Reducing Ripple In Power Supplies.

Thread Starter

recklessrog

Joined May 23, 2013
985
There are frequent requests for information regarding regulation and smoothing in power supplies. Whilst answering a recent thread, I suggested a simple method for reducing ripple that was used a lot a couple of decades ago but seems to have been overlooked nowadays due to the use of voltage regulators and switch mode power supplies. I provided a link to a video that explains the method well ,so will post it again here...........................

https://www.youtube.com/watch?reload=9&reload=9&v=wopmEyZKnYo

This guy also clearly demonstrates many other useful explanations and is well worth following.
After my replies to the thread, I had a call from a member who said he had tried it without success. This prompted me to make a quick lashup to check for myself. Pls excuse the way I have shown it, but it was purely for a quick test
Here are some photo's that show how well it works, the transistor is a darlington pair with a beta of 80. the exact type is not critical and if you watch the video, you will see he describes different configurations using separate transistors. (Do ignore that he shows a couple of polarities in reverse on his drawing for input to output.)
You can put this circuit either before or after a series regulator, but remember, if placed after the regulator, there is a voltage drop across it so your voltage would be lower.

The output is 13 Volts with a 22 ohm load drawing 590 ma. (7.68 watts) Channel 2 is the ripple from the bridge rectifier, around 900mV Peak to Peak, whilst channel 1 is the filter output where the ripple is down in the random noise floor around 1 mV P.P. Some of the "hash" is from the inverter in the bench light.
The capacitor across the bridge rectifier is 4700 mfd, the capacitor to the transistor base is 470 mfd in series with a 4.7K resistor. The small capacitor is 100nF to prevent oscillation and reduce pickup of noise into the base.

When choosing series regulators like the 78XX series, due to differing manufacturing processes, not all seem equal. Many get much hotter than others in the same operating conditions, some do not reject ripple as well as the spec sheets would have you believe.In many instances, I have found that the LM317 is a better choice as with a couple of resistors, or a preset, you can set the output to any voltage within the confines of dissipation limits of VxA. Another very versatile device is the L200.
To protect regulators from reverse voltage damage if the input capacitor is discharged before any capacitor you have on its output, put a normally reverse biased diode across the input to the output pins. That way, should the condition arise, the current flows through the diode.

P1010048.JPG P1010050.JPG P1010003.JPG
 
Last edited:

Thread Starter

recklessrog

Joined May 23, 2013
985
Yeah, but I went to look briefly and didn't see one. How about at least a schematic for your rig as tested?
Link works fine here, When I get a minute I will scan a piccy of the one in my lashup but try the link again
 
Last edited:

sghioto

Joined Dec 31, 2017
2,363
Looks like an amplified smoothing capacitor.
Exactly! Also known as a capacitance multiplier. I remember that equation that stated the effective capacitance was equal to the actual beta of the output transistor under load multiplied by the value of the capacitor. Used that circuit many times to reduce hum in audio output circuits.
SG
 

DickCappels

Joined Aug 21, 2008
7,449
Back in the '70's I would hear this referred to as a "ripple clipper".

We used the to locally clean up the power supplies for noise sensitive curcuits in TV cameras where a few db difference in SNR could easily be a reason a sale was lost.
 

Tonyr1084

Joined Sep 24, 2015
6,000
Funny - - - I recently posted a question about my security cameras where I was connecting the power for those cameras all to one supply. In my videos I got shaded ripples on all screens as soon as I plugged more than one camera into the common supply. I built a standard filter using a 1000 µF and 0.33 µF cap placed across the output. It did nothing to change the issue I was having. Yesterday I came across that same video you posted. I plan on trying that as a solution to my issue. However, the way I resolved the issue was to buy a larger surge protected outlet, one that can house five wall warts, one for each camera. When my cameras are each on their own supply I get nice clear pictures. I still would like to see if this simple circuit (Cap Multiplier) can allow me to hook into a single supply and not get those shaded ripples. It WOULD be nice to have a single supply. However, it may be due to other issues such as ground loop problems interfering with the camera performance. It may continue to be necessary to have individual supplies. Like I said - it's funny you post this same video, the one I just found myself.

Here's my thread:
 

wayneh

Joined Sep 9, 2010
17,152
Here's an LTspice model of the "Ripple Clipper". Sorry it's a bit sloppy - LTspice is doing this weird font-size thing to me - but it works nicely and shows the amazing performance of this design. I used a TIP42 Darlington because I found a model of it at the LT wiki. Those files attached below.

The first question I had was how this compares to just using a regulator. The video partially addresses that by showing the poor ripple rejection of one LDO regulator. I'd have liked to have seen the same test applied to the typical 78xx regulator.

Nevertheless, a cool tool for the toolbox. I like the way the output voltage is buffered against quick changes.
 

Attachments

Last edited:

Thread Starter

recklessrog

Joined May 23, 2013
985
Here's an LTspice model of the "Ripple Clipper". Sorry it's a bit sloppy - LTspice is doing this weird font-size thing to me - but it works nicely and shows the amazing performance of this design. I used a TIP42 Darlington because I found a model of it at the LT wiki. Those files attached below.

The first question I had was how this compares to just using a regulator. The video partially addresses that by showing the poor ripple rejection of one LDO regulator. I'd have liked to have seen the same test applied to the typical 78xx regulator.

Nevertheless, a cool tool for the toolbox. I like the way the output voltage is buffered against quick changes.
Yes the buffering is a very useful Characteristic. I remember I used it once as a soft start for an audio amplifier to eliminate a nasty turn on thump from the speaker. In fact, anywhere a soft start is required, the time constant of the C/R to the base can be tailored to suit.
As a test today, I hooked up a halfwave rectifier to a 35volt rms torroidal transformer with 10,000 mfd main smoothing capacitor. I put 2 of the same darlingtons in parallel with a 0.5 Ohm w/w resistor in each emitter to prevent current hogging by just one of the darlingtons.
With a 44 Ohm load, (2x 22 Ohm 25 Watt in series) I fed the transformer from my variac, and at 40 Volts dc output from circuit, the ripple across the smoothing capacitor was nearly 1.5 Volts, and the output ripple was around 15 millivolts. Current was 0.90 amps, which equals 36.4 watts.
I then hooked it up with the same transformer as a full wave (bi-phase) rectifier, wound the variac until I had 50 Volts across the load. This was quite a test! The ripple across the main smoothing was now 1 volt Peak to peak and with the same load resistors, the ripple was less than 1milliVolt peak to peak. Current was now 1.37 amps which equals 56.8 Watts.
The resistor was 2.2K and the capacitor was again the same 470mfd one used in the first lash up.
I had mounted the two darlingtons on another heatsink the same as shown in my picture, and after 30 mins, they were only fairly warm to the touch although the load resistors were heating my workshop nicely :)
For some reason I can't your files, but may find something online that will.
Regarding standard regulators, my experience has shown that they vary a lot between different manufacturers. I had a batch of 7812's some years ago that were more like random noise generating oscillators! They were marked with a proprietary manufacturers logo and came from a reputable source. We tried everything to suppress them but in the end had to wait until another batch from a different supplier arrived.
I tested some 7805's and they showed large variations in ripple reduction, thermal transfer, and overload protection. I suspect that in most situations, they usually work well enough, and I wonder how many times users actually do more than just check the o/p voltage is correct. In non critical circuits, probably good enough, but if you have a critical circuit that seems to be acting strangely, 'scope the output from your regulator to at least eliminate it as the culprit.
 
Last edited:

wayneh

Joined Sep 9, 2010
17,152
Years ago I repurposed an old train set transformer to run an FM radio pulled out of a car I had. I used a 12V regulator on it. It works great on FM but I get wicked 60-cycle hum on the AM. I'm thinking I may revisit this old build and see if this trick helps. The noise may not be from the power supply, so I'm not holding my breath but I think it's worth a go.
 

Thread Starter

recklessrog

Joined May 23, 2013
985
Funny - - - I recently posted a question about my security cameras where I was connecting the power for those cameras all to one supply. In my videos I got shaded ripples on all screens as soon as I plugged more than one camera into the common supply. I built a standard filter using a 1000 µF and 0.33 µF cap placed across the output. It did nothing to change the issue I was having. Yesterday I came across that same video you posted. I plan on trying that as a solution to my issue. However, the way I resolved the issue was to buy a larger surge protected outlet, one that can house five wall warts, one for each camera. When my cameras are each on their own supply I get nice clear pictures. I still would like to see if this simple circuit (Cap Multiplier) can allow me to hook into a single supply and not get those shaded ripples. It WOULD be nice to have a single supply. However, it may be due to other issues such as ground loop problems interfering with the camera performance. It may continue to be necessary to have individual supplies. Like I said - it's funny you post this same video, the one I just found myself.

You can use ground loop isolators, this is from the CPC catalogue............
Here's my thread:
GB600


SR07971

DEFENDER SECURITY - HD CCTV BNC Ground Loop Isolator

The Video Ground Loop Isolator is a passive device that can reduce Ground Loop Interference with composite video signals. It is used in situations where video signals are being transmitted between points with different ground potentials. Unbalanced power l...
 

Thread Starter

recklessrog

Joined May 23, 2013
985
Here are some more pictures with using two daringtons in parallel and the torroidal transformer with 35-0-35 Volt secondary using half of the bridge as a Bi-phase rectifier.
Left hand meter is input to the filter, right hand is output volts.
The resistors in the emitters are to prevent current hogging by just one of the darlingtons.
Again this is just a lash-up to show the results.
P1010008.JPG P1010015.JPG P1010016.JPG P1010020.JPG
 
Top