EEVblog Electronics Community Forum

Products => Test Equipment => Topic started by: dr_p on February 18, 2013, 09:26:49 pm

Title: SMPS troubleshooting aid
Post by: dr_p on February 18, 2013, 09:26:49 pm
I admit from the start I'm new to switch-mode power supply repairing, but keen to go at it.
This thread is aimed at inspiring other people to build things, maybe even duplicate this type of equipment.

I have, so far, stumbled across some inconveniences:
  - the series light bulb trick ( is annoying without an actual box to hold the bulb, plug and maybe a switch.
  - having to constantly plug-unplug the power supply, having to constantly trace the wire from the SMPS to the plug, making sure IT is unplugged, not something else. My desk is a mess, that's why.

So I decided to build a box with a plug, switchable on-off, with a series light bulb that can be bypassed. Actually I used a normal E27 and a small E14 bulb fixture, so I maximize my chances of having a compatible bulb. Added fuses and an analog voltmeter on the output, since I was dying to use it somewhere :D
Since I was at it I bundled a DIY isolation transformer, made from two identical toroidal 100VA 48VAC transformers. I think it's enough for most SMPS.

Now I find it so easy to work on any mains project: I have a master on-off switch, crude current limiting from the bulbs and a low value, fast blow fuse to keep things safe.

Below you have the schematic (for what it's worth) and a picture of the assembled unit.
Title: Re: SMPS troubleshooting aid
Post by: PA4TIM on February 18, 2013, 09:46:49 pm
Nice build. Some switchers refuse to start with a series lightbulb.
I have done about the same. But just one lightbulb. I am still thinking about an alternative for a lightbulb or sense current and at a adjustable level a lightbulb switches in as protection.
I use a 1:1 isolation transformer and a variac. Also 32VDC, 6,3VAC, 250VDC and 0-1200VDC ( this is mine, but not as nice looking as yours
Title: Re: SMPS troubleshooting aid
Post by: SeanB on February 19, 2013, 04:30:32 pm
Instead of the 3A fuse place a DIN breaker of 5a or a 3A unit if you can get it in it's place. Resettable and both fast acting and will handle surges. I have a whole box of them in a drawer, still trying to use them up.
Title: Re: SMPS troubleshooting aid
Post by: Alana on February 19, 2013, 05:04:36 pm
I sometimes "do" PC PSUs and i made smaller version - just lightblub socket and bypass switch in small case.
You may want to get as big lightblub as you can find - 60W is too low for most ATX PSUs, especially those with active PFC.
Fuses - I'd get rid of them or replace with resettable beakers above 6A. Entire thing about this box is that you have light blub as overload protection and fuses - especially those values may be problematic and burn out when not needed - for example because of initial current surge of unknown SMPS.
Title: Re: SMPS troubleshooting aid
Post by: Analogtech on February 19, 2013, 06:33:19 pm
Thats a very nice build and the schematic was bloody well made but you do need to make sure that the lightbulb can stand up to the inrush which can be pretty high for some SMPSs  and you might want to get rid of the fuses and use breakers instead it can get annoying replacing fuses
Title: Re: SMPS troubleshooting aid
Post by: PA0PBZ on February 19, 2013, 07:16:23 pm
...but you do need to make sure that the lightbulb can stand up to the inrush...

Well, it's not going to be much worse than a short and that is what they survive every day isn't it?
Title: Re: SMPS troubleshooting aid
Post by: dr_p on February 19, 2013, 08:00:15 pm
Well, I just used whatever parts I had, except the enclosure which I bought locally for the equivalent of 6$.
As I said, I'm new at mains powered stings, so I went for glass fuses thinking at least they're safe. I'll just have to wait and see if they do blow often.

I do however have a 3A resettable fuse that I salvaged from an old UPS and I might consider replacing the output fuse with it.

As for the bulb size - that's exactly why I used two sizes of bulb fixtures - I can use whatever bulb I want - 15, 25, 40 for the small one and 40, 60, 75, 100, 150, 200 for the large one.

I also wanted an ammeter, but there wasn't any room on the front panel. Oh well...

Now I'm thinking of another DIY tool for troubleshooting SMPSs - a mixed ESR+capacitance+leakage meter ( and a ring tester ( And as a bonus it might just fit on top of my bulb box :D
Title: Re: SMPS troubleshooting aid
Post by: PA4TIM on February 19, 2013, 08:43:44 pm
I see no use for the ringtester, Q is depending on storage of energy and a ratio between resistance and reactance. Reactance is depnding on self inductance, frequency and voltage ( in case of non-air coils) but maybe this thing measures the parameters and does the math.

