Designing dual linear power supply

[networkingdude]

I was hoping someone could give my schematic a once over to see if I've made any terrible mistakes.

I used the data sheet as a reference for some of it and other implementations that I thought were good.

The design goal was a power supply with low noise, +-15v and should have no issues powering a headphone amplifier or dac.

I welcome all criticism good or bad.

[eblc1388]

If you want 15V output from LM350, then you should choose a transformer secondary of 18V or more instead of 15V, otherwise the output voltage may not remains regulated at full current output.

The output capacitor value of 1500uF is especially large, against the 10uF usual value, which might not be needed or even bad to the overall performance of the regulator.

[Kibabalu]

D1 and D4 are bypassed by a shortcut

Transformer wiring secondary side is wrong. 4 and 5 should be ground and you need just one bridge rectifier. For the negative side you have to use a LM333.

[ledtester]

Here's a thread discussing a LM317/LM337 design:

https://www.eevblog.com/forum/beginners/transformer-voltage-for-dual-power-supply-(12v-12v)/msg1773620/#msg1773620

(see the link in the 3rd post for the schematic)

You can also get kits on ebay which implement this design -- just search for "LM317 LM337 power supply".

[fcb]

Ditch D3 and D6.  You've got other problems if you can get a reverse polarity on these sections.

Your o/p capacitor is unnecessarily large and maybe counterproductive (poor transient response).  If it's for audio, you might want to add some lower value capacitors to improve the response at higher frequency, capacitor type might be more important than value - we are straying intocontroversial territory here. Datasheet recommends solid tants, but it's an old ds.

[Wolfgang]

Hi,

looks OK to me. Now provide some adequate heatsinking. The input caps could be a bit larger, but thats it.

Have fun
  WG.

[networkingdude]

If you want 15V output from LM350, then you should choose a transformer secondary of 18V or more instead of 15V, otherwise the output voltage may not remains regulated at full current output.

The output capacitor value of 1500uF is especially large, against the 10uF usual value, which might not be needed or even bad to the overall performance of the regulator.

I was worried about the voltage so making it 18 is not a big change.

As for the output cap, the 1uf are all solid tants and the large capacitor idea was on a commercial product so I figured it was helpful.

[networkingdude]

D1 and D4 are bypassed by a shortcut

Transformer wiring secondary side is wrong. 4 and 5 should be ground and you need just one bridge rectifier. For the negative side you have to use a LM333.

D1 and D4 are like this in all examples of ldo's I could find. Also the idea was to create 2 independent positive regulator power supplies. I know you normally would just use a negative regulator but then you only need one bridge rectifier.

This configuration gives lower noise value.

[networkingdude]

Ditch D3 and D6.  You've got other problems if you can get a reverse polarity on these sections.

Your o/p capacitor is unnecessarily large and maybe counterproductive (poor transient response).  If it's for audio, you might want to add some lower value capacitors to improve the response at higher frequency, capacitor type might be more important than value - we are straying intocontroversial territory here. Datasheet recommends solid tants, but it's an old ds.

I will remove D3 and D6, I can see how they are redundant. The 1uf caps are solid tants, I will lower the output capacitor.

Should I increase or decrease input caps?

[Shannon]

Hi,

looks OK to me. Now provide some adequate heatsinking. The input caps could be a bit larger, but thats it.

Have fun
  WG.

strongly agree

[tunk]

For the input caps, I think a rule of thumb says ~2000uF/A,
which gives a ripple around 4V at full load.

[Labrat101]

Hi Your Ground loop is really not a good . and not recommended  .
Not to Say It won't work
You would be better making the transform into a Center tap arrangement and use just 2 diodes . from a centre tap with a common
ground direct to the centre tap will give a full wave rectification . Also it will increase affiance .
using LM317 LM337 . The large cap replace with 2 x 480uf  . The 1uf at the input of the LMs 0.1uf
 also this arrangement with a 15volt transform will work better on full load centre taped
As someone else mentioned there are dozens of designs on line that are proven .
 As your circuit stands It will suffer . Theory and Practice don't always work well in real life.

Have fun 

[networkingdude]

For the input caps, I think a rule of thumb says ~2000uF/A,
which gives a ripple around 4V at full load.

I will make this change.

I have also seen some power supplies with a 2watt or larger resistor with a capacitor across the rails. Should this be considered? Is there a rule of thumb for this?

[networkingdude]

Hi Your Ground loop is really not a good . and not recommended  .
Not to Say It won't work
You would be better making the transform into a Center tap arrangement and use just 2 diodes . from a centre tap with a common
ground direct to the centre tap will give a full wave rectification . Also it will increase affiance .
using LM317 LM337 . The large cap replace with 2 x 480uf  . The 1uf at the input of the LMs 0.1uf
 also this arrangement with a 15volt transform will work better on full load centre taped
As someone else mentioned there are dozens of designs on line that are proven .
 As your circuit stands It will suffer . Theory and Practice don't always work well in real life.

