Kungber Bench PS Tear Down Images - As Requested

[DW1961]

Here you go. Any comments especially welcome:

Edit: 06/02/21, just noticed they do list ripple:
Constant Voltage State    Voltage stability≤0.1%±3mV Low Voltage:0.2~0.3%±3mV Load stability≤0.5%±3mV Ripple noise≤30mVrms
Constant Current State    Current stability≤0.2%±3mA Load stability≤0.2%±3mV Ripple noise≤20mArms

Link: https://www.amazon.com/gp/product/B0923LXSM7








[ Specified attachment is not available ]

[DW1961]

It's not like the stuff you guys posted on the other thread. It's a metal shell that screws together with 6 screws. the plastic face plate is secured to the front by 3 of those screws that also attach the cover. It's not like molded plastic shell that is specifically molded for the electronics, like the HP that was posted, and other high end PSUs. I wasn't really impressed with the build quality of the face plate/housing. I mean, it will work, probably forever, but it's not like a molded case specifically made for the PSU. The internals looked nicely laid out and the CB soldering attachments were nice.

Let me know what else I can do and I'll post it.

No, I don't have a scope to test the power in puts out, sadly.

If any of you live in Sacramento, CA (USA) or area, and have a scope, and you would like to test it, let me know.

[beanflying]

Well it is earthed at least  Also a handful of what looks like Y class Capacitors so better than some already.

On the Earthing front some (me included) will dislike soldered ring terminals and also snip the Earth out of the cable tie and tie the Active and Neutral with a new one. The Sharp kink on the Earth copper combined with the solder isn't a good thing in terms of stress so reduce it. Better thing would be to replace the ring terminal with a crimped one and make sure there is a shakeproof washer under the screwhead (not clear if there is one).

What was alluded to in your 'should I buy thread' by one of the later responders was to also check with your DMM if there is an AC coupling from the Earth to the DC Ground terminal. While I would hope you are ok with this one there have been plenty of examples where the DC Ground is referenced to mains and can be dangerously high. This can fry both humans and electronics. So Test it using your DMM on AC and with the supply turned on measure Earth to DC Ground.

[xavier60]

What's the part numbers on U18 and U19?

[DW1961]

Well it is earthed at least  Also a handful of what looks like Y class Capacitors so better than some already.

On the Earthing front some (me included) will dislike soldered ring terminals and also snip the Earth out of the cable tie and tie the Active and Neutral with a new one. The Sharp kink on the Earth copper combined with the solder isn't a good thing in terms of stress so reduce it. Better thing would be to replace the ring terminal with a crimped one and make sure there is a shakeproof washer under the screwhead (not clear if there is one).

What was alluded to in your 'should I buy thread' by one of the later responders was to also check with your DMM if there is an AC coupling from the Earth to the DC Ground terminal. While I would hope you are ok with this one there have been plenty of examples where the DC Ground is referenced to mains and can be dangerously high. This can fry both humans and electronics. So Test it using your DMM on AC and with the supply turned on measure Earth to DC Ground.

I'd like to make those changes. Before I begin I'll upload images to make sure I understand what you are saying.

As far as testing the earth to DC ground, I want t make DAMN sure I have clarification as to what parts you are referencing inside the PS.  I think I get it, but I want to make sure.

[DW1961]

This is a real noob question, but when I connect the PSU to a load, it gives me a reading of watts and amps. Given that the unit is accurate, if that going to be the same amp reading  would get testing amperage with a DMM?

[tautech]

This is a real noob question, but when I connect the PSU to a load, it gives me a reading of watts and amps. Given that the unit is accurate, if that going to be the same amp reading  would get testing amperage with a DMM?

Why wouldn't it be ?
Obviously this PSU has an internal current shunt in series with the output and if the Vset takes its reference from the load side of the shunt it should track fairly close irrespective of the load.

[DW1961]

This is a real noob question, but when I connect the PSU to a load, it gives me a reading of watts and amps. Given that the unit is accurate, if that going to be the same amp reading  would get testing amperage with a DMM?

Why wouldn't it be ?
Obviously this PSU has an internal current shunt in series with the output and if the Vset takes its reference from the load side of the shunt it should track fairly close irrespective of the load.

I remember  back when I was using a DMM test total AMPs for custom LED strips I was making for my computer. I would plug in the LEDs power brick, then connect the DMM to that and yadda yadda. So, I did it, but it was a pain. Then I saw the bench PSU and they had amp readouts, so that's when I really wanted one.

