eHam
eHam Forums => HomeBrew => Topic started by: N3KXZ on March 30, 2023, 01:36:27 PM
-
A radio related audio project I am working on requires 12VDC @ 1A. I thought I could just use one of the several cheap 12VDC switching power supplies I have on hand, but discovered that none of them are filtered good enough for this particular use case (I tried six different ones). Surprisingly with this project even a little ripple is heard at the output. Besides, if I am to use any of these supplies for future projects I will need to easily filter them properly.
Not wanting to go through the trouble of building a supply from scratch, I thought I could simply add some filtering caps on the output. Out of curiosity I hooked the scope to the supply output (no load) to see what difference some filtering caps would make. The answer is, with 2200uF cap for filtering: it makes only a little improvement. Ripple went from about 60mVp-p to under 10mVp-p. Additional capacitance brought minimal improvement.
I haven't actually tried this with a load yet. I want to ask you if there is an easy/better way to filter these cheap switching power supplies that will give me better results than what I have done so far. I don't currently have on hand any filtering sized inductors to experiment with, besides, I don't know enough about these switch mode supplies to confidently address the issue.
See attached photo.
Thanks,
Keith
N3KXZ
(https://card16.com/switching_supply_8666.jpg)
-
Running it through any sort of voltage regulator will
help a lot, as long as the regulated voltage is far
enough below the minimum voltage of the waveform.
In your case, low-dropout regulator (typical drop
less than 0.5V) may give you a fairly clean 11 VDC
output. If you can adjust the output voltage up a
bit, you may be able to use a 12V regulator, but
check the specs on voltage drop at full current.
I seem to remember some active smoothing circuits,
perhaps in Experimental Methods in RF Design. I think
they filtered the voltage applied to the base of a pass
transistor, but would have to go look it up to be sure.
-
Hmm,
Are you sure of your timing? If 10mS/div, then each cycle looks to be about 0.050s long which is 20Hz.
73
-
Mark N7EKU - Oops, yep I guess you are correct, I mis-calculated the frequency, though in this case I am not sure how it matters in regards to smoothing out the supply?
@WB6BYU - I don't happen to have any of those LDO regulators to play with. I did try a set of two 5.6V zeners in series (with a resistor) but that did not help, not sure why... maybe I would need two of those sets, one for positive going and one for negative going direction?
-
@WB6BYU - I don't happen to have any of those LDO regulators to play with. I did try a set of two 5.6V zeners in series (with a resistor) but that did not help, not sure why... maybe I would need two of those sets, one for positive going and one for negative going direction?
Where did you measure the voltage in your
test? The junction of the resistor and the
top of the upper zener should be a fairly
constant 11.2VDC in that case, assuming that
the minimum voltage of the waveform is high
enough (I can’t tell what the average voltage
is from the plots). But if you just stuck the
zeners + resistor across the power supply, it
won’t provide any filtering. And a zener
probably isn’t what you want for a 1A current
draw, since it has to pass through the resistor
to be regulated. Any resistor larger than
about half an ohm is going to drop the
voltage too much, and you will lose regulation.
And when properly implemented, the zeners
likely will dissipate a lot of heat, especially if
the current draw is intermittent.
Are you measuring the ripple voltage under
full-load conditions?
One thing you could do, of course, is to use
a much larger filter capacitor, like 30,000 uF
or so. But the supply may have problems
starting up with the capacitor attached, as
it will look like a short circuit until it charges
up.
You can make your own regulator using a
big enough PNP transistor (rated at several
amps) with the input voltage applied to
the emitter and output from the collector.
Then use an op amp to compare the
output (usually through a voltage divider)
to a reference (often a zener) and use that
to drive the base, do the base drive is
reduced when the voltage is too high.
That’s the principle of some of the low
dropout regulators, but they usually
include other protective circuitry. You
could build your own if you have the
parts handy, otherwise you may need
to start with 15+ V and a standard
7805-style regulator with sufficient
current capacity.
On the other hand, one could argue
that such an audio device should not
be sensitive to ripple on the power
input, but perhaps it cane with another
supply initially that was more quiet.
-
"And a zener
probably isn’t what you want for a 1A current
draw, since it has to pass through the resistor
to be regulated. Any resistor larger than
about half an ohm is going to drop the
voltage too much, and you will lose regulation."
Yea, I guess that is what is happening, and thank you for pointing that out. I was just playing around with parts I have on hand out of curiosity. I wasn't really wanting to spend the time building a supply at this time, I was just wondering if there was some easy quick fix that I wasn't aware of after being surprised that 2200uF was no where near enough filtering for an supply that is already supposed to be regulated DC... hah!
And yes, this may be a somewhat uncommon case where a small amount of ripple is evident in the output. I may want to go back through and check my ground connections and such. Or just live with it.
Thank you for helping,
Keith
N3KXZ
-
I would just repeat the filtering scheme of a high end audio unit power supply.
-
You might find the simple capacitance multiplier useful, which may be the configuration Dale was recalling. The image below shows the basic idea, which works well enough in many applications, with the value of C1 being effectively multiplied by the current gain of the transistor. The page from which the diagram is lifted (https://www.electronics-notes.com/articles/analogue_circuits/transistor/capacitance-multiplier-circuit.php) discusses some limitations and refinements.
