1 or 1:10 probe or Z0 probe) does not matter. The load may make a small difference.

8 V and 50 ohm would be 160 mA. So a setting to 40 mA would have the supply in CC mode.
So one would more tend towards a high load resistor (e.g. 100 or 220 ohm) and a slightly higher current limit like 100 mA or 500 mA.
Not so sure why 8 V. More relevant may be 5 V or 3.3 V as a supply to quite some low voltage circuits that may care about an overshoot.

[nctnico]

With a 8 V supply  and possibly higher it is not a good idea to use the scope internal 50 ohms. Depending on the scope this may already be damaging from overload, especially over longer time.

Agreed. For simplicity, a 47 Ohm resistor might be a better choice as this is a standard E12 series value

[Fungus]

With a 8 V supply  and possibly higher it is not a good idea to use the scope internal 50 ohms. Depending on the scope this may already be damaging from overload, especially over longer time.

"5) Power Supply Current Limit to 40mA"

But yeah, a 47 Ohm resister seems more sensible.

[Fungus]

So one would more tend towards a high load resistor (e.g. 100 or 220 ohm) and a slightly higher current limit like 100 mA or 500 mA.

This, too.

[nctnico]

8 V and 50 ohm would be 160 mA. So a setting to 40 mA would have the supply in CC mode.

This is exactly the purpose of this thread. See the last few pages from this thread: https://www.eevblog.com/forum/testgear/power-supply-for-home-lab-do-i-really-need-a-rs/ for the discussion leading to this more generic thread. It turns out quite a few lab power supplies (including those from A-brands), don't start in CC mode properly and have a huge current spike when the output is enabled. This huge spike can destroy sensitive components when wanting to use the supply as a current source instead of a voltage source.

[mawyatt]

^ Exactly!!!

47Ω, 50Ω doesn't really matter.

DSO internal 50Ω should be fine as Current Limit is 40mA and DSO should be able to handle short period over current issue we are looking for

Best,

[Aldo22]

Is that correct?
I asked a very similar question today about the DPS-150 so I had the test setup more or less ready.
I knew the Longwei PDS-3010G was really bad in this regard.
Here is the proof. 

47Ω resistor.
The DMM is there to display the 40mA (when it has settled), the LW can't display such small currents.
Osc. Acq. mode is peak, therefore somewhat bold.

[mawyatt]

8 V and 50 ohm would be 160 mA. So a setting to 40 mA would have the supply in CC mode.

This is exactly the purpose of this thread. See the last few pages from this thread: https://www.eevblog.com/forum/testgear/power-supply-for-home-lab-do-i-really-need-a-rs/ for the discussion leading to this more generic thread. It turns out quite a few lab power supplies (including those from A-brands), don't start in CC mode properly and have a huge current spike when the output is enabled. This huge spike can destroy sensitive components when wanting to use the supply as a current source instead of a voltage source.

A good example of this effect is when one is using an expensive IC and somehow gets the supply leads reversed, or the part plugged into the protoboard upside down, or the prototype wiring messed up, and so on: We've all been-there-done-that

Assuming one has set the Current Limit to a respectable value (always a good idea), this mishap hopefully would allow the expensive chip to survive. However, the high current overshoot on Power Supply Turn ON could destroy or severely damage the device, of course as Murphy would have it this will be the most expensive chip, not a LM358, or LM7805, but the ADC or DAC or Reference

Best

[Furna]

Is that correct?
I asked a very similar question today about the DPS-150 so I had the test setup more or less ready.
I knew the Longwei PDS-3010G was really bad in this regard.
Here is the proof. 

47Ω resistor.
The DMM is there to display the 40mA (when it has settled), the LW can't display such small currents.

Is the multimeter connected to the circuit some how?
You do not want anything in between ... especially in series mode as a multimeter measuring A.
It is altering the circuit.

[Aldo22]

Is the multimeter connected to the circuit some how?
You do not want anything in between ... especially in series mode as a multimeter measuring A.
It is altering the circuit.

Yes, there are 400mΩ extra so I can see what I'm doing. 
That doesn't make much difference in this case.

[mawyatt]

Is that correct?
I asked a very similar question today about the DPS-150 so I had the test setup more or less ready.
I knew the Longwei PDS-3010G was really bad in this regard.
Here is the proof. 

47Ω resistor.
The DMM is there to display the 40mA (when it has settled), the LW can't display such small currents.
Osc. Acq. mode is peak, therefore somewhat bold.