Have fun 

I'm making this with positive voltage regulators only as they appear to have a lower noise. This necessitates building 2 independent power supplies.

Could you elaborate on ground loops? I'm still learning the basics.

[Labrat101]

Hi Your Ground loop is really not a good . and not recommended  .
Not to Say It won't work
You would be better making the transform into a Center tap arrangement and use just 2 diodes . from a centre tap with a common
ground direct to the centre tap will give a full wave rectification . Also it will increase affiance .
using LM317 LM337 . The large cap replace with 2 x 480uf  . The 1uf at the input of the LMs 0.1uf
 also this arrangement with a 15volt transform will work better on full load centre taped
As someone else mentioned there are dozens of designs on line that are proven .
 As your circuit stands It will suffer . Theory and Practice don't always work well in real life.
Update .. As per diagram They are Not independent due to the ground arrangement 
Have fun 

I'm making this with positive voltage regulators only as they appear to have a lower noise. This necessitates building 2 independent power supplies.

Could you elaborate on ground loops? I'm still learning the basics.

You Have got the output of ic2 feeding to the negative of the first stage bridge . this will cause a imbalance  and will properly make
the second IC2 run very hot and fail . and if it does it will take out your transform as well an the bridge rectifiers.
 Do not use this type of ground loop its an accident waiting to happen

[networkingdude]

Hi Your Ground loop is really not a good . and not recommended  .
Not to Say It won't work
You would be better making the transform into a Center tap arrangement and use just 2 diodes . from a centre tap with a common
ground direct to the centre tap will give a full wave rectification . Also it will increase affiance .
using LM317 LM337 . The large cap replace with 2 x 480uf  . The 1uf at the input of the LMs 0.1uf
 also this arrangement with a 15volt transform will work better on full load centre taped
As someone else mentioned there are dozens of designs on line that are proven .
 As your circuit stands It will suffer . Theory and Practice don't always work well in real life.
Update .. As per diagram They are Not independent due to the ground arrangement 
Have fun 

I'm making this with positive voltage regulators only as they appear to have a lower noise. This necessitates building 2 independent power supplies.

Could you elaborate on ground loops? I'm still learning the basics.

You Have got the output of ic2 feeding to the negative of the first stage bridge . this will cause a imbalance  and will properly make
the second IC2 run very hot and fail . and if it does it will take out your transform as well an the bridge rectifiers.
 Do not use this type of ground loop its an accident waiting to happen

How is this different than connecting a 2 channel bench power supply together to get a positive and negative voltage? I'ts connected the same way?

[Labrat101]

Hi Your Ground loop is really not a good . and not recommended  .
Not to Say It won't work
You would be better making the transform into a Center tap arrangement and use just 2 diodes . from a centre tap with a common
ground direct to the centre tap will give a full wave rectification . Also it will increase affiance .
using LM317 LM337 . The large cap replace with 2 x 480uf  . The 1uf at the input of the LMs 0.1uf
 also this arrangement with a 15volt transform will work better on full load centre taped
As someone else mentioned there are dozens of designs on line that are proven .
 As your circuit stands It will suffer . Theory and Practice don't always work well in real life.
Update .. As per diagram They are Not independent due to the ground arrangement 
Have fun 

I'm making this with positive voltage regulators only as they appear to have a lower noise. This necessitates building 2 independent power supplies.

Could you elaborate on ground loops? I'm still learning the basics.

You Have got the output of ic2 feeding to the negative of the first stage bridge . this will cause a imbalance  and will properly make
the second IC2 run very hot and fail . and if it does it will take out your transform as well an the bridge rectifiers.
 Do not use this type of ground loop its an accident waiting to happen

How is this different than connecting a 2 channel bench power supply together to get a positive and negative voltage? I'ts connected the same way?

 
    A lot first your bench power supply does not have 2 sets of bridge recertifies .
Plus your Bench power supply outputs are configured differently to what you are doing.
If you read the recommend way of doing
this on the Manufactures site it will explain how to connect as duals .
https://www.ti.com/lit/ds/snvs772b/snvs772b.pdf?ts=1612018138744&ref_url=https%253A%252F%252Fwww.google.com%252F

There is all the info you require .  But of course remember its your Project so  you are free do do what ever you wish .
 If you ask for guidance fine . your are a Novice for sure . No hard feeling we all started there .
 But follow the Manufactures recommendation and that way you will learn Why.

