More Bench power supply questions

Thread Starter

spinnaker

Joined Oct 29, 2009
7,830
I has planning on building a bench power supply based on this design.

I want to make some modifications with it by having 2 outputs (plus ground) .

One would be variable. The other a fixed voltage that is selectable (3.3, 9, 12).

Am I better off using two LM317s and using trim pots to get the fixed voltages like he is doing here or would be better off using the individual regulators for each of my desired voltages?


Also I think I understand why he has individual C1 and C2. If I remember correctly, this helps with heat dissipation as opposed to one large cap?
But what is the purpose of C3?
 

studiot

Joined Nov 9, 2007
4,998
Comments:_

The design seems under specified for a heavily driven bench psu.
The pcb tracks are not heavy enough.
The regulator heatsink seems very lightweight.
I would recommend some 10n capacitors in parallel with the input and output electrolytics to the reg.


That said it has the basis of a useful supply.

I would recommend using an LM337 which is the negative version regulator if you want twin supplies. That way you can have a common floating zero and earth it if required. You cannot do this if you stack identical supplies on top of the other.
However you need to be aware that the pinouts for the 317 ( Vout is the centre pin is not the same as for the 337 (Vin is the centre pin) so you cannot unfortunately simply reverse all the capacitors and plug in the other polarity on the same pcb.
This need not be so bad if you use a 337/317 combination as you can then reduce the output below the 1.2 volt minimum to a true zero by referencing each from the other's rail.
 

Thread Starter

spinnaker

Joined Oct 29, 2009
7,830
Comments:_

The design seems under specified for a heavily driven bench psu.
The pcb tracks are not heavy enough.
The regulator heatsink seems very lightweight.
I would recommend some 10n capacitors in parallel with the input and output electrolytics to the reg.


That said it has the basis of a useful supply.

I would recommend using an LM337 which is the negative version regulator if you want twin supplies. That way you can have a common floating zero and earth it if required. You cannot do this if you stack identical supplies on top of the other.
However you need to be aware that the pinouts for the 317 ( Vout is the centre pin is not the same as for the 337 (Vin is the centre pin) so you cannot unfortunately simply reverse all the capacitors and plug in the other polarity on the same pcb.
This need not be so bad if you use a 337/317 combination as you can then reduce the output below the 1.2 volt minimum to a true zero by referencing each from the other's rail.
Thanks. I really did not want a heavy duty supply right now anyway. I'd probably build it on a perf board too (for now).

So my question still is, what is the purpose of the smaller capacitors in parallel with the larger ones?

Should I use individual fixed regulators for my fixed voltages?

And one I think might be no. If I do not have a large enough center tap transformers, can I use 2 transformers instead?
 

studiot

Joined Nov 9, 2007
4,998
Yes indeed you can use two transformers.

C3 is the same as my recommended 10n only at 100n it is ten times larger - the value is less important than the type. It is meant to bypass the electrolytics at higher frequency where they have high impedance, to prevent unwanted oscillation or other instability of the power supply. Note I recommended one at the output in parallel to C5 as well.

Additional fixed output regulators are useful, but the choice is yours.

Inadequate tracks will affect the regulation. By heavily driven I was referring to 1.5 amps which is within the capability of the 317. It is simple enough to provide adequate tracking or cabling, you only need do it for the main current and return paths.
 

Thread Starter

spinnaker

Joined Oct 29, 2009
7,830
Yes indeed you can use two transformers.

C3 is the same as my recommended 10n only at 100n it is ten times larger - the value is less important than the type. It is meant to bypass the electrolytics at higher frequency where they have high impedance, to prevent unwanted oscillation or other instability of the power supply. Note I recommended one at the output in parallel to C5 as well.

Additional fixed output regulators are useful, but the choice is yours.

Inadequate tracks will affect the regulation. By heavily driven I was referring to 1.5 amps which is within the capability of the 317. It is simple enough to provide adequate tracking or cabling, you only need do it for the main current and return paths.

Thanks. Just to be clear. He already has a .1uf across the input. You are recommending a 10nf instead?


Are there set rules for sizing the proper cabling?
 

SgtWookie

Joined Jul 17, 2007
22,230
If you are talking about filter capacitors after a rectifier bridge, then you should use 3,000uF per ampere of output current - that is, if you are using a full-wave bridge rectifier on 60Hz.

If you are using a half-wave rectifier, double the capacitance.

If you are using 50Hz into a full-wave rectifier, use 3,600uF per ampere output.
If you are using 50Hz into a half-wave rectifier, double the above.

You should always have a small capacitor right at the input to the regulator; 0.33uF-1uF.
 

studiot

Joined Nov 9, 2007
4,998
If you look at the pictures I posted towards the end of this thread

http://forum.allaboutcircuits.com/showthread.php?t=31158

You will see a power supply I made a while back with not a pcb or perf board to be seen.

Although only for fixed outputs it has lots of similarities to your proposal.

It has two transformers
Two bridge rectifiers
Two TO220 regulators

Look at the construction it may provide inspiration

You can see the anti oscillation capacitors mounted directly on the main bus wires. Standard 2.5square mm domestic ringmain wiring was pressed into service to form these - rated at 30 amps in service. Everything was done on a budget.
 
Last edited:

Wendy

Joined Mar 24, 2008
23,415
If you are talking about filter capacitors after a rectifier bridge, then you should use 3,000uF per ampere of output current - that is, if you are using a full-wave bridge rectifier on 60Hz.

If you are using a half-wave rectifier, double the capacitance.

If you are using 50Hz into a full-wave rectifier, use 3,600uF per ampere output.
If you are using 50Hz into a half-wave rectifier, double the above.

You should always have a small capacitor right at the input to the regulator; 0.33uF-1uF.
Where did you arrive at those numbers? Just asking.
 

Thread Starter

spinnaker

Joined Oct 29, 2009
7,830
If you are talking about filter capacitors after a rectifier bridge, then you should use 3,000uF per ampere of output current - that is, if you are using a full-wave bridge rectifier on 60Hz.

If you are using a half-wave rectifier, double the capacitance.

If you are using 50Hz into a full-wave rectifier, use 3,600uF per ampere output.
If you are using 50Hz into a half-wave rectifier, double the above.

You should always have a small capacitor right at the input to the regulator; 0.33uF-1uF.
Thanks Sarge. I was really talking about wire / track size but this is very useful information too.

I am going to have to make up a database with all of the useful information I learn on this forum!

I did some searching and found this useful calculator for PCB tracks.

I found a few calculators for wire size but they all seem to be for AC 120V and up.

Found a few tables but they don't seem to consider braided vs solid, material etc.
 
Top