Korad KA3003D (Tenma 72-10480) power supply - start up voltage spike

[Rogersan]

This power supply can produce a 19 volt spike at the output terminals as it is powered on. This is in spite of the supply having been set previously to 5 volts, and the On/Off switch not having been operated. The On/Off switch is not isolating the output terminals at power up

Any sensitive circuitry connected to this supply can be destroyed at power on. I tried a 1 watt 5.1 Zener diode across the output terminals and this was shorted out by the surge so there is a fair bit on energy available in the start up spike.

                 Roger Sanderson

[nrxnrx]

Thank you for the warning!

I just checked mine and it also happens when you turn it off.

[Andreas]

The On/Off switch is not isolating the output terminals at power up

Are you shure that the relay for the output is not defective (hanging).
I usually have no spikes with a KA3005 as the relay switches off during power off.
Only if I actuate the power on switch very fast from on to off to on (before the relay can switch off) I can generate some overshoot.

With best regards

Andreas

[nrxnrx]

I can only find two relays inside the unit and none seem to be used for the output.
I have't traced the circuit, but I hear no click at power-up or when turning the output on or off.

--
Mihai

[Andreas]

Ok,

so it seems they only have a relay on the KA3005.
I hear it clicking on switching the output on/off.
Default on power on is output off.

With best regards

Andreas

[tautech]

This power supply can produce a 19 volt spike at the output terminals as it is powered on. This is in spite of the supply having been set previously to 5 volts, and the On/Off switch not having been operated. The On/Off switch is not isolating the output terminals at power up

Any sensitive circuitry connected to this supply can be destroyed at power on. I tried a 1 watt 5.1 Zener diode across the output terminals and this was shorted out by the surge so there is a fair bit on energy available in the start up spike.

                 Roger Sanderson

Welcome to the forum.

You really need to offer a screenshot for proof of your claims.
Not that I don't believe you but something like this would be in order:
Borrowed from this thread:
https://www.eevblog.com/forum/testgear/cheap-variable-switching-power-supply/msg666082/#msg666082



Appropriate trigger settings, timebase and input coupling required. 

[nrxnrx]

Attached are two screenshots of probing across a 10ohm resistor and a 100ohm one - when powering off.
(at the end of 2 ~30cm leads)

DC coupling for the signal & trigger.

[Kleinstein]

The spikes look like noise from a switcher - many pulses at constant frequency with massive ringing. It so high in frequency that it may get to the scope through the air as well. The shown waveform may not fully show the true voltage though.
This has not much to do with the normal power on-power / power-off spikes seen with some supplies. It more like a massive EMI problem.

A first test would be to turn the supply of / on by pulling the plug. It might be the switch is bad, producing excessive arcing - though it should be rather bad for this.

I would have a good look at the capacitors at the output.

Is there this type of RF interference also present during normal operation ?

[tautech]

@ nrxnrx
Use single shot triggering.
Have a fiddle to find the right V/div and timebase settings.
You'll notice my screenshot shows Stop, that's the single shot doing it's job.....my finger isn't that fast. 

[nrxnrx]

@ nrxnrx
Use single shot triggering.
Have a fiddle to find the right V/div and timebase settings.
You'll notice my screenshot shows Stop, that's the single shot doing it's job.....my finger isn't that fast. 

I can't trigger manually for microsecond events either . I did use single shot for all the screenshots I posted here (top left of the pictures, to the right of "RIGOL").
As for the V/div settings, what do you want to see?

It's a linear supply and during normal operation the output is nice and clean.

--
Mihai

[tautech]

@ nrxnrx
Use single shot triggering.
Have a fiddle to find the right V/div and timebase settings.
You'll notice my screenshot shows Stop, that's the single shot doing it's job.....my finger isn't that fast. 

I can't trigger manually for microsecond events either . I did use single shot for all the screenshots I posted here (top left of the pictures, to the right of "RIGOL").
As for the V/div settings, what do you want to see?