Not sure how the interpret the scales on your DSO, but this doesn't look good!! Seems the supply takes ~ 6 divisions to return to the Current Limit 40mA set point. If that's 500ms/div then 3 seconds!! Even if the entire span is 500ms, then ~200ms which is quite long.

Best,

[Aldo22]

Not sure how the interpret the scales on your DSO, but this doesn't look good!! Seems the supply takes ~ 6 divisions to return to the Current Limit 40mA set point. If that's 500ms/div then 3 seconds!! Even if the entire span is 500ms, then ~200ms which is quite long.

The scope does not lie.
It takes this long with 47Ω.
It is faster with a “short circuit”.

[Furna]

Is the multimeter connected to the circuit some how?
You do not want anything in between ... especially in series mode as a multimeter measuring A.
It is altering the circuit.

Yes, there are 400mΩ extra so I can see what I'm doing. 
That doesn't make much difference in this case.

You do not know what kind of resistor the multimeter has ... you are mixing type of resistors ... resistors have inductance.
Could you please repeat the test without the DMM in series?

[Aldo22]

Could you please repeat the test without the DMM in series?

Yes, it's a bit different.
But the overall impression of this PSU doesn't really change.

[mawyatt]

Not sure how the interpret the scales on your DSO, but this doesn't look good!! Seems the supply takes ~ 6 divisions to return to the Current Limit 40mA set point. If that's 500ms/div then 3 seconds!! Even if the entire span is 500ms, then ~200ms which is quite long.

The scope does not lie.
It takes this long with 47Ω.
It is faster with a “short circuit”.

What is the time scale, is it 500ms/Div?

That looks like ~160mA for ~6 horizontal divisions!!!

Best

[Aldo22]

What is the time scale, is it 500ms/Div?

Yes. (Top: Hantek -> Stop -> H -> 500ms)

That looks like ~160mA for ~6 horizontal divisions!!!

Yes. The DMM shows a Max of 174mA

[Furna]

Yes, it's a bit different.

... and the 100µs/Div?

[mahi]

6 ) Set DSO to 0.5V/Div for full scale with 4V.

Shouldn't we take 1V/div because we already know certain power supplies will go up to 8V for a brief moment. Or don't we care about same scales for everyone?

[mawyatt]

What is the time scale, is it 500ms/Div?

Yes. (Top: Hantek -> Stop -> H -> 500ms)

So ~160mA with 40mA Current Limit for 3 seconds

Best

[Aldo22]

So ~160mA with 40mA Current Limit for 3 seconds

Yes. The DMM shows a Max of 174mA.

It's all about the measurement. There was no question that this PSU is bad in this respect.

That's why I asked today whether the DPS-150 is better.

[mawyatt]

Here's some measurements with 50Ω and Voltage set to 8V and Current Limit set to 40mA.

#64 Siglent CH1 3303X @ 50ms/div
#65 Siglent CH1 3303X @ 100us/div
#66 Siglent CH2 3303X @ 50ms/div
#67 Siglent CH2 3303X @ 100us/div

#68 Instek CH1 GPP4323 @ 50ms/div
#69 Instek CH2 GPP4323 @ 50ms/div

#73 Korad KWR102 @ 50ms/div

#74 Korad KWR103 @ 50ms/div

BTW the Korad KWR102 is a SMPS with 30V/30A capability, and KWR103 is SMPS with 60V/15A capability.

Best

[mawyatt]

Here's the Turn OFF characteristics, all 50ms/div.

#75 Siglent 3303X CH1

#76 Instek GPP 4323 CH1

#77 Korad 102

#78 Korad 103

Best

[Furna]

50Ω wire wound resistor; Voltage set to 8V and Current Limit set to 40mA.

#31 Korad KA3005P @50ms/div PowerON
#32 Korad KA3005P @50ms/div PoserOFF

No "overshoot" but voltage stay at 2.09V => 0.0418A (multimeter confirms) while PSU display shows 2.07V and 0.039A.
Anyway for the price seems tollerable to me.
@mawyatt could you add measurments to your caputure? Are the KWR102/KWR103 also "lying"?

Why I insist on specifing wire wound? Because I am putting the PSU in the worst scenario.

[mawyatt]

@mawyatt could you add measurments to your caputure? Are the KWR102/KWR103 also "lying"?

KWR102 reads 0.039A when set to 0.040A and KS34465A DMM reads 42.042mA, KWR 103 reads 0.039A when set to 0.040A and DMM reads 42.03mA.

Best