So, anyway, I was just wondering if they would be as accurate as a DMM, at least to 2 decimal places. I didn't test it yet with a DMM hooked up to it, but I did calculate the average amps per 5050 SMDs, which is .01875 amps @ 12V, per SMD. It was dead on. Pretty awesome. I'm excited about it.

I am not excited about how the face plate is mounted. The bottom kinda clicked into place, but the sides and top have one screw each side, which is from the cover. The screws do go through the steel sides of the cover, but face plate screw holes are plastic. I don't like that. They should have threaded inserts in the plastic.

[beanflying]

Well it is earthed at least  Also a handful of what looks like Y class Capacitors so better than some already.

On the Earthing front some (me included) will dislike soldered ring terminals and also snip the Earth out of the cable tie and tie the Active and Neutral with a new one. The Sharp kink on the Earth copper combined with the solder isn't a good thing in terms of stress so reduce it. Better thing would be to replace the ring terminal with a crimped one and make sure there is a shakeproof washer under the screwhead (not clear if there is one).

What was alluded to in your 'should I buy thread' by one of the later responders was to also check with your DMM if there is an AC coupling from the Earth to the DC Ground terminal. While I would hope you are ok with this one there have been plenty of examples where the DC Ground is referenced to mains and can be dangerously high. This can fry both humans and electronics. So Test it using your DMM on AC and with the supply turned on measure Earth to DC Ground.

I'd like to make those changes. Before I begin I'll upload images to make sure I understand what you are saying.

As far as testing the earth to DC ground, I want t make DAMN sure I have clarification as to what parts you are referencing inside the PS.  I think I get it, but I want to make sure.

In your case as the Earth is on the front panel si it's simple and safe. DMM set to maybe 200V AC to start. One probe on the Earth and one on the DC Ground. Check it and if it is flicking up or down by a volt or two drop the range to say 20V AC and re test and so on down the range.

We ran into this with the Feeltech Frequency generators and found the isolation was complete crap and the ground was floating at 80-90V. If by comparison as I just did boot up my 'Siglent' to keep Tautech happy  I get about 40mV AC. While I was at it quick check of my Agilent 6632B Power supply was under 0.1V AC. This is sort of where you want to be in terms of numbers and not 10's of volts.

[DW1961]

Well it is earthed at least  Also a handful of what looks like Y class Capacitors so better than some already.

On the Earthing front some (me included) will dislike soldered ring terminals and also snip the Earth out of the cable tie and tie the Active and Neutral with a new one. The Sharp kink on the Earth copper combined with the solder isn't a good thing in terms of stress so reduce it. Better thing would be to replace the ring terminal with a crimped one and make sure there is a shakeproof washer under the screwhead (not clear if there is one).

What was alluded to in your 'should I buy thread' by one of the later responders was to also check with your DMM if there is an AC coupling from the Earth to the DC Ground terminal. While I would hope you are ok with this one there have been plenty of examples where the DC Ground is referenced to mains and can be dangerously high. This can fry both humans and electronics. So Test it using your DMM on AC and with the supply turned on measure Earth to DC Ground.

I'd like to make those changes. Before I begin I'll upload images to make sure I understand what you are saying.

As far as testing the earth to DC ground, I want t make DAMN sure I have clarification as to what parts you are referencing inside the PS.  I think I get it, but I want to make sure.

In your case as the Earth is on the front panel si it's simple and safe. DMM set to maybe 200V AC to start. One probe on the Earth and one on the DC Ground. Check it and if it is flicking up or down by a volt or two drop the range to say 20V AC and re test and so on down the range.

We ran into this with the Feeltech Frequency generators and found the isolation was complete crap and the ground was floating at 80-90V. If by comparison as I just did boot up my 'Siglent' to keep Tautech happy  I get about 40mV AC. While I was at it quick check of my Agilent 6632B Power supply was under 0.1V AC. This is sort of where you want to be in terms of numbers and not 10's of volts.

Well, bad news. At first I measured the current with my Amazon DMM. It only has two settings, 200 and 600V, but on the 200V setting, it will measure down to 0.1V. It measured between 000.1 and 000.0. I then plugged in my Unity DMM and it measured 35V. So I used another DMM and it also measured 35V. So I got the Amazon DMM specifications out and it said it will measure down to .1V at 50-60Hz. I then stuck it into a 120V 60hz outlet and it measured 121V and so did the other two DMMs. So, I have no idea why the Amazon DMM isn't showing the same voltage as the two other DMMs. But 35V I guess is pretty shitty.

