Switched mode power supplies

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Whats the current status on SMPS ?

During the last couple of weeks ive read a few comments like "no big deal" and "ive made several that works fine".

I need a lot of power, like 40A 12 to 30V.

Anybody made something like that that actually works?

AND with a reasonable reliability and expected life to it???

Circlotron you mentioned that SMPS is a nogo no matter what, since i know theyre part of your profession, id like to have a short explanation why it is so.

Magura
 
I did an off-line, 30V/1kw supply for a customer. It was a 200khz, full bridge forward converter. Full bridge is a bit complicated, but the transformer was so small. I used a small PQ transformer with a center-tapped foil secondary. The primary was wound with triple insulated wire - great for meeting safety requirements. Rubadue, Furukawa and others make this kind of wire.

:eek:
 
I can get it to you in a few days. Yes, a bit expensive. It was part of a much bigger system: 1kW PFC front end, four independent full bridge outputs (for driving thermo-electric coolers). Basically a big pain in the ***, but also a great learning exerience. Each pice of the system was pretty straight forward, but making them all play together on one board was tough. It did meet class B conducted emissions though.:rolleyes:
 
Sch3mat1c said:
All I've heard is that they are noisy as all hell but nobody's yet given me a reason why this noise can't be filtered, so go for it. I've tried a little work with flyback supplies and have had up to 10W output, I need a larger "spring"...
Tim

as you filter a square wave, eventually you get a pretty good sine wave but then you reduce efficiency -- this is how I use a PIC as a low distortion oscillator --

take a look at Application Note 70 on Linear Tech's website -- they have the noise down to where it can be used in hydrophones, medical applications, etc. You can order the parts from Linear directly, they have a low noise gate driver using the same slew-control technology.
 
Magura said:
A. How about getting +30-0- -30V, is that possible in a reasonable way?

B. How about an option to regulate the output voltage, lets say from 12V to 30V ?

Magura

Ans. A -- yes -- with two transformers, or one very good center-tapped transformers, or one transformer with two secondary windings.

Ans. B - by regulating the duty cycle you can regulate the output (but other things come into play.) use an optocoupler to sample the output, compare to an adjustable voltage reference.

you can also throw a linear regulator on the output -- an LM317 with a pass transistor.
 
Disabled Account
Joined 2003
Sch3mat1c said:
All I've heard is that they are noisy as all hell but nobody's yet given me a reason why this noise can't be filtered,
Tim


I used one (12v 3a) to power my active heatsink. the smps sits next to my jlh without any shielding or additional filtering. Zero noise that I can tell - the smps runs north of 150khz if I recall correctly.

it has 5 parts including the smps regulator.
 
millwood said:


I used one (12v 3a) to power my active heatsink. the smps sits next to my jlh without any shielding or additional filtering. Zero noise that I can tell - the smps runs north of 150khz if I recall correctly.

it has 5 parts including the smps regulator.

The ancient Sorensen switchers which I bought off EBay have less % noise than a 7805 -- a few millivolts .

whether they are good for music remains to be decided --- I know that some ham radio ops are not pleased with the results on CW and SSB using switchers due to the big pulse demands and the comparatively slow response time (which can lead to RFI.) I guess bigger reservoirs are in order here.
 
Noise on the output of a switcher is always difficult to 'beat down' to a decent level for audio or instrumentation purposes. It usually takes a small lossy choke and some extra capacitors. After that, a linear regulator can do an amazing job of reducing noise. Most people (I was one) are skeptical of the efficacy of the linear as a filter with its limited bandwidth, but it does work with up to 100kHz noise (maybe higher).
It is also important to prevent radiated noise from coupling back onto the output.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.