Recognise this trace?

Do you have a question? Post it now! No Registration Necessary

Translate This Thread From English to

Threaded View
Gentlemen,

The link below shows a scope trace of the output of a linear PSU (voltage  
4.7VDC) with about 300mV of 100Hz ripple riding on it. This PSU is in-
circuit under load from the boards it supplies. Normally I would assume a  
filter capacitor to be at fault here, but they all check out fine so  
something else is causing this ripple. Note there is a characteristic  
'knee' on the high peaks and I'm thinking this must be indicative of  
*something* trouble is, I don't know what. If anyone recognises this  
waveshape and knows what causes it, that'd be "awesome" - as our American  
friends describe everything. Check it out:


https://www.flickr.com/photos/128859641@N02/35684682551/in/dateposted-
public/

Re: Recognise this trace?
On Sunday, July 9, 2017 at 5:14:43 PM UTC+2, Cursitor Doom wrote:
Quoted text here. Click to load it
  
Quoted text here. Click to load it
  
Quoted text here. Click to load it
  
Quoted text here. Click to load it

What's odd about it? The filter capacitor is discharging for about 6msec (w
hile the half-sine from the bridge rectifier is less than voltage across th
e capacitor), and charging for the 4msec when it is higher. The "knee" is m
ostly likely the internal resistance of the capacitor, which adds on top of
 the capacitor voltage when the capacitor is charging, and subtracts from i
t when the capacitor is discharging.

If that was all it was you'd expect something similar on the low point of t
he rippe, but if you throw in a little leakage inductance in the transforme
r the turn-on is gradual and the corresponding "knee" gets smoothed out whi
le the charging current at the turn-off hangs on for a bit.

I can't be bothered plugging this into LSpice - it's a question from Cursit
or Doom after all - but it feels plausible.

--  
Bill Sloman, Sydney

Re: Recognise this trace?
On Sun, 9 Jul 2017 15:10:46 -0000 (UTC), Cursitor Doom

Quoted text here. Click to load it

Surely, you jest.

Quoted text here. Click to load it

Is the mystery load the hardware appearing in adjacent photos?  EXIF
says that they were taken 4 days earlier, which suggests a connection.

These photos looks odd:
<https://www.flickr.com/photos/128859641@N02/34892442244/in/photostream/
<https://www.flickr.com/photos/128859641@N02/35564536392/in/photostream/
because C17 (the 2nd blue electrolytic from the right) seems to have
been soldered into the wrong PCB holes.  The white silk screen circle
makes a good target for installing capacitors.  If my crystal ball is
correct, this capacitor might be presenting a higher than normal
current load to the power supply, which is reacting by producing
excessive ripple that you're seeing on the scope.

Is this the capacitor you removed from the board?
<https://www.flickr.com/photos/128859641@N02/35050368241/in/photostream/
If so, you ripped out the plated through hole.  4 terminal radial caps
are not common, so I guess you substituted something else.  What did
you put in its place?  Did you test and fix the PCB traces?  

Oh, never mind.  I found your rather ugly solder wick PCB fix:
<https://www.flickr.com/photos/128859641@N02/35602405551/in/photostream/
You might want to double (or triple) check your soldering and PCB
through hole connections.

From this photo:
<https://www.flickr.com/photos/128859641@N02/35564539582/in/photostream/
there's a tuning screw, semi-rigid coax cable, and cavity filter on
the left side.  Too bad the power supply board is missing.  There are
some gold connector pins in the photo.  So, what is it?  


--  
Jeff Liebermann     snipped-for-privacy@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com
We've slightly trimmed the long signature. Click to see the full one.
Re: Recognise this trace?
On Sun, 09 Jul 2017 10:20:47 -0700, Jeff Liebermann wrote:

Quoted text here. Click to load it

If the adjacent photos were of any relevance, Jeff, I'd have mentioned it  
and linked to them.


Quoted text here. Click to load it

All I did was replace an electro with one of identical parameters; it  
just happened to be a bit smaller than the original (which is not at all  
uncommon).

Quoted text here. Click to load it

You're complicating this unduly and massively, Jeff. I simply posed a  
question about an oscilloscope trace, that's all.

Re: Recognise this trace?
If it's your signal generator thing, yes, it's very possible / likely to  
have dried out caps (excess ESR).

Yes, the nippley appearance is likely ESR.  But, if you say it's okay, then  
it must be okay.  Right?

