History of this webpage:

August 12, 2009: This is what I like

August 10, 2009: There is hope: 3 pin SMPS's are there

 August 8, 2009: small video on Ycap-effect  and a bit larger video of scoped effect

Updated August 7, 2009     More scope pictures     Click to go there

Update August 6, 2009    Six wallwarts more tested     Click this link to go there

August 5, 2009: start of this webpage Inspired by Doc's posting on AVRfreaks
http://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&t=82257&postdays=0&postorder=asc&start=0&sid=c69ca828fe2e2f4026c340e6d1123118

Curious as allways, I did some tests on a bunch of SMPS and Transformer Power Supply modules

The (in)-famous capacitor as I found in the Li Ion-chargers:




Same thing in Topview






The modules I tested:      < right-click picture to store in full size format  >


See Sht 1 and 2 of schematic


See Sht 3 and 4 of schematic




Wish to edit / use the file: Results in ExpressSch-format
Don't have- or use Express Schematic: see my Tools-page

Results as .png files





previous revision

Today ( August 6 2009) I tested six more. Since you know by now how I test these things, I'll stick to figures and pictures only.
These are the results::



From left to right:

1. Low cost replacement cellphone-charger:
Leakage measurement:   82V AC and 40uA ; quite standard

I also wanted to know a bit more on how that leakage-current looks like.
The currents are small, so I used a 1k resistor as 'shunt' and put the scope-probe over this shunt



The first scope picture: 1. Low cost replacement cellphone-charger:

2. Chinese LiIon-charger for 2 cells, 500mA: 
Leakage measurement:  80V AC and 40uA
Same values as the previous one, but look how different the waveforms are !

3. Hapé universal wallwart, upto 2.5A, adjustable outputvoltage from 3 to 7.5V
This one tickles ...
Leakage measurement: 92V AC and 160uA !  
This is the first one that is showing more leakage than I like

It also shows how wrong multimeters can be: the peak lies around 500mV over 1kOhm, so that's 500uA
If it was a full sine, RMS would be around 350uA. But since half of the sine is missing: 175uA
More on multimeters here

A sidetrack: let's have a look at the transformer PS that is on Sht 3, the top one: it's 12V DC, 1A
I measured 3 or 4 uA, depending on how it is plugged in.



The transformer beats them all huh ? But wait ...

4. Very recent Samsung cellphone battery charger
Leakage measurement:  12V or 3V AC, depending on plug-orientation,  and only 2 uA !! The winner so far.

5. Very recent Sony Ericson cellphone battery charger
Leakage measurement:  12V AC and 5.3 or 7.5 uA , depending on plug orientation

6. Chinese LiIon-charger for 4 cells, 1000mA: 
Leakage measurement:  78V AC and 87 uA


Then a note on the used instruments: none of mine is capable of measuring True RMS. And even when I would have had such an instrument, it would not be able to cope with the high frequencies in the spectrum. For one SMPS wallwart though it would be usefull: nr 3. Hapé universal wallwart, the tickling one.
I tested with 3 different multimeters: all three capable of measuring small AC currents, but basically they double rectify the measured current, take the avarage, and then scale it to the corresponding RMS value, assuming a sinewave form signal of 40 to 400 Hz

Conclusion so far:

- The tested wall warts are not too bad  .... except for nr. 3 in the latest batch, the Hapé Universall wall wart.
- To keep it safe, connect one side of the secundary to safety ground. In almost all cases we'll choose the minus-pole
   With doing so, a new problem arises: (safety) ground loops. Frankly, I am aware of those and used to it. But that may be different for you, reader :-)
- Important thing rberger writes about:
The main problem with switchmode plugpacks with a capacitor from output to input rectifier is not necessarily the leakage current but the instantanious voltage on the capacitor when the plugpack is connected to one of our applications when there is a pre-existing grounded connection such as a CRO or PC Serial port.