Just to be clear, this is with the power supply AC on and the power to the DC side enabled also. It has the ability to power on, but not supply power to the DC circuitry. You can se in the images there are two power buttons. Again, I tested it with the entire unit energized.

 I'll contact the vendor and let them know. We'll see what they say.

[radiolistener]

So, I have no idea why the Amazon DMM isn't showing the same voltage as the two other DMMs. But 35V I guess is pretty shitty.

difference can be strong indication that the output has significant ripple and pulses. Different DMM have different filters.

That's switching mode power supply....

[Kleinstein]

The supply is resonably isolated from ground and mains. So when measuring the AC voltage relative to ground one has a very high impedance voltage source to measure.  Different DMMs can be quite different in the input impedance in the AC ranges. Some are close to 10 or 11 Mohms and some are more like 1 M Ohms. So the different readings are likely just because of different impedance of the meters.

The residual voltage is likely due to the EMI suppression capacitors from both sides of mains to the ground at the output side. With 2 equal caps this would be halve the main voltage (some 60 V). The meters read less, as the impedance of the parasitic voltage seeems to be about as higher as the impedance of the meter - so reading only half of the open circuit voltage.  One could test this effect by using the 2 meters in series. Instead of some 30 and 35 V I would expect something like 20 and 24 V for the 2 meters.

[DW1961]

The supply is resonably isolated from ground and mains. So when measuring the AC voltage relative to ground one has a very high impedance voltage source to measure.  Different DMMs can be quite different in the input impedance in the AC ranges. Some are close to 10 or 11 Mohms and some are more like 1 M Ohms. So the different readings are likely just because of different impedance of the meters.

The residual voltage is likely due to the EMI suppression capacitors from both sides of mains to the ground at the output side. With 2 equal caps this would be halve the main voltage (some 60 V). The meters read less, as the impedance of the parasitic voltage seeems to be about as higher as the impedance of the meter - so reading only half of the open circuit voltage.  One could test this effect by using the 2 meters in series. Instead of some 30 and 35 V I would expect something like 20 and 24 V for the 2 meters.

Thanks for that explanation.

One of the meters show no voltage, or almost no voltage. Should I use the meters that read 30-35V or one that shows 35V and one that shows .1 volts?

Also, what does this mean? Is the PSU faulty or badly engineered? Or, without the suppression caps, would be seeing very low voltage?

[beanflying]

Not faulty for a start far more likely by design which is what Kleinstein alluded to. What you are looking at is ripple on top of the DC voltage similar to this below.



In your case what you are likely getting is something with more large spikes in part due to it being a switching supply. You might want to have a look at this thread just for some mixed random images of noise https://www.eevblog.com/forum/chat/show-me-your-psu-noise/

Drop a load on it now preferably resistive (so a Fat Resistor, light globe or LED string) and retest it and see how it goes under load with the same tests.

[DW1961]

Drop a load on it now preferably resistive (so a Fat Resistor, light globe or LED string) and retest it and see how it goes under load with the same tests.

I got 43V now under a very light load 0.2A.

[shakalnokturn]

Although they do play a big role in getting some of the measurable unwanted AC on the secondary side I don't think the EMI suppression capacitors are the only cause.
(In any case I wouldn't say that DC ground is "referenced" to mains if it isn't ground referenced.)
The transformer itself has parasitic capacitance between primary and secondary, the switching frequency also has some effect on the amount of stray AC on the secondary side, under load there will be more ringing in the transformer which can also affect the amount of AC capacitively coupled through the transformer.

While it's nice to know what amount of mains leaked AC to expect on the secondary, the more relevant measurement for user safety is the amount of leakage current available.

For component safety, if you're using MOS components in the wrong conditions even 35V will be enough to fry them.

In most cases it's probably wiser to permanently tie the negative or common output terminal to ground on SMPS lab supplies.

[DW1961]

Although they do play a big role in getting some of the measurable unwanted AC on the secondary side I don't think the EMI suppression capacitors are the only cause.
(In any case I wouldn't say that DC ground is "referenced" to mains if it isn't ground referenced.)
The transformer itself has parasitic capacitance between primary and secondary, the switching frequency also has some effect on the amount of stray AC on the secondary side, under load there will be more ringing in the transformer which can also affect the amount of AC capacitively coupled through the transformer.

While it's nice to know what amount of mains leaked AC to expect on the secondary, the more relevant measurement for user safety is the amount of leakage current available.

For component safety, if you're using MOS components in the wrong conditions even 35V will be enough to fry them.

In most cases it's probably wiser to permanently tie the negative or common output terminal to ground on SMPS lab supplies.