Tim

--  
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Contract Design
We've slightly trimmed the long signature. Click to see the full one.
Re: Recognise this trace?
On Sun, 09 Jul 2017 13:24:04 -0500, Tim Williams wrote:

Quoted text here. Click to load it

Do I detect a note of scepticism here, Tim?
I've checked all the electrolytics with my ESR/Capacitance tester which  
is claimed to be able to test for ESR in-circuit so I didn't remove any  
of the caps in order to test them. I'm relying on the meter  
manufacturer's claims that it's not necessary to do so and in the absence  
of any evidence to the contrary, I don't see what more I can do.

Re: Recognise this trace?
Cursitor Doom wrote:

------------------------
Quoted text here. Click to load it

** Electro cap ESR meters can be relied on to work normally with the caps in circuit.  

Diodes and transformer windings in parallel have no effect.  


....  Phil  

Re: Recognise this trace?
On 7/9/2017 8:10 AM, Cursitor Doom wrote:
Quoted text here. Click to load it
You might get this when you have two caps separated by an inductor.
First cap usually gets more peak charge current than the second one
and develops high ESR first.  Even if both caps are good, the inductor
(or resistor) in between can cause this effect.

Re: Recognise this trace?
On Sun, 09 Jul 2017 12:34:51 -0700, mike wrote:

Quoted text here. Click to load it

Someone on SER said a non-linear load could explain the shape of that  
trace and my initial reaction was that sounded quite plausible. However,  
on reflection I think if that *were* the case, then the knee would span  
600-700mV (when the semis start conducting) and that isn't what we see  
here.
The obvious thing to do next is whip off the output plugs from the PSU  
one by one and see if that clears the problem and if so, which sub-
circuit is responsible for the overload as it may be nothing to do with  
the PSU at all.

--  
This message may be freely reproduced without limit or charge only via  
the Usenet protocol. Reproduction in whole or part through other  
We've slightly trimmed the long signature. Click to see the full one.
Re: Recognise this trace?
On 9.7.17 23:44, Cursitor Doom wrote:
Quoted text here. Click to load it


The peak is definitely from the charge pulse from the rectifier. A
capacitor-input PSU charges only 5 to 10 % of the half-cycle time,
depending on the resistances and inductances in the charge path.

--  

-TV


Re: Recognise this trace?
Cursitor Doom wrote:
Quoted text here. Click to load it
   Cause: stored charge in the rectifiers.


Re: Recognise this trace?
Robert Baer wrote:

----------------------
Quoted text here. Click to load it


** Bollocks.  


..... Phil  

Re: Recognise this trace?
wrote:

Quoted text here. Click to load it

  a soft diode in the bridge rectifier?

--  
Using Opera's mail client: http://www.opera.com/mail/

Re: Recognise this trace?
On Sun, 09 Jul 2017 19:37:39 -0400, tom wrote:
-
Quoted text here. Click to load it

When I tested this psu out-of-circuit with a dummy load I got less than  
5mV of ripple on that rail. Now I may have too lightly-loaded the output,  
I can't say for sure, but that does tend to indicate to me I should check  
for shorted components down the line from the psu before I do anything  
else.

Re: Recognise this trace?
Cursitor Doom wrote:

Quoted text here. Click to load it



** Normal waveform for a supply frequency, unregulated DC supply.  

  Scope a few others and see how dumb you are being.  



.... Phil  

Re: Recognise this trace?
On Sunday, July 9, 2017 at 11:14:43 AM UTC-4, Cursitor Doom wrote:
Quoted text here. Click to load it

Regulator is not working. It's broken, doesn't have enough headroom, something wrong with its feedback path, reference circuit is corrupted, something is broken.

Re: Recognise this trace?
On Mon, 10 Jul 2017 06:41:45 -0700, bloggs.fredbloggs.fred wrote:

Quoted text here. Click to load it

Yup, that should cover it.

Re: Recognise this trace?
On 7/10/2017 12:33 PM, Cursitor Doom wrote:
Quoted text here. Click to load it

Well, just "something is broken" would cover it.

Re: Recognise this trace?
wrote:

Quoted text here. Click to load it

Or broken as designed.

Re: Recognise this trace?
On Sunday, 9 July 2017 16:14:43 UTC+1, Cursitor Doom  wrote:
Quoted text here. Click to load it

What's the problem with it? Why do you think it ought to be behaving any different? It looks like a standard unregulated 50Hz PSU.


NT

Site Timeline