Scoping a SMPS primary side...

The best place to connect the ground of your 'scope for primary side monitoring, is the negative terminal of the main primary side smoothing cap. But please, please be very careful if you're not an experienced engineer. I know you say that you're using an isolation transformer, but the belief that this makes you totally safe against electrocution, can lead to complacency. If you manage to get the mains across you, even if it's being supplied by an isolation transformer, this can still at the very least give you a very nasty shock, and under extreme conditions, result in death. Also be aware that under certain circumstances, with SMPS's, the mains earth on your scope can be an issue.

If you have startup voltage, but no chopper drive output from the '3842, suspect first any small electrolytic caps associated with the chip, particularly the one connected to the supply pin. It's also not at all uncommon for the chip itself to fail.

BE CAREFUL - Keep one hand in your pocket, whilst taking readings.

Arfa

Look here for the datasheet on the ic. The first shows ac waveforms. The last one is a 138 page manual on PWM switch mode power supplies. Sky.

signals

on

cap.

I

that

complacency.

an

scope

Yeah, isolation transformer or not, it's still lethal amps. I don't like to probe with the line voltage on. I'll power it off, hook things up, stand back and slowly bring up the power. I checked the small electrolytic cap on the supply pin and it checks fine. I have no reference voltage from the chip either. The symptoms were longer and longer startup times which has me thinking that although the esr reading on that cap is ok, it may still be dead. Might as well replace it first and see what happens. I don't get a tweet, tweet, tweet of a cycling supply. I hate just swapping parts however it might be the easiest thing to do...

I know that working around things like this can lead to accidents. Line voltage is nothing to play with.

I like reading the electronics repair faq as well. Lots of good info and constant warnings about discharging caps, using isolation transformers, one hand. A little reminding doesn't hurt, either...

Thanks for the info!

Kirk S.

BTW, you better power your circuit, not the scope from that isolation transformer. That will isolate your entire construction from the hot wire thus making your work much safer.

--
******************************************************************
*  KSI@home    KOI8 Net  < >  The impossible we do immediately.  *
*  Las Vegas   NV, USA   < >  Miracles require 24-hour notice.   *
******************************************************************

I've worked on these a lot - I don't use the scope anymore. When I did, I made sure the ground of the scope was floating - use one of those 2 prong to 3 prong adapters to isolate the ground. BUT - since the IC isn't working - there really aren't any signals to see. The +16v is good - suppy voltage. No ref voltage says the chip is shut down. Also, if memory is good, the

16v has to go up over 18v to turn the chip on. It gets this via feedback from the flyback [usually]. No feedback voltage - chip doesn't turn on. Check all the diodes around the 3842 chip. And all the other electrolytic caps around the 3842 too.

The next thing I'd check would be all the rectifying diodes on the secondary of the power supply circuit. If one of those is bad - no low voltages - no flyback voltage - chip doesn't turn on. Then I'd check the horizontal output transistor.

There were a couple of times I actually had to replace the 3842; but, usually it was something else that caused the 3842 to shut down. They used to be pretty cheap at Dalbani; but, it looks like they've gone out of business....????

later, The Schneids

The Schneids wrote in news: snipped-for-privacy@corp.supernews.com:

The WRONG way. You're supposed to isolate the DUT,not the test equipment.

Leaving the scope chassis potentially HOT,a clear hazard.

--
Jim Yanik
jyanik
at
kua.net

Interesting to see that you run your 'scope with the ground disconnected. I do too, but hesitated to suggest that the OP did this, because unless you're well experienced, as I am, and clearly you are too, it can be a dangerous practice.

I actually have the ground wire disconnected from inside the plug, and brought out through the crack between the two halves of the plug, so that it can be clearly seen to be disconnected, and I also have a Dymo label on the plug top saying " Caution. No Earth "

One day, we're both gonna fry .... |:-(

Arfa

WARNING... DANGER.... what is the point of floating the AC ground on the scope, it is not only dangerous, but totally unneccessary as long as the equipment that you are troubleshooting is being powered via an isolation transformer..... that is the whole reason to use an isolation transformer so that things like a hot chassis television power supply can be easily hooked up to properly grounded test equipment without making sparks fly and exposing yourself to unneccessary hot ground hazards... keep the safety ground (3 prong plug) connected on your test gear... in a large shop or manufacturing and testing facility that has accountability to OSHA, disconnecting the safety ground for any reason is a flagrant and dangerous violation. electricitym . . .

I agree - isolation is best, but if the 'scope has a 3-wire power cord, with a properly grounded chassis, and the outlet is properly grounded, should you not be able to just use the positive probe of the 'scope to look at the primary, and not even connect the test lead ground wire to anything?

Mark Z.

"Mark D. Zacharias" wrote in news:Wxr3f.865$ snipped-for-privacy@newssvr29.news.prodigy.net:

Not for the primary side of switchers.There is no "ground" for those,the low side is a "common",and not connected to true earth ground,and it can be elevated up to -170VDC .Some switchers use a doubler arrangement for 120VAC operation and reconfigure for 240VAC operation.

Tektronix used to have a great booklet about isolation and why certain methods like "ground-busting" should not be used.I used to include one with every scope that I serviced that arrived with no ground pin on the plug. Maybe they still have it,I do not know.

--
Jim Yanik
jyanik
at
kua.net

Since the equipment under test is being powered via an isolation transformer you should ground the scope probe ground to the negative side in the hot chassis in order to obtain proper waveform measurements.... no problem, no sparks, not an issue. Obviously voltage and current are still present and standard personal safety proceedures are still required, but because the scope is safeyy grounded there are not hot chassis issues since they are floating thanks to the isolation transformer. electricitym

All correct. An isolation transformer is always a good part of your personal safety.

The UC384x type supplies I hadto repair have had these usual sympthoms:

1. Dead UC 3842. Change.
2. Freid resistor at chip output (in the connection to the gate of the FET). Replace.
3. Dead FET. Replace.
4. Capacitor about Ucc of UC384x: Leaky or lost capacity. This is often the "twit..twit...twit" sound of the startup.

google for the article of Keith Lofstrom (or similar , may be my memory is not correctly remebering): "fixin scitching poer supplies".

hth, Andreas

Hi all,

Thanks for the help...

I hooked the monitor up to the isolation transformer and variac, brought up the voltage slowly while watching the VCC with the scope hooked to the (-) of the large filter cap and the lead hooked to VCC on the 3842. Voltage climbs to +17vdc and sits. If the chip was working, I would see some sort of drop when the oscillator started running. Nothing on any of the output lines either. All diodes around the IC test ok. All diodes on secondary test ok.

Removed the 3842, put in strip sockets, installed new 3842 and the montor fired up. Used the light bulb in place of fuse and as the monitor would cycle up, the light bulb would get bright as the filter cap charged and then settle down when the monitor came up.

The original was a KA3842A and I replaced it with a UC3842B. They both cross to the same NTE and from the little I can garner from the specs, the B is a higher tolerance version of the A.