Electronics > Projects, Designs, and Technical Stuff
Wide input and output range power pre-regulator (85–265 Vac / 2.5–52.5Vdc / 5A)
<< < (3/7) > >>
Sylvi:
Hi Prasimix

What a project ! :D

I've had only a little bit of experience using SMPS modules from meanwell and recom. Both put out a lot of noise and the AC-DC converters are the worst. Although radiated noise is practically zero the conducted noise is pervasive and goes everywhere. I did not see this with my old CRT scope but recently I got an OWON XDS21012A DSO and the noise is clearly visible. I was shocked how much there was.

A friend was using these modules, too, but his assembly was modular with the PSU on its own board. For him, changing the layout, adding more chokes and filter caps, did not cost too much per board revision, so eventually we got his sytem to be very quiet. He made some measurements using a computer-based real-time analyser and got things to where the RTA was showing no apparent noise in the audio path (our application is for audio). The DSOs seem very sensitive to how the leads are grounded, so he did not do much inspection with his DSO since everything seemed noisy through it. At first i had a similar problem using the DSO, but fixed a ground on my test bench and then saw clean where it should be clean and SMPS noise wherever it was present.

Anyway, I abandoned the SMPS for now and I'm looking at an alternative high-frequency PT but using linear instead of switching methods. My friend found a supplier at Aliexpress selling resonant converters suitable to power a solid-state power amp. They used the term "zero voltage switching" interchangeably with "resonant". maybe you cannot have rhe one without the other?

The noise output from your supply is impressive. Do you use any kind of soft switching? or do you think the low-noise is the result of better filtering than Meanwell uses?

Sylvi:
Hi again

I also wonder about the use of half-wave rectification on the secondaries. Isn't it better to use a bridge? ordoeis the fact of the "square" output negate the pitfalls of half-wave?
prasimix:

--- Quote from: Sylvi on March 06, 2019, 08:07:12 pm ---My friend found a supplier at Aliexpress selling resonant converters suitable to power a solid-state power amp. They used the term "zero voltage switching" interchangeably with "resonant". maybe you cannot have rhe one without the other?

The noise output from your supply is impressive. Do you use any kind of soft switching? or do you think the low-noise is the result of better filtering than Meanwell uses?

--- End quote ---

I'm also new in this area, and this is my first SMPS project, but learned a lot and got some "first hand" experience. I've learned, for example, a hard way how PCB layouting is of paramount importance. I believe that even better results can be achieved if wisely/masterfully routed 4-layer PCB is used.
As far as I know resonant are usually ZVS or ZCS. But, it is possible to achieve soft-switching with non-resonant topologies. In that case we can talk about ZVT and ZCT (that some still calling ZVS and ZCS).
Currently I'm not using neither ZVT or ZCT in CF-DIC, but that could be possible (and add another level of complexity). Its schematics is publicly available on GitHub and on the project site where you can check what I have on the output. I cannot talk about Mean Well since I never try to reverse engineering it. Take also into account that additional QR flyback is also deployed and its also to some extent participate in output noise (if the power output is fixed it can be removed, but that is not the case for pre-regulator).

Please note that you cannot employ resonant topology for such wide output range, and that is the primary requirement for the power pre-regulation :). If you are aware of such resonant solution that is reasonably complex please let me know. When you add to this requirement another one, to have "auto-switcher" for input voltage (i.e. wide input voltage range), than it become a pretty challenging mission.

I also believe that is possible to reach such level of noise reduction on SMPS output that is can be used at least for powering audio circuits if not also RF appliances. There is an increasing number of "D-class" (and similar) Hi-Fi amplifiers that works pretty well even for demanding audio users. Possibly that will become one of my next "mission impossible" project ;).
prasimix:

--- Quote from: Sylvi on March 07, 2019, 04:03:48 pm ---Hi again

I also wonder about the use of half-wave rectification on the secondaries. Isn't it better to use a bridge? ordoeis the fact of the "square" output negate the pitfalls of half-wave?

--- End quote ---

The CF-DIC output is full-wave not half-wave. Additionally a synchronous rectifier is used to further improve efficiency.
Sylvi:
Hi Prasimix

The schematic PDF with TR1 shows four secondaries and all of them are half-wave rectified as far as I see - one diode each. Maybe this is just the auxiliary supplies, not the main output?

Interesting about the resonant converters being more optimal for constant loads. I wondered about that. My initial foray with SMPS modules was an audio preamp, so the load is pretty constant and known, so should be an easy load.

Class-D is still very immature and suitable for subwoofers and not much else. It is suspicious when you see THD specs that the test frequency is often very odd - not the standard 1kHz or 20kHz, which strikes me as trying to get a good figure at some "null point" of frequencies between the switching PSU and the switching amplifier :) It is ironic that digital circuitry has a good noise immunity level to supply noise, considering most computers operate at quite low voltages. True linear audio circuitry can be forgiving, too, inasmuch as you might not hear the switching noise but you can see it on the scope.

Over the past few years, my friend has used SMPS modules to power mic preamps and headphone amps. The headphone amp was where we had to apply the first "bandaid". The circuit would power up but when he tried to run a signal through it the SMPS would shut down. We added an impedance converter to the supply then everything was happy. Linear wants a low-z supply with wide frequency response and the output of the SMPS is constantly changing and high-ish. The z-converter is just an active hum filter circuit with a current clamp.

Also, when you use a master AC:DC converter then use secondary converters for other voltages, powered from the master converter, there is a whole mess of noise harmonics with the interaction between them. My friend successfully got his latest equipment audibly quiet, and as far as the limited equipment he has shows, but he is still uneasy about the stability of interaction of the four converters needed for some of his audio boards. he does in fact have a DSO - bought one after I got mine - but encountered the same sensitivity to RF noise and didn't try to fix it. I fixed the grounding on my bench and still suspect he would see noise in his gear with the DSO.

What are you powering with your new PSU?
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod