Dual adjustable power supply

Thread Starter

Doros

Joined Dec 17, 2013
144
Hello everybody,

I have a couple of power supplies (SMPS) coming from salvaged printers giving up to 32V-1A and I was thinking to make a dual adjustable power supply, using the LM317/LM337 voltage regulators.

Is any drawback using these SMPS power supplies to do so, compared to use a center tapped transformer, rectify the output etc?

I am intending to use the dual power supply for op amp circuits

Thanks for your help

Doros
 

DickCappels

Joined Aug 21, 2008
10,152
Using an SMPS to drive the analog regulators might result in a significant amount of high frequency noise, particularly if you are not very careful about grounding. In the case of using the simple 50Hz/60Hz transformer-bridge-capacitor source you are likely to have more ripple at 2X the line frequency than with the SMPS source -those regulators have limited ripple rejection capabilities.
 

crutschow

Joined Mar 14, 2008
34,285
Note that with a 32V input, the power dissipated by the regulator could be greater than 30W@1A and low output voltages, so it will need to be on a good heatsink (perhaps with a fan) unless you intend to draw only small currents from the output.
 

vishnumaiea

Joined Jul 9, 2017
1
About one and half year ago I built a simple variable linear power supply using a 36V, 2A SMPS as source and LM317 as voltage regulators. Why I used the SMPS becasue I didn't have any use for such high voltages and it was made in Hungary - so stable and well built unlike the Chinese ones. You can see my project here - https://www.instructables.com/id/LM317-Based-DIY-Variable-Benchtop-Power-Supply/ - It won second prize (around USD 500) at the Instructables Power Supply Contest 2017. I used bypass transistors to reduce the heat dissipation at the voltage regulators. But if you're intending to use the PSU for op-amp circuits you might need a split rail one with least amount of ripple for which you'll need to use center tap transformers.
 

Thread Starter

Doros

Joined Dec 17, 2013
144
Thank you all very much for your input.

to put things in perspective.

I intend to power op amps (dual power) for educational purposes and projects, so the current won't be high. Not more than 80mA in extreme cases. I will also use heat sinks if for any reason, I get ahead the 100mA limit.

Concerning noise, the SMPS are in metal cases, to be grounded. Isn't possible to observe/measure the noise with my oscilloscope, and if is not acceptable dump the project and use center tap transformer instead?

vishnumaiea, if I understood well you are more for a center tap transformer, even though the ripple will be higher than with the SMPS

Thanks for your help
Doros
 

Thread Starter

Doros

Joined Dec 17, 2013
144
Hello crutschow

I have two identical SMPS. I thought, from threads I read also from this forum, that I cannot use them as the photo below.
Because they are never "identical" and may fight each other. I didn't understand it, but took it for granted.

If I can do it and will be a better solution I can go for it

Thanks again Doros

upload_2018-2-4_18-44-1.png
 

crutschow

Joined Mar 14, 2008
34,285
I have two identical SMPS. I thought, from threads I read also from this forum, that I cannot use them as the photo below.
Because they are never "identical" and may fight each other. I didn't understand it, but took it for granted.
I see no problem with that.
They don't have to be identical since they are completely separate circuits.

Perhaps you are referring to the problem of putting them in parallel.
 

Thread Starter

Doros

Joined Dec 17, 2013
144
In parallel you mean, if someone wanted to get from two power sources 32VDC each, 64VDC total?
 
The only reason why you can't connect them as above is if one side is grounded.

Series (double the voltage) is usually no problem; https://www.electro-tech-online.com/parallel doubles the current.

Some Lab supplies have the ability to have Master/Slave relationship. In series mode, one control, controls both supplies (2V out, means 1V from each supply). Current mode is handled separately.


One supply is actually operated in CV mode and the other CC mode. The CC set point of the other supply is off a little bit. So if the master (CV) is set for 30 v (CV) 1A, the slave might be set for 32 V and 0.5 A (CC).
The second supply never sees 32V and delivers about 1/2 the current.

Some circuits do much better with "tracking" power supples. This generally means that ground is centered between the rails at all times. You don't get +15 and -14.7. You might get +-14.7 or +-15.
 

Thread Starter

Doros

Joined Dec 17, 2013
144
If no power supply is grounded can I use the configuration below? Given that I will not connect the VGND to physical ground

upload_2018-2-4_19-38-14.png
 
One would have to have it's positive grounded and the other, it's negative grounded. So, no.

You can always use an isolation transformer on one of them. If you had say A few 24:120 xformers around, you can connect them like 120:24 24:120 and either get 120 isolated or zero Volts out. If zero, swap the 24 VAC lead connections.

Isolation transformers have other properties, but that will work.
 

Thread Starter

Doros

Joined Dec 17, 2013
144
Thanks for the info, I have a direction to go after make some tests. (continuity-noise etc)

So for laboratory op amp projects there are options for dual power supplies.

But if you want to make a handheld instrument, then you have two options. Either, if is suitable for your application, use a single supply op amp, or if you need dual supply but in a hand held instrument, use something like the TLE 2426 rail splitter, if you are happy with the 20mA

thanks again for your help
 
Top