Author Topic: toroidal share = 50hz noise.  (Read 1328 times)

0 Members and 1 Guest are viewing this topic.

Offline harpsTopic starter

  • Regular Contributor
  • *
  • Posts: 73
toroidal share = 50hz noise.
« on: December 11, 2017, 12:01:27 pm »
Hi

I need dual power for 2 x analogue preamp circuits.

My system all works fine when I use 2 x toroidal's, one for each preamp.

however  when sharing a single toroidal to these two pre amps a very low 50hz is
noticed on the output.

power being used is 230v 50hz.

Test 1.

Toroidal (1)> rectifier> caps>regulator>  caps> preamp  =  o.k  next to no noise.
Toroidal (2)> rectifier > caps> regulator>caps> preamp  = o.k, next to no noise.

So the system works fine on test 1.

Test 2 exactly the same circuit except now only one toroidal is shared

Toroidal (1) shared>
                             > rectifier> caps> regulator> caps> preamp  =  low noise at 50hz
                              >rectifier> caps>regulator> caps>  preamp  =  low noise at 50hz

Can i add a little circuit  to tame this?


Thanks for help.
 

Offline drussell

  • Super Contributor
  • ***
  • Posts: 1855
  • Country: ca
  • Hardcore Geek
Re: toroidal share = 50hz noise.
« Reply #1 on: December 11, 2017, 12:21:13 pm »
You have very likely created a ground loop somewhere.

Perhaps some photos of your apparatus would be helpful.
 

Offline Nitrousoxide

  • Regular Contributor
  • *
  • Posts: 156
  • Country: au
Re: toroidal share = 50hz noise.
« Reply #2 on: December 11, 2017, 01:20:01 pm »
Just a quick observation, Make sure you have the requisite input filtering which is usually a PI shaped configuration of C-L-C.
Also, why do you need two rectification stages? Can't you just add linear regulators after the first set of capacitors?

The buzz could come from a poor ground return path or be coupled in from somewhere. Photos/schematic would help.
 

Offline harpsTopic starter

  • Regular Contributor
  • *
  • Posts: 73
Re: toroidal share = 50hz noise.
« Reply #3 on: December 11, 2017, 02:13:49 pm »
How do you mean by "requisite input filtering which is usually a PI shaped configuration of C-L-C"

I could connect the regulators to one rectifier , but they are on separate printed circuit boards at the moment.  it shouldn’t matter for a test.
i'm not trying to make that section compact yet.


 

Offline Andy Watson

  • Super Contributor
  • ***
  • Posts: 2085
Re: toroidal share = 50hz noise.
« Reply #4 on: December 11, 2017, 02:56:49 pm »
it shouldn’t matter ...
Clearly it does matter.
As  drussell has pointed out, you have created a ground loop - probably including the input to your preamps - that is being used to distribute the relatively high charging currents of the reservoir capacitors.
 

Online Zarhi

  • Contributor
  • Posts: 24
  • Country: bg
Re: toroidal share = 50hz noise.
« Reply #5 on: December 11, 2017, 03:05:31 pm »

Toroidal (1) shared>
                             > rectifier> caps> regulator> caps> preamp  =  low noise at 50hz
                              >rectifier> caps>regulator> caps>  preamp  =  low noise at 50hz

Can i add a little circuit  to tame this?


Thanks for help.

Use only one rectifier for both boards and 50Hz noise will gone.
 

Offline Brumby

  • Supporter
  • ****
  • Posts: 12298
  • Country: au
Re: toroidal share = 50hz noise.
« Reply #6 on: December 12, 2017, 12:14:40 am »
Use only one rectifier for both boards and 50Hz noise will gone.

That would be my first suggestion.

I suspect there would be slightly different forward voltage drops across all the rectifiers which could cause problems if there is a cross connection between the 0V rails.  Such a cross connection could be created by a common chassis connection, for example, or it could occur when feeding signals to both channels where the shields/common on those connections are joined within the source device.  (If this latter is the case, do you still get the hum if you disconnect one of those inputs?)
 

Offline Nitrousoxide

  • Regular Contributor
  • *
  • Posts: 156
  • Country: au
Re: toroidal share = 50hz noise.
« Reply #7 on: December 12, 2017, 05:29:33 am »
How do you mean by "requisite input filtering which is usually a PI shaped configuration of C-L-C"

It shouldn’t matter for a test.

How did you come to that conclusion? Was it just a guess. Indeed it does matter, as per mentioned due to variations in silicon, the two ground potentials are different.

A PI topology filter is usually implemented to suppress EMI and in general is a good practice as you will see it in many power supplies. It is a capacitor in parallel, followed by an inductor in series, followed by another capacitor in parallel.

https://en.wikipedia.org/wiki/Capacitor-input_filter

Your best bet is to just copy (within reason). Look at supply topologies and how most input stages for power supplies are designed, that should give you a few hints.

What type of regulators are you using? Make sure you do not violate the input and output capacitance stated by the datasheet.
Do you have decoupling for the pre-amps?
« Last Edit: December 12, 2017, 05:34:52 am by Nitrousoxide »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf