| Author |
Message |
roborobert
Joined: Jan 15, 2013
Posts: 24
Location: austria
|
 Posted: Sun Oct 13, 2013 12:16 pm Post subject:
how to get rid of power rail noise |
  |
|
hey there,
around the finish line of my first big DIY-Synth Project, i'm running into a power rail problem.
for the last week i've been hooking up several modules (TH 555-VCO, TH XR-VCO, TH CRTL-LFO, MFOS DUAL AR, MFOS Noise Cornucopia, JH WASP-VCF, TH VCF-1 so far - more to come).
Sadly, the waveforms of Oscillators, AR's and LFO are coupling back into the power rails. Not very much, but enough to be audible on the VCOs. which is bad.
i used ferrite beads on all modules, adding two 1000uF Capacitors at the power bus doesn't change anything. Also adding them directly at the VCO didn't cause any improvement, so i don't think i can approach it this way.
the power supply itself is rather oversized.
are there more standard-procedures to follow? apart from beads and caps?
i'm totally clueless
hope someone experienced and solved this before.
best,
rob |
|
|
Back to top
|
|
 |
JingleJoe
Joined: Nov 10, 2011
Posts: 878
Location: Lancashire, England
Audio files: 14
|
| Posted: Sun Oct 13, 2013 1:46 pm Post subject:
|
|
|
Sometimes wires from one module passing by another, cause signal bleedthrough by induction, to remedy this, if you cannot keep the wires at a distance, you could try ground-sheilding sensitive modules. I recently acheived this in a ramshackle way with tinfoil glued between two peices of paper and wired into a near by GND or 0V connection. That cut down interference from induction by half which was good enough for my specification.
Are there any parts of your circuit with high gain/amplification in the VCO? That would probably explain the signal bleedthrough there: the circuit is amplifying the power supply noise!
To test for power supply noise I use an oscilloscope on AC setting with a sheilded probe. Sometimes the inductance of the probe wires and sheilding can add extra noise so if I were you I'd watch the VCO output aswell and see if connecting the scope affects any noise you were seeing there previously.
As a last resort, you could also try to make your noisey modules draw less current.
_________________
As a mad scientist I am ruled by the dictum of science: "I could be wrong about this but lets find out"
Green Dungeon Alchemist Laboratories |
|
|
Back to top
|
|
|
elektrouwe
Joined: May 27, 2012
Posts: 143
Location: Germany
|
| Posted: Sun Oct 13, 2013 2:02 pm Post subject:
Re: how to get rid of power rail noise |
|
|
| roborobert wrote: | ... enough to be audible on the VCOs...
|
does that mean you get an audible small FM from AR and LFO ?
in this case it's obvious that ferrite beads and other HF tricks (shielding) will not work, because it's a LF (=DC) problem.
Do you use thin power supply wiring (flat ribbon cable?) with module to module connection instead of a "star connection"
with seperate wires from the power supply to every single module ? |
|
|
Back to top
|
|
|
roborobert
Joined: Jan 15, 2013
Posts: 24
Location: austria
|
| Posted: Sun Oct 13, 2013 4:14 pm Post subject:
|
|
|
wow, thanks for your interest, that was fast!
@ jinglejoe
i don't think it's the wires passing since i'm using shielded cables nearly all over (at least for all connections that are or might be transporting signals of say 1Hz or more) - but i'll give it a try with the tinfoil shielding if i'm desperate 
your high gain question is beyond my knowledge.
the oscillators i built are thomas henry 555-vco and 2206(xr)vco.
the lfo is thomas henrys crtl-lfo, the envelopes are musicfromouterspace dual ARs.
i do know that i have quite some party going on on my power rails, envelopes, vcos and lfos are visible there when i check the power bus on my scope in ac mode. the signals i see on the scope are clearly the signals i hear changing pitch of the oscillator. if i can read my scope right the intensity is about 5-10mV for the negative rail and a bit more for the positive. is this much?
@elektrouwe
yes, that's exactly what i get.
i do use flat ribbon cable from the power bus to the modules.
the powerbus itself is rather fat (two stripboard pieces for connectors and 1mm (coppercore) wire for interconnection between the two pieces.
i do connect each module seperately to the bus.
also i took care that grounds on the shielded cables are connected only on the modules side, so there are no ground-loops from one module to another when patching.
i'll double check on this one since the cable from my power supply to my synths power bus does not meet these specifications.
thanks again to both of you for your fast answers!
best, rob |
|
|
Back to top
|
|
|
roborobert
Joined: Jan 15, 2013
Posts: 24
Location: austria
|
| Posted: Sun Oct 13, 2013 4:26 pm Post subject:
|
|
|
quick update:
better cable from my power supply to the bus improves the situation a lot. thanks for pointing me the right direction!
still, i have about 3mV mixed signals on the rails - still audible, but not disturbing anymore. great  |
|
|
Back to top
|
|
|
JingleJoe
Joined: Nov 10, 2011
Posts: 878
Location: Lancashire, England
Audio files: 14
|
| Posted: Mon Oct 14, 2013 6:00 am Post subject:
|
|
|
| roborobert wrote: |
also i took care that grounds on the shielded cables are connected only on the modules side, so there are no ground-loops from one module to another when patching. |
hold on hold on, where did you get that information, can you elaborate on how that create a ground loop? I ask because from my understanding, your proposed wiring will create a ground loop. I assume you are patching with 1/4 inch telephone jacks or similar?
in my experience they allways need the ground connected at both ends. I'll go into more detail if you can explain your reasoning first.
_________________
As a mad scientist I am ruled by the dictum of science: "I could be wrong about this but lets find out"
Green Dungeon Alchemist Laboratories |
|
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004
Posts: 24083
Location: The Netherlands, Enschede
Audio files: 278
G2 patch files: 320
|
| Posted: Mon Oct 14, 2013 7:08 am Post subject:
|
|
|
| JingleJoe wrote: | | in my experience they allways need the ground connected at both ends. I'll go into more detail if you can explain your reasoning first. |
When connecting separate pieces of equipment that do not yet have a shared ground connection, yes.
But when they are in a rack, and (thus) already have a shared ground connection, no.
_________________
Jan
also .. could someone please turn down the thermostat a bit.
 |
|
|
Back to top
|
|
|
roborobert
Joined: Jan 15, 2013
Posts: 24
Location: austria
|
| Posted: Mon Oct 14, 2013 8:47 am Post subject:
|
|
|
first of all, a big thanks to all here!
i went on building last night and the use of a larger and more sturdy power bus, as well as stronger power connections to the modules lowered my mixed signals on the bus down to around 2mV.
is this still much? or is it what you would expect?
concerning jinglejoes question: i couldn't explain it shorter than blue hell did.
the modules share their ground connection on the power bus.
each module has only one ground connection to that bus.
the shields of the cables (that are going to the jacks/pots) are connected to ground on the modules side.
the jack that connects my audio output to my mixer does have a ground connection (to the bus), of course.
best, R |
|
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004
Posts: 24083
Location: The Netherlands, Enschede
Audio files: 278
G2 patch files: 320
|
| Posted: Mon Oct 14, 2013 8:53 am Post subject:
|
|
|
2 mV is not bad 
Good enough can be determined from the synth behavior - I'd expect that to be all OK now.
_________________
Jan
also .. could someone please turn down the thermostat a bit.
|
|
|
Back to top
|
|
|
JingleJoe
Joined: Nov 10, 2011
Posts: 878
Location: Lancashire, England
Audio files: 14
|
| Posted: Tue Oct 15, 2013 12:39 pm Post subject:
|
|
|
| Blue Hell wrote: | | JingleJoe wrote: | | in my experience they allways need the ground connected at both ends. I'll go into more detail if you can explain your reasoning first. |
When connecting separate pieces of equipment that do not yet have a shared ground connection, yes.
But when they are in a rack, and (thus) already have a shared ground connection, no. |
yeah but why? surely it would make no difference?
_________________
As a mad scientist I am ruled by the dictum of science: "I could be wrong about this but lets find out"
Green Dungeon Alchemist Laboratories |
|
|
Back to top
|
|
|
Cynosure
Site Admin
Joined: Dec 11, 2010
Posts: 966
Location: Toronto, Ontario - Canada
Audio files: 82
|
| Posted: Tue Oct 15, 2013 1:30 pm Post subject:
|
|
|
Do you use bypass caps on your opamps and cmos chips? They are often either omitted from schematics or are shown all together on the bottom of the last page.
