-
I am putting together a power supply with several outputs (so far 3.3, 5, 12, and adjustable voltage) using a dc buck convertor for each voltage. I will be running them at 2 amps max with the exception of the adjustable one.
One is --> http://www.aliexpress.com/item/Wholesale-1pcs-DC-Step-Down-Converter-DC-4-0-38V-to-1-25V-36V-5A-75W/32285636716.html
It has been running very well for several days now.
The picture shows the output from the laptop ps (19v and about .15 amps) and the module running at 12v at 2 amps (the spikes are higher at 3.3v 2.0a) The vert is 100 v/div and the hor is 2 usec/div. The freq is 174k.
Maybe it would help the outputs from the modules if I can reduce the ripple from the Toshiba 90 watt laptop power supply. Any suggestions on a simple fix? Remember the module costs $6.83 so the fix needs to be inline with this (ie less than $6). Maybe a 317 or a cap ??
thanks
-
Looks like most of it is ringing on the measured input from the switching of the converter.
When the output ramps up (converter turned on) the voltage of the input drops, when it ramps down (converter turns off) it jumps up.
So the "174kHz" signal you're worrying about is actually induced by your converter itself, whereas the much higher frequency ringing is something you could much more easily battle.
One thing you can do is have a capacitor bank close to your converters that they can suck some power out, allowing the PSU to keep operating closer to its desired voltage on switching peaks, but it's very likely not going to help your output one bit. Possibly only improve total or DC/DC efficiency slightly. Or if the bank is too large for the PSU too handle, worsen the system efficiency (although I'm thinking at $6 with a Toshiba PSU you're not going to achieve too large very easily).
EDIT: (Although there seems to be some intermittent ringing that might also come from the PSU itself, but that doesn't seem to affect your output all that much, which is to be expected in a loaded, regulated buck converter switching at far below ringing frequencies, really)
-
Quote
Looks like most of it is ringing on the measured input from the switching of the converter.
I did not think of this. The attached pic is just the laptop ps running at 2.35 amps into a 8 ohm resistor. What a difference but to be honest I do not know if better or worse.
V = 200mV/div H = 1 uS/div
-
Much better.
This is high frequency blabber on the voltage. That could be filtered out with affordable components if you wanted to.
But since the switching frequency of the DC/DC is 170kHz-ish, it'll have little to no influence.
Why?
If the input and output caps and coils of the DC/DC can filter the buck dumps at 170kHz, it is very unlikely any of the ringing will make it through to the output in any peak-to-peak value significant compared to the DC/DC's own generated ripple.
The little spikes you saw on your output earlier are ringing generated by the buck'ers. Remember, you're not paying a shitload for the bucks, so they are going to have okay performance likely, but surely not prime performance. -
thanksQuote
but surely not prime performance
I know but I want to see how good I can make the ripple using some low cost methods. I am waiting on some 317 modules and will try placing them before and after the modules and study the performance vs power lost using them. And in the meantime keep reading Chap 9 from Art of Electronics 3rd Ed, to see if there are any other methods to use.
thanks again
-
Laptop PSUs are not precision regulated and filtered, since the voltage gets stepped down by buck converters on the motherboard anyway.
-
I think the answer is here somewhere
https://e2e.ti.com/support/power_management/simple_switcher/w/simple_switcher_wiki/2243.understanding-measuring-and-reducing-output-voltage-ripple
It seems using 47uF caps in parallel along with some xx pf ceramic caps.
I just have to figure out this:
"Choose a ceramic capacitor that has an impedance null (self resonance) at the ring frequency or higher."
oh boy
-
I know but I want to see how good I can make the ripple using some low cost methods. I am waiting on some 317 modules and will try placing them before and after the modules and study the performance vs power lost using them. And in the meantime keep reading Chap 9 from Art of Electronics 3rd Ed, to see if there are any other methods to use.
Great to see you experimenting to find the answers on your own! Good on you for ordering the 317's.
You can get an rough preview of the results by looking at this graph in the 317 datasheet:
This shows that the 317 is great at filtering noise up to about 10kHz, and then performance goes down as the frequency goes higher.
Another filter to consider is an RC low pass filter. However, you'll quickly realize that these are unsuitable for any decent amount of current (the resistor has to get smaller and smaller to let enough current through, which means the caps have to get bigger and bigger ($$$)).
The next filter to consider is an LC low pass filter. This filter is much better suited for passing larger amounts of current. If you exceed the saturation current of the inductor, is becomes much less effective as a filter, so be sure to check that value in the datasheet. When choosing a cap, the lower the ESR, the more effective they will be in the filter. Also, be sure to take a glance at how much ripple current the caps are rated for.
Another filter you might check out is a capacitance multiplier low pass filter. However, for larger currents the transistor (and heatsink?) might become more expensive than an LC filter.
You might also experiment with come ferrite beads for the really high speed ringing. Again, you'll have to keep an eye out for the saturation current spec for the ferrite beads (you might need to parallel several of them).
I recently fooled around with both an LC and a capacitance multiplier filter when trying to filter the noise from an ICL7660 switched capacitor voltage converter. https://www.eevblog.com/forum/projects/an-evening-with-the-icl7660/msg675242/#msg675242
-
"Choose a ceramic capacitor that has an impedance null (self resonance) at the ring frequency or higher."
So, in theory, a low pass filter's performance gets better and better as you go up in frequency.
Unfortunately, that's only true for "ideal" components, which don't exist in the real world.
If you search for "ceramic capacitor self resonance" as a google image search, you'll see results like this:
The straight lines in that graph represent what the performance would be for ideal components, and the lines which dip and then go back up are the kind of performance you can get with real-world components.
The lessons here are that ceramic caps have a self-resonant frequency, and right at that frequency their performance gets a little bit better (the dip), but above that frequency their performance starts getting worse and worse (the lines start curving back up). Also, you can see from the graph that difference sized capacitors have different self resonant frequencies (and the larger the cap, the lower the frequency).
In EEVBlog #343 ( ) Dave tests out an LC filter using a spectrum analyzer and tracking generator. There, he demonstrates this effect. The filter works great up until about 75 MHz, then starts getting less effective as the frequency gets higher.
So the reason they say to pick a cap which has a resonant frequency right where you want to filter is that you get a slight boost in performance right at that frequency. But "don't best be the enemy of good" (don't let not understanding that slow you down. Order some 1uF and 0.1uF ceramic caps and try them out!).
(I should have also mentioned that if you are going to try an LC filter, you'll find that small toroid inductors from china on ebay are ridiculously cheap, especially compared to more trustworthy parts found on digikey. Because your switcher is only $6, I'd recommend buying a handful of 100uH toroid inductors rated for a few amps from ebay and fool around with them (and then follow them with a 317). e.g. http://www.ebay.com/sch/i.html?_from=R40&_sacat=0&_nkw=toroid+inductor+100uh&rt=nc&LH_BIN=1 )
-
See also this post and this app note.
https://www.eevblog.com/forum/projects/ferrites-in-power-supply-good-or-bad-how-to-reduce-noise-on-vcc/
file:///Users/cell/Downloads/5582.National_Input_Filter_Design_for_Switchers.pdf