Electronics > Beginners

Improving small unregulated power supplies

<< < (4/8) > >>

basinstreetdesign:
The biggest problem for me is the absence of ground out of the WW.  This has caused me problems in making small signal generators for the bench.  For instance I made a ovenized 10 MHz source a couple of years ago and used what I thought was a good WW to power it.  It did power it well since the design had 7X05/7X12 inboard.  The problem was that the output signal had some 60 Hz added to it that when connected to other equipment would bleed 60Hz and harms into my counter/scope/audio spec an. etc.
Now I am making a synthesized signal generator to replace it and will use a 3-wire grounded cord and a linear supply with Xfrmr to try to avoid that.
It bugs me that I cannot justify using up a spare WW even when I have one with suitable V/I output and I have about 25 lbs of the things in a box doing nothing. :palm:

soldar:

--- Quote from: james_s on September 26, 2019, 05:29:03 am --- Efficiency regulations is the reason the iron transformers have gone away. The small iron wall warts are not very efficient, only about 50% IIRC, they are "impedance protected" by having thin wire windings that cannot supply enough energy to catch fire if there is a short but this also leads to high losses and significant sag under load.
--- End quote ---

I do not think "efficiency regulations" have much, if anything, to do with it.  The losses in very small units, say under 5 or 10 W, are negligible. More powerful units get too big and heavy. A 50 or 100 W SMPS is very manageable but it would be too heavy with a mains transformer.

In the very small units range I believe the main reason is economic. SMPS are cheaper and provide stable regulated output. If you want to add a linear regulator to a conventional PSU then it gets too big, too heavy and too expensive.

Another point in favor of SMPS is wide range voltage input.

In summary, in the very small power range, say under 10 W, I believe market forces are the reason SMPS have replaced conventional PSUs and not any regulations (which I have not seen or heard of).

soldar:

--- Quote from: basinstreetdesign on September 26, 2019, 07:04:14 am --- The biggest problem for me is the absence of ground out of the WW. 
--- End quote ---

I do not understand this. Adding a ground to a floating PSU is trivial. Am I missing sumpting?

Gyro:

--- Quote from: Zero999 on September 24, 2019, 11:26:14 am ---
--- Quote from: Gyro on September 23, 2019, 09:32:25 pm ---
--- Quote from: Zero999 on September 23, 2019, 08:30:02 pm ---One issue I've found is the smoothing capacitor is almost always far too small. It's often something stupid like 2200μF in for a 1A unit, which will result in a huge ripple. I slap in a 4700μF capacitor before the regulator in anything run from one of those old transformer PSUs and the capacitor should be rated to 25V, for reliability's sake.

--- End quote ---

Oh yes, I place a premium on the ones that have case fixing screws rather than being ultrasonically welded shut. Quite a surprising number (of UK ones) do. Always worth checking when scanning ebay.

--- End quote ---
I'm talking about the unregulated type. No need to open it up, just put the capacitor on the PCB of the device you're designing.

--- End quote ---

I disagree, there's no point in losing an unregulated PSU brick just because its cheap overstressed internal electrolytic (is about to) spill its guts. You can put snubber caps on the rectifiers too - although I've found that most reputable ones already have them fitted.

Also, you have the opportunity to convert them to AC/AC, sometimes centre tapped.

Zero999:
Gyro, I see your point, but it's not worth my time or effort opening up a linear wall wart to replace the capacitor.


--- Quote from: basinstreetdesign on September 26, 2019, 07:04:14 am ---The biggest problem for me is the absence of ground out of the WW.  This has caused me problems in making small signal generators for the bench.  For instance I made a ovenized 10 MHz source a couple of years ago and used what I thought was a good WW to power it.  It did power it well since the design had 7X05/7X12 inboard.  The problem was that the output signal had some 60 Hz added to it that when connected to other equipment would bleed 60Hz and harms into my counter/scope/audio spec an. etc.
Now I am making a synthesized signal generator to replace it and will use a 3-wire grounded cord and a linear supply with Xfrmr to try to avoid that.
It bugs me that I cannot justify using up a spare WW even when I have one with suitable V/I output and I have about 25 lbs of the things in a box doing nothing. :palm:

--- End quote ---
Was it the transformer type or SMPS? Presumably the latter, in which case, it sounds like you've had problems with the leakage through the Y capacitor. This shouldn't be an issue for low impedance nodes, but can be a problem for low level high impedance stuff. The easiest way to resolve this is to connect the 0V of the circuit you're powering from the wall wart, to the other devices connected to its output. It does surprise me you had problems with a signal generator, because the wall wart's ground would normally go the the output ground, which would be connected to the device under test.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod