Can DC be dirty?

[Che_Guitarra]

I understand how AC can be dirty, but DC is a constant value and i'm unsure how it would carry signal impurity. 

I ask as I have a DIY +/- 24VDC power supply which i'm almost certain is the root cause of noise in my EQ builds, but I don't have means for testing or analysing the PSU beyond my DMM.  The DMM reflects a steady output voltage from the PSU, but i'm not convinced it's telling the whole story.


Can anyone confirm or dispel my concerns - can DC be dirty?

 

[millzners]

Without a doubt. Dc is never really just Dc.  There is always an ac signal in there if you look close enough.  Put your Dmm on volts ac and look at your Dc rails.

 

[Che_Guitarra]

20 mVAC.  Interesting. What should I read into that?


So how would dirty DC manifest itself in a circuit - are there typical signs/symptoms?  Or are the possibilities infinite?

 

[CJ]

120 cycle hum would be what you are hearing with an analog supply,

switchers make all kinds of other stuff but sounds like you went analog,

add another pi filter, you want about 1 to 2 mv AC max on that DC supply,

maybe a choke would help too if the additional RC network does not cure the problem,

how much current are you drawing?

 

[Che_Guitarra]

The PSU is fused for 500mA @ the IEC input, and 500mA again on each DC rail.  I believe each EQ unit needs about 150mA.  I'd take a current reading but killed my last spare fuse when flicking through the bypass modes on the unit (trafo engaged, EQ engaged, full bypass), trying to make more detailed observations.  That makes it 6 fuses that have blown at the hand of the bypass switch.  Guessing by their fast fail rate i'm guessing they're not slow blows.  The perceived noise is at 240Hz i'd imagine, but it's not a tone as much as it just plain static noise.

BTW my EQ's are Bluzzi 1084s.  PSU is a Five Fish 2448mk2.

I'll add - Five Fish has sent me two PCBs by accident - one has already had a cap blow off it, this other one doesn't seem to offer clean DC.  I may end up throwing them both in the bin in favour of a Vintech PSU.

 

[millzners]

If your noise problem is primarily a harmonic of the mains AC, then it's usually a problem of either not filtering enough, or a ground loop, or cross-talk from power supply wires to audio wires.  If your DMM has a counter on it you could stick that on the DC rails to see what the dominant frequency is; the reason it isn't a tone is that mains AC always shows up as harmonics which sounds like buzzing.

Things to try would be:

Follow CJ's advice and filter some more.  Pi networks are great if you have inductors.  If all you have are caps, that could work too, just put some caps to ground on your DC rails and see if it helps.  Maybe a 10R series 100uF to ground filter could help although you'd have a voltage drop.

I'd also suspect a ground loop.  If you disconnect every external connector from your box, everything but mains power, then measure the volts AC on your DC rails what do you read?  Then once you connect your audio cables, etc. again what do you read? 

 

[My3gger]

Is psu on the picture giving you problems? It's probably the other one because this has blown cap. I wouldn't throw them in the bin, many people seem to use them without problems. I would check wiring on eq and ask Five Fish how to test his psu on it's own. It seems strange that he would send you psu with blown cap, ask him about that too.

 

[Che_Guitarra]

I build up the PCBs from a kit using the provided assembly guide.  The lid blew off that cap as I was adjusting the 48V phantom power trimmer.  Lucky I had the second PCB at my disposal.

I'd also suspect a ground loop.  If you disconnect every external connector from your box, everything but mains power, then measure the volts AC on your DC rails what do you read?  Then once you connect your audio cables, etc. again what do you read? 

20mA is the 'unplugged' reading.  I can't measure 'plugged in' until tomorrow when I pick up some more fuses. 

I wouldn't be surprised to find a ground loop occurring either.  When I first powered down the EQs (before I turned off the PSU)  I noticed the units became a radio receiver for a few seconds unit they fully discharged.

Should I be concerned that i'm blowing fuses so readily?  I'd assume it should be OK to switch between bypass modes whilst the units are on.  For the last few attempts every time I tried to engage the signal thru trafo, EQ bypassed mode it was instant death for the +24VDC rail fuse.

My last observation is that I can't draw a full 24V to both units when everything is powered up.  The most I can screw out is 23.3V, with EQ's off that jumps up to 29V at the +VDC rail.  Should I drop the resistor value paired with that rail, or should I resolve other symptoms first?

Hmmm.


BTW - I appreciate your thoughts guys

 

[My3gger]

You can change that cap, no problem. When i wrote check your psu on it's own i meant loaded with resistor so you can see if it is psu.

 

[horvitz]

I don't really know this PSU but if you look at the two small caps in your pic next to the blown cap, is one of them backwards?  It seems that the - side of every cap on the board is going the same way.  I'd be surprised to see a designer put just one facing the other direction.  Could explain your trouble.

  Brian

 

[My3gger]

Yeah, if you check polarity of diode bridges they seem to have + side in that direction. Take that cap out, clean the board and markings for cap polarity.

 

[Che_Guitarra]

I don't really know this PSU but if you look at the two small caps in your pic next to the blown cap, is one of them backwards?  It seems that the - side of every cap on the board is going the same way.  I'd be surprised to see a designer put just one facing the other direction.  Could explain your trouble.

  Brian

Wow, massive oversight... no wonder the other cap popped! 

Gosh, i'm paranoid now :-[  Time to go over the other circuit with a fine tooth comb.  I think this PCB is fit for work... unless it gives me grief for another two weeks