It's a linear supply and during normal operation the output is nice and clean.

--
Mihai

The thread is about power on spikes. 

Then try normal triggering and single shot, set the trigger to a value the signal will cross at power on and you should have a power on spike capture.

[Andreas]

Hello,

first I have to correct myself. The clicking that I heard is only at higher output voltages than 5V.
(So it is the relay for the transformer winding at higher voltages).

I repeated the measurements with 5V (usually I do not use a 30V supply for a 5V load).

On switching mains on I see 0.8V (low energy pulse < 1us) at the output. (I guess capacitive coupling from the mains transformer to the output).

Switching the output on gives up to 0.4V overshoot to the 5V for around 5 ms on my device.

Mains off gives some contact bounce at the mains switch together with the transformer inductivity (burst impulses).
Zooming in shows that they are low energy. (although up to 4V above the 5V)
I guess that most of those pulses travel by the mains line or capacitive coupling into the scope.
(The scope may be disturbed by those pulses).

Finally especially for Kleinstein the noise in normal operation (AC coupled)
first with 20 Mhz the 2nd with 100kHz bandwidth.

So there are no switching noise spikes from the linear power supply.

with best regards

Andreas

[nrxnrx]

The thread is about power on spikes. 

Then try normal triggering and single shot, set the trigger to a value the signal will cross at power on and you should have a power on spike capture.

The problem as I understand it is that the thing outputs highish energy pulses when the unit is turned on and off, even though the output is nominally "off".

The On/Off switch is not isolating the output terminals at power up

Perhaps I misunderstand?

[Andreas]

The problem as I understand it is that the thing outputs highish energy pulses when the unit is turned on and off, even though the output is nominally "off".

Perhaps I misunderstand?

The question is high energy pulses. (19V (how long?) instead of 5V).

On my unit it is max overshoot to 5.4V for 5ms with light loads.

On your unit it is most probably a measurement problem of your scope with switch bouncing.
Disconnect the cable on the Korad side and leave it connected to the scope and measure again.

With best regards

Andreas

[Mr. Scram]

I've recently received one of these power supplies, a two channel model, and noticed the same issue along with a few others.

• When the unit powers on, you often hear a loud "twang" sound, like you used to hear in CRT monitors. A vibration coinciding with the sound can be felt. Sometimes dimmed lights nearby flicker slightly.
• There's 3,3 mV on one output and -5 mV on the other when the outputs are off. The supply display shows nothing.
• There seems to be a far amount of voltage spiking when turning the unit on and off. Voltages as high as 80 Vpp can be seen. When loaded down with a 20 ohm resistor, the spikes drop to about 10 Vpp. Turning channels on and off isn't a problem. There's very minor overshoot when turning on a channel, but otherwise the unit is well behaved.

The spiking isn't always present and doesn't necessarily coincide with the twang noise. I've repeated the measurement with the leads disconnected from the power supply but with them next to it, with the leads still attached to the oscilloscope. Even though there is a signal induced, it's negligible in duration and amplitude compared to the signal when the leads are connected. It's not a muted mirror image either, so the signal definitely seems conducted.

[Andreas]

Hello,

did you use proper shielded cables/connectors?
E.g. similar to this connector togeter with a BNC-cable?

https://www.digikey.com/products/en?keywords=BKCT2939-ND

That what you show (50ns pulses) is nothing that can be generated by the (slow) power transistors of the power supply.
These are common mode pulses generated by bouncing of the mains switch together with the inductivity of the transformer.

with best regards

Andreas

[Mr. Scram]

Hello,

did you use proper shielded cables/connectors?
E.g. similar to this connector togeter with a BNC-cable?

https://www.digikey.com/products/en?keywords=BKCT2939-ND

That what you show (50ns pulses) is nothing that can be generated by the (slow) power transistors of the power supply.
These are common mode pulses generated by bouncing of the mains switch together with the inductivity of the transformer.

with best regards

Andreas

I didn't use those connectors. I've ordered them, but haven't received them yet. However, instead of user two leads, I've now probed the binding posts directly with the scope probe. I think I'll repeat the test later, as results haven't been completely consistent in regards to the voltages generated, but this seems to be a representative sample of spikes during turn on. Turning the power supply off generally generates a lot more spikes. Probe at 1x, DC coupling on channel and trigger.

