XLR DC blocker

Thread Starter

frenchie29

Joined Nov 28, 2008
134
I want to build a xlr dc blocker (female to male xlr) to use for intercom and other audio systems that run voltage.

When testing on my mixing board at home I still get 47v when I turn the phantom power on. I can see it getting down to 1.5v after a few minutes.

What's missing on that schematic to really block the dc?

 

t_n_k

Joined Mar 6, 2009
5,455
Are the capacitors a non-polarized type - presumably they are electrolytics? If they are polarized types (or tantalums) then you need to have the correct orientation of the terminals for the given DC voltage(s) you are blocking.

Also on the DC isolated side it would help if you connected the ("DC free") wires through a 100k (say) resistor to a DC common / ground return point. This would shunt any leakage voltage / current back to the DC source.

Perhaps indicate exactly what is connected on each wire.
 

Thread Starter

frenchie29

Joined Nov 28, 2008
134
yes polarized the 48v source is from the XLR-f connector (from the mixing board) and there's voltage on both pin 2&3 (current goes from left to right on the schematic) and pin 1 is the ground.
 

t_n_k

Joined Mar 6, 2009
5,455
So the left side terminal of the capacitors would be the +ve lead in each case. Hopefully the capacitor voltage rating is well above the 48V level in each case.

As suggested you could consider adding a 100k bleed resistor from both XLR-m pins 2 & 3 to pin 1. This shouldn't load any of your signals but would conduct any DC leakage current away.
 

Thread Starter

frenchie29

Joined Nov 28, 2008
134
Caps are 10uf 50v

So a 100k resistor between pin 2&3 and then how do I hook up pin 1 to the same resistor? But wouldn't that make the caps to share the common now? I taught that those caps had to be in series and seperated and share nothing.
 

t_n_k

Joined Mar 6, 2009
5,455
2 resistors, one from pin2 to pin 1 and another resistor from pin3 to pin1???

That's right - although I'm not sure your most recent drawing shows that.

One resistor soldered between pin 2 & pin 1. The other soldered between pin 3 and pin 1.

Your goal is to isolate the DC. The capacitors block the DC from left to right. Because the caps aren't perfect some current can leak across and a residual voltage can appear. It all depends on what is connected to the isolated side. If that plug is just sitting there unconnected on the right hand and you measure the DC on the plug terminals you will possibly see a voltage on your meter owing to the aforementioned leakage current. That seems what may have happened by your earlier account.

When you actually plug the right hand plug into the rest of the circuit the residual voltage may possibly disappear - depending on what else is connected on that side. The leakage current may find its way to ground through the rest of the now connected circuitry. If that turns out to be the case then you would not require the two added resistors.

I'm a bit wary about the 50V rating - even though it's above 48V. I always prefer a bigger margin. Particularly if what's connected on the isolated side could be damaged by the 48V suddenly appearing where it wasn't supposed to be.
 

Thread Starter

frenchie29

Joined Nov 28, 2008
134
I think you're right it was 2:30am when I draw the last one. As for the caps I was thinking the same thing a bit to close from the max I have 10uf 63v.
 
t n k said:
Because the caps aren't perfect some current can leak across and a residual voltage can appear.
Too right. I see an awful lot of commercial designs that omit these crucial tiedown resistors, and that's very, very bad. If nothing else it makes for very noisy and microphonic cable. Imagine your coaxial cable has 50 V of DC on it because of leaky caps and no terminating resistors. That coaxial cable acts as a charged capacitor, and if it's moved, crushed or vibrated then the distance between the outer braid and the inner signal core (effectively the two "plates" of a cap) will be altered. The charge will remain constant, so the voltage will have to change to accomodate the effective change in the coax capacitance - this changing voltage will be heard as nasty noise.

As for electrolytic ratings, a good rule of thumb is to run them at no more than 2/3 the maximum rated voltage. If you run an electrolytic near the maximum voltage, ripple current and temperature ratings then the lifetime is abysmal.
 

thatoneguy

Joined Feb 19, 2009
6,359
The above will have an impact of the line impedance. According to This site the impedance isn't as large of concern as it is in other applications, so the design by t_n_k should eliminate DC biasing. The only other option is an isolation/impedance matching transformer.

You should use 63 or 100V rated non-polarized electrolytics (+1/3 to double the anticipated working voltage)
 
frenchie29 said:
I still see a "spike" at around 7v when I turn the phantom power on.
Yup. You've now got an RC differentiator, and the spike is the differentiated step change when the phantom power is applied. It's quite normal, and can either be minimised by lowering the RC time constant, at the expense of a higher low-frequency rolloff, or better still, having a gently ramping phantom power upon switch-on.
 

Thread Starter

frenchie29

Joined Nov 28, 2008
134
Actually it's gone now.... I didn't change anything. before the pos of the cap and pin 1 I get 48v then after the resistor I get no voltage at all :D now my next test is to see if audio still can run trough
 
Top