The ESR meter explanation in the pdf stinks. He mixes up impedance and reactance and more ( like meauring ESR at other frequency as 100 KHz is wrong) but the idea probably wil work. Itis the way to measure Z and Z is what the datasheets tell you at 100 KHz. So the outcome of his calculations will give you ESR but the datasheets most times gives ESR at 1KHz or below in the form of D or tan d.
But like all ESR meters it is usefull and the c measure is a nice extra. ( about ESR

The leakage tester makes no sense. Rp is not what you want to know, Rp is a mathematical thing. You can measure C in series or in parallel, this comes from the old bridges. They often measured Cp and G (as Rp) and that you could recalculate to Rs and Cs.
Leakage is an isolation fault of the oxidelayer and is measured at the working frequency
The norm is most times 0.001 x C( in uF) x V in uA (  about leakage
Title: Re: SMPS troubleshooting aid
Post by: dr_p on February 19, 2013, 10:21:28 pm
I think the ring tester is checking for shorted turns in a coil that would severely lower Q. I understand it sends a pulse into the coil and sees how long of a ring it gets. I attached another Bob Parker article that seem to be more clear.

As far as the leakage, the ESR meter pdf says
Measuring leakage current
1. connect the cap via 25 ohms to 5V
2. wait till the cap is charged to 5V
3. keep the cap connected to 5V via a 220kOhm resistor
4. measure the voltage drop across the 220kOhm resistor
5. calculate the leakage current and display it on the display (first line, right hand side)


Measuring the leakage current showed to be uncomplicated. But it's only makeshift because the lc
is always measured at 5V.
The lc should usually be measured at the maximum voltage of the capacitor.

So I understand from you wrote on your site that they do measure leakage correctly, but at 5V, so pretty useless, unless it's a 6V cap :-//.
Either way, I think that the bonus of capacitance measuring is justifying this build. Otherwise I would go for a Bob Parker style ESR meter.

Honestly, I'm just looking for a go/no go indication, rather than a quantification of the parameter (ESR, leakage, capacitance, Q). Something like "magic still seems to happen in that transformer", "electrolytic is in decent shape"  %-B :-DD
Title: Re: SMPS troubleshooting aid
Post by: PA4TIM on February 20, 2013, 12:17:47 am
Leakage is here indeed for 5V caps only.

Yes, it is a coarse way to measure Q, or rather compare Q, you find that methode on my side too, and also what the real Q was from the tested coils and that was way off but it is veryusefull if you have a spare transformer or choke. Both should give the same reading.
If you only have one, you do not know nothing because without knowing the inductance, frequency  current and resistance you do know what the Q should be. But if Parker tells it is true it probably will be ;-) But do not forget, it is made for flybacks in televisions, those run on selfresonance and this orobably will work as  the flybacks will problably all have a Q in the same ballpark.

But for other inuctors:
There is no such thing as an absolute High Q, like there is no such thing as an absolute high ESR. It is relative.
If Xc is 1 Ohm and ESR is 1 Ohm then D is 1, that is a really bad cap, if Xc is 100 Ohm and ESR is 1 Ohm then D is 0.01 and that is a good cap. But ESR is in both cases the same, and C can too, if F changes. If F is a contant then C is not the same, so 1 Ohm tells you nothing. You must calculate D from it and then you can say, good or bad. But if C is OK, ESR is low it is most tiimes Ok and if you calculate D a few times you get a feeling for what value is go/nogo.

About coils
If Xl is 100 Ohm and Rs is 1 Ohm then Q is 100, rather good, but if Xl is 1 Ohm and Rs is 1 Ohm, Q is 1, an inductor with a big loss. But this can be the same inductor on an other frequency or an other inductor on the same frequency, or the same inductor, on the same frequency with shorted windings and there for a lower selfinductance and lower Xl. But without a reference you do not know.

I needed a choke for a desing I'm working on. It had to be around 80 uH upto 500 KHz. The first one I made was 88 at 5KHz and 0,6 Ohm, but at 500KHz selfinductance dropped to 43 uH and Rs was 24 Ohm. ( core loss, not a shorted coil)  The final coil was 86 to 82 uH over this range and Rs from 0.8 to 2.8 Ohm.

Title: Re: SMPS troubleshooting aid
Post by: dr_p on February 20, 2013, 05:45:53 pm
thanks a lot  for your input and advice  :-+