Either way, it seems the voltages could be problematic for certain parts and devices attached to the power supply.

[Andreas]

Either way, it seems the voltages could be problematic for certain parts and devices attached to the power supply.

Hello,

these short pulses even if they should be differential mode (between + and - pole of supply) do not carry much energy.

But they are most probably common mode (between poles and earth ground).
Thats why you see large differences when connecting different ways.

with best regards

Andreas

[Mr. Scram]

Hello,

these short pulses even if they should be differential mode (between + and - pole of supply) do not carry much energy.

But they are most probably common mode (between poles and earth ground).
Thats why you see large differences when connecting different ways.

with best regards

Andreas

What would be a relevant method to test the nature and energy of the spikes?

I guess I could add a resistor across the leads. With a known voltage and resistance, current and energy could be deduced.

[Andreas]

I guess I could add a resistor across the leads. With a known voltage and resistance, current and energy could be deduced.

Hello,

the measurement at these frequencies is not trivial.
You need a coaxial resistor suitable for some 100 MHz.
Scope or power supply has to be in a shielded cabinet.
(the coaxial resistor through the wall).

With best regards

Andreas

[ziplock9000]

I just bought one of these from eBay that is supposed to be in excellent condition. This version has the binding posts, which I understand is the older of the two models (although that are both Tenma 72-10480)
If I find I get the spikes outlined above when I check it with my scope, is there anything I can do to mitigate the effects of this? Maybe replace a component?

[Andreas]

If you mean the high frequency pulses from arcing of the mains switch:
Try to convince that he should not arc. (e.g. with a snubber).

With best regards

Andreas

[ziplock9000]

If you mean the high frequency pulses from arcing of the mains switch:
Try to convince that he should not arc. (e.g. with a snubber).

With best regards

Andreas

What type & values of snubber would work well for this PSU?

[ziplock9000]

It arrived. I did a power on (via hardware switch) test when set to 5v / 200mA and no load and I don't get a spike >5v, although it's close. I've tried this in various modes (peak, normal, hi-res, 20m BW limit).

I'll do some more testing later with different voltages and loads. So far I'm very happy with it and it's a great upgrade to my Farnell L30-2

[Mr. Scram]

What did you end up installing?

[Terry01]

I have the Tenma 30v 5A model of this PSU, it's been a great wee unit for the year or two I've had it. I got it new when I started out with electronics. I have noticed the fan getting a little bit louder lately. I've given it a check and blow/hoover out so there is nothing foreign in there. I would like to replace the fan with a silent one, or as near to silent as possible. Now I have noticed it I think I hear it more than i normally would anyways if that makes sense?

Does anyone know offhand which fan I need to replace the "noisy" one in it now? I really could use someone who knows better than me to guide me a little. I have never replaced one of these before and know when it comes to mains powered stuff it's for sure better to ask and get it right first time than try myself and fluff it and find out the hard way I as wrong.

Any help will be much appreciated!

[Mr. Scram]

I have the Tenma 30v 5A model of this PSU, it's been a great wee unit for the year or two I've had it. I got it new when I started out with electronics. I have noticed the fan getting a little bit louder lately. I've given it a check and blow/hoover out so there is nothing foreign in there. I would like to replace the fan with a silent one, or as near to silent as possible. Now I have noticed it I think I hear it more than i normally would anyways if that makes sense?

Does anyone know offhand which fan I need to replace the "noisy" one in it now? I really could use someone who knows better than me to guide me a little. I have never replaced one of these before and know when it comes to mains powered stuff it's for sure better to ask and get it right first time than try myself and fluff it and find out the hard way I as wrong.

