low noise switching power supply design

[SSShocked]

i want to make a laboratory switching power supply
problem is
low noise-less than 300uV
high eff. not lower than 90%
output 0-100v 0-25a 2000va max(reach 25a at 80v)
high regulation 0.1% voltage and current

i don't know is it possible
maybe NCP1034 or L4981
https://www.onsemi.com/PowerSolutions/product.do?id=NCP1034
https://www.st.com/internet/analog/product/63215.jsp

any help will be appreciated

sorry for my english

 

[mtwieg]

You should find this app note interesting:
http://cds.linear.com/docs/Application Note/an70.pdf

Basically their is a direct trade off between efficiency and noise.

But your application requires a wide output range with high power. Even without the noise requirement, 90% would be very, very difficult. With the noise requirement I don't think it would be possible without huge amounts of filtering and shielding.

 

[SSShocked]

what about low frequency? 400hz
i found in a bulgarian book :
low frequency power amplifier
the losses are 10w maximum at 400hz and 1kw output(99.9% eff)



i think because of this frequency there will be maximum 200uV noise
50 and 60hz transformers almost have no noise
i don't care about the size

 

[RCinFLA]

80v @ 25 amps with 300 uV of noise! Your only chance is to use the switcher to pre-regulate followed by linear regulator. The switcher should maintain a few volts of overhead above the linear regulator.

Even with that, if your loads are varying, wire resistance will likely cause more then 300 uV of ripple voltage.

 

[SSShocked]

the dude who designed this amp. say'd:
at 1000w load and 400hz the noise will be less than half uV-
this is because of the switching frequency

so my idea is
two amps. bridged
the first is the error amp. (using op. amp.)
the second one is through potentiometer with zener diode...

so will it work

- - - Updated - - -


the schematic

 

[godfreyl]

View attachment 76636
the schematic

That circuit uses very old Germanium transistors. I don't think you could build that now because the transistors are obsolete and have not been available for many years.

The reason for the low switching frequency is probably because the transistors used were very slow.

 

[SSShocked]

i can replace them with Si
it will be better
also i have them all and in bulgaria you can buy them from every electronics shop

- - - Updated - - -

the reason for the low switching frequency is the high efficiency and low noise
(the dude say'd so)

- - - Updated - - -

and they are very cheap too

- - - Updated - - -

i will have to calculate the inductor for 100v
maximum voltage of this circuit is 50v
i have the formulas

 

[SSShocked]

any idea what will be the regulation?

 

[mtwieg]

what about low frequency? 400hz
i found in a bulgarian book :
low frequency power amplifier
the losses are 10w maximum at 400hz and 1kw output(99.9% eff)

Wait, are you talking about a DC power supply or a signal amplifier? They're completely different things.

And I don't think there's such a thing as a 99.9% efficiency power supply/amplifier, regardless of frequency.

 

[SSShocked]

dc power supply
В. I. Cooper, 23. 0 6 . 59
he designet this power amp.
my question is-can i convert it to dc power supply

- - - Updated - - -

On fig.I-73 is a principal scheme of the DC
power amplifier with 1000 W, working in key mode. For the clock
frequency of 400 Hz losses of about 10W. Load resistance is related to the general
collector output terminal connected in parallel transistors
whose bases are connected through a resistor (about 1Ω) to the emitter
the front transistor for equalization modes. End
core transistors are controlled also by key amplifier
composed of transistors T5, T6, T7 and. Blokinggeneratorat T3 works
a coil with ferrite core and creates a saw-tooth pulses
that control switching of Schmitt trigger composed
of T3 and T4.
For greater linearity instead blokinggenerator can
different type of generator with more transistors and more complex
scheme.
In Fig. I-74 a, b are shown the dependencies of the output load
current from the DC input and loss in function
of the clock.
The amplifier can be amplified and variable signal, but in this
If it is necessary to clock higher number
times the highest frequency band gain
tension. The amplifier can be used for control and management
engines, DC power supplies and other purposes.

- - - Updated - - -

translated with Google translator

 

[schmitt trigger]

SSShocked:
The schematic you provided is only a small section of a complete converter.
T1 is a blocking oscillator
T2 is a width modulator
T3,T4 are a one shot
T5,T6 and T7 are a triple darlington to increase the current drive
and T8, T9 are predrivers for.......something

Missing is the power stage (comprised of several transistors, magnetic components, etc) and error amplifier (comprised of many additional transistors and references).
Do you have any additional schematics?

On other topic, I'm interested in purchasing germanium transistors for historical value. Do the shops in Bulgaria that stock them have a web page where I could purchase them?