(https://www.electronics-notes.com/images/transistor-capacitance-multiplier-01.svg)
These days there are a variety of low-noise low-dropout regulators and the like which work well. However, to get some filtering into a smallish space using a cheap transistor(s) the capacitance multiplier is quite useful when used with a decent regulator.
73, Peter.
-
That sharp leading waveform.....
That leads me to believe that as you load that down, you are going to start to see more and more squaring of the lagging end of the waveform. PWM
Switching power supply. Square waves.
Your fix is a common mode choke at the output of the "transformer".
No amount of capacitance is going to fix this.
--Shane
WP2ASS / ex KD6VXI
-
Your fix is a common mode choke at the output of the "transformer".
Good idea.
Do you have a suggestion for a small common
mode choke for audio frequencies that will
handle 1A?
Might be able to rewind a power transformer...
-
For wall wart noise you can check these:
https://palomar-engineers.com/ferrite-products/ferrite-cores/Wall-Wart-RFI-Noise-Filter-Bulk-Pack-of-10-Filters-p90491965 (https://palomar-engineers.com/ferrite-products/ferrite-cores/Wall-Wart-RFI-Noise-Filter-Bulk-Pack-of-10-Filters-p90491965)
For noise suppression sometimes, one very large capacitor is not as good as one large one and one small one. You might try adding a 0.1 or 0.01 uF cap too. But as mentioned, the go-to method for this type of noise is to wrap the DC line many times around the apprppriate toroid core.
73
-
For wall wart noise you can check these:
https://palomar-engineers.com/ferrite-products/ferrite-cores/Wall-Wart-RFI-Noise-Filter-Bulk-Pack-of-10-Filters-p90491965 (https://palomar-engineers.com/ferrite-products/ferrite-cores/Wall-Wart-RFI-Noise-Filter-Bulk-Pack-of-10-Filters-p90491965)
Only 500 ohms impedance at 1 MHz, so
unlikely to help at the audio frequencies
that he is trying to filter. Going to need
a much higher permeability core, and
likely more turns of wire.
-
For wall wart noise you can check these:
https://palomar-engineers.com/ferrite-products/ferrite-cores/Wall-Wart-RFI-Noise-Filter-Bulk-Pack-of-10-Filters-p90491965 (https://palomar-engineers.com/ferrite-products/ferrite-cores/Wall-Wart-RFI-Noise-Filter-Bulk-Pack-of-10-Filters-p90491965)
Only 500 ohms impedance at 1 MHz, so
unlikely to help at the audio frequencies
that he is trying to filter. Going to need
a much higher permeability core, and
likely more turns of wire.
I suspect that the OP doesn't have the frequency correct,
N3KXZ, was your scope really set to 10 ms/div? Normally SMPS are running somewhere up in the 40kHz region or higher.
73
-
From Pacific Antenna
Description
The Wall Wart Tamer lets you utilize those surplus computer and wall pack power supplies as a adjustable sources of clean DC power for radios and equipment.
The Wall Wart Tamer is a simple device designed to filter input power and remove most of the receiver “hash” noise that results from common wall warts and laptop supplies, especially switching supply types.
The project itself is straightforward and if you follow the steps you will have a working Wall Wart Tamer module putting out CLEAN DC power when you finish.
https://qrpkits.com/wallwarttamer.html (https://qrpkits.com/wallwarttamer.html)
-
@ KG4RL - I saw that wallwarttamer ... I would build my own supply before paying $20 + shipping to build a fix for a wallwart... wouldn't you?
@ Mark N7EKU - Yes, I ran the test again and the scope was set to 10 ms/div. I agree that seems weird. I tried several different wallwarts and got different forms of noise (some which looked like high frequency ringing) but all of them (except one) had significant levels of noise that could be heard in my device. My current junkbox is limited, but I did try a few toroids as chokes with no luck... this is an area I want to pursue further as I don't have much experience working with them. Also, my scope (and skill level) isn't the best.
@ WB6BYU - yea, filtering out such low frequencies is kinda like building a speaker crossover with those huge coils!
@ VK6HP - thanks for that circuit on the simple capacitance multiplier and the link! I looked that over and will try it sometime when I have more time to experiment.
Just for context, my bench supply worked fine in this application with no noise passing through to my device. And I did finally find one wallwart that works OK with almost no ripple and does not cause issues with my device- it's rated for 3A and I guess is just better made.
Thanks to everyone for your time and input! I need to study common mode chokes, and not rely on wallwarts so much ...
Keith
N3KXZ
-
Hi Keith,
What is the make and model of your scope? What are you using for a probe on that? To get 20Hz waveforms like that on a wallwart is too strange! Switching types will give much higher frequency noise, and linear types can give 120 or 60Hz noise. Maybe it is some kind of behaviour happening when it is not loaded. I would load it down with a resistor and see what happens.
Be careful of wall warts. Some of the older types may be unregulated linear types. The output voltage stated is only the voltage as a result of loading, so the unloaded or very lightly loaded voltage can be quite a bit higher than listed. Also they can sometime be center negative on the plug.
Are you experimenting with DC receivers? These can be susceptable to PS noise.
73
-
@ KG4RUL - I saw that wallwarttamer ... I would build my own supply before paying $20 + shipping to build a fix for a wallwart... wouldn't you? ...
Keith
N3KXZ
Actually, I bought one along with a few other items from them.