Any help will be much appreciated!

The original model has a bit of a peculiar heatsink setup which causes most of the noise. Users who swapped the fan report not much improvement. I understand the later model has a different heatsink, so that might actually benefit from a different fan.

[Terry01]

Thanks for the help.
I did notice that mine has the back plate of the fan bolted on to the rear of the unit and the front plate of the fan has a large heatsink bolted on to it so maybe it's the older model. I can take a photo tomorrow and post it up for you to see. It's not hellish noisy but now it's caught my attention it annoys me! Silly eh?

[Mr. Scram]

Thanks for the help.
I did notice that mine has the back plate of the fan bolted on to the rear of the unit and the front plate of the fan has a large heatsink bolted on to it so maybe it's the older model. I can take a photo tomorrow and post it up for you to see. It's not hellish noisy but now it's caught my attention it annoys me! Silly eh?

I got my information from the video below. There's another guy reporting a good improvement with a similar modification in another video, so YMMV.

https://youtu.be/Hmc-2NBIzts

https://youtu.be/fgdS_3ZApIQ

[Terry01]

Interesting results!
1 is a waste of time and 1 is a good bit quieter. I'll order up a new fan and give it a go. Even if it turns out to be no good at least I can say I tried to quieten it down and I can moan like crazy about it and kick it now and then! Even call it "a P,O,S fan"!
I can hardly moan and whine about it if I don't at least even try to fix it right?
It may even turn out that mines is the type to benefit from a new fan and end up silent.

We'll see! I'll post my results when I've replaced it and let anyone who has the same problem know if I think it's worth the time or not. I doubt it but maybe a better fan will be what works. The first guy who tried it, his fan was blootered in hot snot. That can't be good for the fan either.

[ziplock9000]

Interesting results!
1 is a waste of time and 1 is a good bit quieter. I'll order up a new fan and give it a go. Even if it turns out to be no good at least I can say I tried to quieten it down and I can moan like crazy about it and kick it now and then! Even call it "a P,O,S fan"!
I can hardly moan and whine about it if I don't at least even try to fix it right?
It may even turn out that mines is the type to benefit from a new fan and end up silent.

We'll see! I'll post my results when I've replaced it and let anyone who has the same problem know if I think it's worth the time or not. I doubt it but maybe a better fan will be what works. The first guy who tried it, his fan was blootered in hot snot. That can't be good for the fan either.

You may want to have a look on a computer builder / modder forums around the internet. They know more about quiet, high throughput, efficient etc fans on there than anyone else really.

[Andreas]

Hello,

I am using this fan to reduce the noise:
https://www.reichelt.de/arctic-housing-fan-f8-tc-80-mm-ac-fan-f8tc-p131300.html?

the temperature sensor is fixed to the cooling plate.

with best regards

Andreas

[Terry01]

Hello,

I am using this fan to reduce the noise:
https://www.reichelt.de/arctic-housing-fan-f8-tc-80-mm-ac-fan-f8tc-p131300.html?

the temperature sensor is fixed to the cooling plate.

with best regards

Andreas

Does this fan reduce the noise by a fair bit? Also do you have the older model or the newer one?

[Andreas]

Hello,

at least at low temperatures / low power operation.
So for me it works good.

Another known mod is to place a 47uF capacitor across the fan.
But this did not reduce noise in my case sufficiently.

What do you mean with old / new model?

with best regards

Andreas

[Terry01]

Hello,

at least at low temperatures / low power operation.
So for me it works good.

Another known mod is to place a 47uF capacitor across the fan.
But this did not reduce noise in my case sufficiently.

What do you mean with old / new model?

with best regards

Andreas

Thanks for the help. 
I'm sure the rear board design was changed at some point. I think the way the heatsink is attached or something like that. I'll look out where I read that, it may even have been someone here who told me that.