So what does this mean for the power supply discussed here?

[shakalnokturn]

Not much... Just probably wiser to tie negative output and ground together either internally or externally with something permanent.

[DW1961]

Not much... Just probably wiser to tie negative output and ground together either internally or externally with something permanent.

Can't you just do that manually? Also, if it were permanent, it wouldn't be floating and the benefits of a floating PSU would be lost?

I'm just wondering if anyone found anything questionable in the pictures, or would like some more detailed images for the unit? That is, did anyone see anything that was alarming, or anything that was positive (like the Y caps).  For instance, how would I test it for EMI, or is it already acceptably protected?


Does it get a pass from the community?

[beanflying]

Without a scope to have more of a look at the actual noise we are like you taking a bit of a stab in the gloom.

Back before I got a few decent Linear supplies I was things like a PC ATX or my 0-15V 60A Manson as a raw source then I had made a couple of small Linear 78xx based convertors for powering anything delicate but fed from the SMPS. So much as it is probably not needed just for the safety of those sorts of devices in your place I would make something similar and feed it off this supply. These days maybe even look at some low drop out Regs over the 78xx series.

There are some options to reduce that floating voltage a bit and they were discussed buried in the Feeltech mega thread (bound to be others if you search) so you might like to have a read before and after this and also and in particular the Keysight link in the post https://www.eevblog.com/forum/testgear/feeltech-fy6600-60mhz-2-ch-vco-function-arbitrary-waveform-signal-generator/msg1344167/#msg1344167

[DW1961]

Without a scope to have more of a look at the actual noise we are like you taking a bit of a stab in the gloom.

Back before I got a few decent Linear supplies I was things like a PC ATX or my 0-15V 60A Manson as a raw source then I had made a couple of small Linear 78xx based convertors for powering anything delicate but fed from the SMPS. So much as it is probably not needed just for the safety of those sorts of devices in your place I would make something similar and feed it off this supply. These days maybe even look at some low drop out Regs over the 78xx series.

There are some options to reduce that floating voltage a bit and they were discussed buried in the Feeltech mega thread (bound to be others if you search) so you might like to have a read before and after this and also and in particular the Keysight link in the post https://www.eevblog.com/forum/testgear/feeltech-fy6600-60mhz-2-ch-vco-function-arbitrary-waveform-signal-generator/msg1344167/#msg1344167

Is the floating voltage on this unit bad or could cause problems?

[shakalnokturn]

Can't you just do that manually? Also, if it were permanent, it wouldn't be floating and the benefits of a floating PSU would be lost?

Is the floating voltage on this unit bad or could cause problems?

If it's permanent the floating output is lost, sure you can do it manually.
It all depends on your typical usage. I just wouldn't trust myself to leave it floating because usually a SMPS isn't floating enough for me.
It isn't a criticism of this PS, but more a general "be aware of what you're using"...

I mentioned MOS previously because one of my last traps was using a "floating" SMPS for a quick and dirty Gate drive to beefy MOS transistors used to discharge beefy capacitors into circuit breakers to simulate and scope short-circuit opening.
I killed a couple of transistors before understanding that when replacing the circuit breaker touching the MOS Drain wire I was putting the "floating" SMPS to ground through the Gate...

Considering something "floating" is not that simple, ideally you'd consider what impedance it's floating at and breakdown voltage. If I want something floating I go for a SMPS behind a isolation transformer or a linear PS.

I'm just wondering if anyone found anything questionable in the pictures...

Nothing awful for me other than the earthing that could be a bit better as others have mentioned.

[DW1961]

Nothing awful for me other than the earthing that could be a bit better as others have mentioned.

You mean the wire angles and the soldered wire vs crimped?

[DW1961]

I just noticed they listed they ripple on this unit:

Constant Voltage State    Voltage stability≤0.1%±3mV Low Voltage:0.2~0.3%±3mV Load stability≤0.5%±3mV Ripple noise≤30mVrms
Constant Current State    Current stability≤0.2%±3mA Load stability≤0.2%±3mV Ripple noise≤20mArms

[beanflying]

Well a few weeks later and I had a cheapo "Wanptek' branded 120V 3A power supply arrive with a powerful inner core  This job is fine with some ripple and noise as it will just be powering a hot wire cutter.



Interesting having a compare it along with yours are either the same or a close cousin mainboard and likely front display board (different orientation). Similar slightly suspect Earthing connections but at a quick first look somewhere close.

I will clear the bench and fire up some proper testing over the next day or two including a look at ripple/noise and loaded and unloaded.