Request for a DIY bench +/- 15V variable voltage and variable current supply circuit

[Mosaic]

Well I guess the title says it.

I need to do a linear lab supply...the tired old ATX approach is a bit too noisy now as I getting into some audio type work.

For now I am looking at using the LM337/317, 1.5Amp combo to do it along with a fixed 5 & 3.3V regulator output.

I was wondering if there is a convenient circuit delivering both variable voltage and variable current limiting for a dual supply.

 

[MrAl]

Hi,

If you find a variable voltage and current supply made from an LM317, all you have to do is 'reverse' the circuit to get the same supply for negative voltage using the LM337.

If you do not need super super accurate current setting, i have a simple circuit for you which you can easily modify to work with the LM337 if you are interested, or i could show you the required changes.

 

[Mosaic]

That would be great!

 

[Mosaic]

Hi all ...
I am scrapping an old Daewoo DV 115 home theatre system that's gotten wonky.
Have a look @ the first and eighth pages here.
**broken link removed**

I am trying to figure out how to repurpose the transformer to make my linear supply.
I see a number of secondaries for 5V , +/-12 V and what seems to be a B+ and B- but no voltage specified. Then there is a -30V for the vac fluorescent display and a 4.3VAC heater coil (?).
I would have liked to get perhaps +/- 24VDC variable but I am not sure how. All the lower voltages seem possible from around 12V down.
Thanks for any advice.

 

[k7elp60]

I see all kinds of possibilities. If you can read the value of the filter capacitors in the +/- 12V and the various 5V supplies we can get an idea of what the
possible current output of those supplies may be. It is possible to connect several of the rectifier circuts in series to get a higher voltage. I would not use
three terminal regulators as your adjustable requlators, as they have internal current limiting that can work against you in that they sense the differential
in/out voltages with the current load and then shut down. I can give you a schematic that uses series pass transistors and current limit transistors.

 

[Joe G]

have you looked at the lm317 data sheet, there is a lab supply that fits what you need in it...

 

[Mosaic]

I'll have a look, I'd like to keep the 'noise' and ripple down as I plan to use it with preamplifiers etc.

 

[ronv]

It sounds like you are turning into an audiophile.
How about +/- 12 at 1 amp? Would that be enough?
If so and you have a radio shack I think they have some 24 volt center tapped transformers for like $10 that you could use.

 

[kinarfi]

Here's what I built, add a couple digital readout volt meters

 

[Mosaic]

That's interesting, but how come you're unbalancing on the +ve side with the FETS? Also, why not keep the ground common and measure currents on the +ve and -ve outputs?

5 milliohm sense resistors would give millivolt drops which are in the same range of the input offsets of the LM358 introducing a scaling error. I'd go with an acs713 linear current sensor.

I still expect noise, possibly into the RF spectrum on the outputs.

I am playing around with an SMPS prereg with a linear post reg (with a precision shunt) via a pi filter and ferrite beads to handle RF with small common mode capacitor grounding to earth and a common mode choke on the outputs.

 

[kinarfi]

The LM2576 is only good for 3 amps, adding and paralleling the fets let it handle around 20 amp, which was the original goal, the - ve just worked out well, 5 milliohm resistor at ground just worked out well for me also, noise is of no concern to me at this point. This is offered as a place to start or as a an idea to modify, as you get your's working to your specs, please update us.
just looked up acs713, cool idea, I like it, I have a LEM Current Transducer that I may use for current or I may invest in acs713 after I look at it some more

 

[Mosaic]

I did some testing on an LM2575 I had on hand for noise....I am seeing about 10mV p-p RF noise on the output which drops to about 4mV with a 5A ferrite bead. That's with 500Mhz, 10X probes. With a 20Mhz BW filter on the scope the noise is around 1mV.
Of course the probe itself picks up ambient noise. I'd have to fire up my spectrum analyzer to see if the 'noise' is from ambient Wifi or FM radio etc. I have near field probes that could sniff the SMPS as well....When I get my LM2576 I can let you know the best approach with it.

I'd be ok with a 3A dual supply with a zero volt offset.

 

[Dean Huster]

By the time you've bought all the components, put everything together and gotten things to sort-of work, you could have bought a commercial unit. For instance, on ebay, the Tektronix PS503A is a triple supply: ±0-20v, current limiting up to 1 a and +5v, fixed, 1 a. The variable supplies can be slaved together at any ratio and made to track one another. Put one in a TM504 power module along with a DMM (e.g., DM502), frequency counter (e.g., DC503A) and an audio oscillator (e.g., SG502), and you have quite a bit of power in a very small, inexpensive space. You can probably put it ALL together for around $250 US, and far less if you bid on and get "junkers" to repair. Tektronix equipment is VERY repairable and worth repairing.

 

[Mosaic]

I don't need a DMM, Frequency counter or audio oscillator.
A 3A dual smps tracking/linear supply will cost me under $50 DIY all new parts.
Since I am not in the USA...shipping and customs costs factor in for finished goods, and I would have learned little.