For reference: this is the URL to the AVRfreaks.net topic: http://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&t=82257&postdays=0&postorder=asc&start=0

August 7, 2009

You will need a          W     I     D     E       screen monitor to view the tables below

I know, I am a lousy webpage designer, but I can live with it :-)

This is how I measured the leakage currents with the oscilloscope:
< as usual, use right click, open in new tab or page if you want the full size, or if you wish to store it >



I checked the oscilloscope-probe specifications and in 10x mode they can handle 300V rms. In 1x mode it's  only 150V rms. So I need to be carefull.
It's too much work to go through all the wallwarts and other PS's again, so I made a selection:

a. Transformer based 5V 800mA stab wallwart for Ethernet hub
b. Transformer based 19V 500mA unstab wallwart for powerlights
c. the best SMPS: Samsung recent cellphone charger
d. the worst SMPS: Hape Univ wallwart
Wallwarts a and b were not tested before, but for AVRfreaks those are good candidates for projects. (Ken's remark)

a. Transformer based 5V 800mA stab wallwart for Ethernet hub



a. Transformer based 5V 800mA stab wallwart for Ethernet hub


a. Transformer based 5V 800mA stab wallwart for Ethernet hub



a. Transformer based 5V 800mA stab wallwart for Ethernet hub: plugged in "one way"


a. Transformer based 5V 800mA stab wallwart for Ethernet hub: plugged in "the other way"


a. Transformer based 5V 800mA stab wallwart for Ethernet hub: there is just a minor, negliable difference in leakage current when the wallwart is plugged-in "one way" or "the other"

For the voltage image: taking the (well-isolated) secundary cable in my hand, made the amplitude increase. So there is very little oompf in that coupling. So very safe to use. Given that the transformer is built in the correct way, with sufficient isolation between primary and secundary windings.
b. Transformer based 19V 500mA unstab wallwart for powerlights


b. Transformer based 19V 500mA unstab wallwart for powerlights

b. Transformer based 19V 500mA unstab wallwart for powerlights: current image over 1k


b. Probe 1x  1 MOhm

b. Probe 10x  10 MOhm

b. Here you can see that the capacitive coupling between primary and secundary winding is very small, and therefor it behaves as a constant current source: increasing the probe-impedance from 1 MOhm to 10 MOhm makes the amplitude increase with 10 as well (bpf)

For the voltage images: same as on a. : taking the (well-isolated) secundary cable in my hand, made the amplitude increase. So there is very little oompf in that coupling. So very safe to use. Given that the transformer is built in the correct way, with sufficient isolation between primary and secundary windings.

c. the best SMPS: Samsung recent cellphone charger


c. the best SMPS: Samsung recent cellphone charger

c. the best SMPS: Samsung recent cellphone charger: current image over 1k


c. Probe 1x  1 MOhm

c. Probe 10x  10 MOhm

c. the best SMPS: Samsung recent cellphone charger:
Also here we find that constant current source effect: and very precise ! No bpf (ballpark figure)
Lookin at the current picture: spiky stuff but very small


d. the worst SMPS: Hape Univ wallwart


d. the worst SMPS: Hape Univ wallwart

d. the worst SMPS: Hape Univ wallwart: familiar picture, isn't it ?


d. the worst SMPS: Hape Univ wallwart   Probe 10x  10 MOhm

d. the worst SMPS: Hape Univ wallwart   Probe 1x  1 MOhm

Even with the highest sensitivity setting on the scope, the amplitude is too big to fit on the screen

d. the worst SMPS: Hape Univ wallwart


The pictures show it all: no very high impedance kind of constant current source behaviour ..... this one tickles, and is not safe !!


That's it for today. Cheers !

August 10, 2009

About 3 pin SMPS modules: there is hope Ron !
The two on the left, though quite compact, have a three wire connection.
The used plugs and receptables are not standard, and quite uncommon in the Netherlands.

Overview:  < right click to open full size etc. >
This SMPS board has been designed by an expert: it's a 5V 6A and 24V 2.4A



On the left we see extensive filtering on the input. And a solid Safety Ground connector (top left)
In the top-right corner: the 2 blue capacitors are both Y-grades: one connects secundary null to Safety Ground, the other the top of the 24V winding to Safety Ground. The very good news is that the first Ycap is the only connection between secundary null to Safety Ground. Think of f.i. a PC-PS: their secundary null is hard wired to Safety Ground. This supply allows me to put secondary null on a DC-level I need. I can even put two of these supplies in series !

Am I just showing off ??
Nope. Finding this way of doing proper EMI-filtering, without constraining the application in its use, was such a pleasant surprise that I felt I needed to share this.
For the Dutch audience: http://www.dickbest.nl/webshop/index.php or http://www.baco-army-goods.nl/
FIN