[networkingdude]

Hi Your Ground loop is really not a good . and not recommended  .
Not to Say It won't work
You would be better making the transform into a Center tap arrangement and use just 2 diodes . from a centre tap with a common
ground direct to the centre tap will give a full wave rectification . Also it will increase affiance .
using LM317 LM337 . The large cap replace with 2 x 480uf  . The 1uf at the input of the LMs 0.1uf
 also this arrangement with a 15volt transform will work better on full load centre taped
As someone else mentioned there are dozens of designs on line that are proven .
 As your circuit stands It will suffer . Theory and Practice don't always work well in real life.
Update .. As per diagram They are Not independent due to the ground arrangement 
Have fun 

I'm making this with positive voltage regulators only as they appear to have a lower noise. This necessitates building 2 independent power supplies.

Could you elaborate on ground loops? I'm still learning the basics.

You Have got the output of ic2 feeding to the negative of the first stage bridge . this will cause a imbalance  and will properly make
the second IC2 run very hot and fail . and if it does it will take out your transform as well an the bridge rectifiers.
 Do not use this type of ground loop its an accident waiting to happen

How is this different than connecting a 2 channel bench power supply together to get a positive and negative voltage? I'ts connected the same way?

 
    A lot first your bench power supply does not have 2 sets of bridge recertifies .
Plus your Bench power supply outputs are configured differently to what you are doing.
If you read the recommend way of doing
this on the Manufactures site it will explain how to connect as duals .
https://www.ti.com/lit/ds/snvs772b/snvs772b.pdf?ts=1612018138744&ref_url=https%253A%252F%252Fwww.google.com%252F

There is all the info you require .  But of course remember its your Project so  you are free do do what ever you wish .
 If you ask for guidance fine . your are a Novice for sure . No hard feeling we all started there .
 But follow the Manufactures recommendation and that way you will learn Why.

This circuit is the same as 2 different fully independent bench power supplies set up in series. I'm using only positive voltage regulators. I think you are misreading the schematic. Pretend the 2 supplies are not tied to one another. They are independent and isolated from each other.

[Labrat101]

Build It . exactly as per your schematic and then see what happens .
 It will work but there will be an unbalanced as per your arrangement .
 Put a scope on the finished project and all will be come clear.
 I have been in electronics for over 60 years done that seen that . 

IC2 VO  is linked back to BR1 AC  ~  VO is DC out  .. Top minus & Bottom Plus
  are not true .    Your Ground
BR1 is coming off the first winding
BR2 is the second winding .
The chances that the transformer you are using the winding are probable going to be out of phase
with each other . ( Chinese stand  i guess And the 2 bridges are identical in all characteristics. yes)
The  Ground as per your drawing can not & will not be true Zero . To Both .
Therefore the top and bottom will not be isolated as per drawing

 Put up the scope results of the finish project .
Good Luck

[bdunham7]

    A lot first your bench power supply does not have 2 sets of bridge recertifies .
Plus your Bench power supply outputs are configured differently to what you are doing.

Multichannel linear bench supplies most certainly do have separate bridge rectifiers with isolated separate secondary windings for each channel.  What you have here is two completely independent isolated channels that can be connected in series for higher voltage, as bipolar (as the OP is doing) or even in parallel with additional controls.  I'm not seeing a 'ground loop' as there is one and only one connection to ground.

[networkingdude]

D1 and D4 are in place as per the datasheet.  Diode D1 and D4 protects against C4,C5,C9 and C10 discharging into regulator.
D2 and D5 protect adjustment pin from C3 and C8.
I will be removing D3 and D6 as they appear redundant.

[julian1]

It *looks* like a really interesting tradeoff. Add another rectifier, but gain the benefit of using identical positive regulators on both top/bottom, with more component choice.

A question, more than a comment. Is there a difference in the ripple phase, between this and the more conventional single rectifier approach? Here, the ripple would be 180 out of phase - which is desirable I think - for stuff like powering bipolar supply ops. eg. the common mode is cancelled.
 

[Wolfgang]

Your diodes are all OK. I also have D3 and D6 to protect the circuit against outside voltages when power is accidentially switched off.

The design with separate transformer windings and rectifiers as you have it has advantages:
- The outputs can be used as independent power supplies if needed
- NPN-Type (positive) regulators are better, cheaper and more robust than negative voltage parts. LM317 is better than LM337, and so on ...
  With a tapped transformer winding and just one rectifier bridge, you must use negative regulators. In your design, you can avoid that.

Otherwise, go ahead.

[Vovk_Z]

Could you elaborate on ground loops? I'm still learning the basics.

There is no need to change schematics to centertap. Everything is fine.

[Vovk_Z]

I will remove D3 and D6, I can see how they are redundant.

Don't get rid of them. It's more reliable with them.