_________________
JacobWatters.com |
|
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004
Posts: 24083
Location: The Netherlands, Enschede
Audio files: 278
G2 patch files: 320
|
| Posted: Tue Oct 15, 2013 2:09 pm Post subject:
|
|
|
| JingleJoe wrote: | | yeah but why? surely it would make no difference? |
As in one case it would make a loop and in the other it would not, the loop consisting of the wire shield and the wires from the connectors to the common ground.
That loop would convert changing magnetic fields into an electric current, which would then change the voltage levels of what is supposed to be ground and this could become audible, especially so with fast changing magnetic fields.
_________________
Jan
also .. could someone please turn down the thermostat a bit.
|
|
|
Back to top
|
|
|
roborobert
Joined: Jan 15, 2013
Posts: 24
Location: austria
|
| Posted: Wed Oct 16, 2013 1:42 am Post subject:
|
|
|
@ buehell
cool, so i consider myself happy with the 2mV on the rails!
but still, because i'm curious and because it's still slightly, slightly audible - what would be the way to get rid of that too?
i guess it could be done by using +/- voltage regulators for each module - eg having a power bus voltage of +/- 18V and regulating it on each module down to +/- 15...
although this seems a bit overkill.
@ Cynosure
yes, i'm using bypass caps - but afaik they are for stabilizing the ICs (especially that logic doesn't get false triggers) and not so much for having a clean power rail. is that true? |
|
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004
Posts: 24083
Location: The Netherlands, Enschede
Audio files: 278
G2 patch files: 320
|
| Posted: Wed Oct 16, 2013 2:33 am Post subject:
|
|
|
| roborobert wrote: | i guess it could be done by using +/- voltage regulators for each module - eg having a power bus voltage of +/- 18V and regulating it on each module down to +/- 15...
although this seems a bit overkill. |
That would be one method yes, and not a bad one - regulators have good noise rejection usually. This would need a bit of attention though as regulators do not like to be reversed biased - that is output voltage being higher than input voltage - which could happen when powering off the synth. So add a diode there too. Also it needs the two capacitors mentioned in the data sheets (usually).
An other way would be to add some two way power filtering for each board, two small resistors (or inductors) and a capacitor in a T configuration. The capacitor probably would not have to be that larger, as its mostly the clicks on the power lines that would need to be smoothend a bit. This would need some care too, as you don't want too much resistance (to avoid large voltage drops), so this may need to be tuned a bit for each module.
_________________
Jan
also .. could someone please turn down the thermostat a bit.
|
|
|
Back to top
|
|
|
JingleJoe
Joined: Nov 10, 2011
Posts: 878
Location: Lancashire, England
Audio files: 14
|
| Posted: Thu Oct 17, 2013 12:13 pm Post subject:
|
|
|
| Blue Hell wrote: | | JingleJoe wrote: | | yeah but why? surely it would make no difference? |
As in one case it would make a loop and in the other it would not, the loop consisting of the wire shield and the wires from the connectors to the common ground.
That loop would convert changing magnetic fields into an electric current, which would then change the voltage levels of what is supposed to be ground and this could become audible, especially so with fast changing magnetic fields. |
OH you could have just said via induction, but that doesn't sound like a ground loop to me, it sounds like bog standard interference.
my understanding of a ground loop is that it is due to a disconnected ground, where the loop of the circuit then goes through the mains, resulting in a hum from the AC power supply- the common point is now an AC high voltage mains line and so introduces the familiar hum. I don't see how this could be created by induction?
_________________
As a mad scientist I am ruled by the dictum of science: "I could be wrong about this but lets find out"
Green Dungeon Alchemist Laboratories |
|
|
Back to top
|
|
|
forshee
Joined: Dec 16, 2010
Posts: 46
Location: Hopkins, MN
|
| Posted: Fri May 30, 2014 1:03 pm Post subject:
|
|
|
| Just as a test you may want to plug in module one at a time and see if its back feeding from one of the module. If adding a bunch of bulk didn't help it might be back feeding from a noisy circuit. It's always best to treat noise at the source and some added/improved decoupling and a CLC pi filter on noisy lines might clear things up. |
|
|
Back to top
|
|
|
|
Forum index » DIY Hardware and Software » Developers' Corner
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
