Silencing an el-cheapo USB phantom power supply.

[k brown]

Got a couple of these to see if they'd be usable powering non-critical auxiliary mics - they weren't. Though surprisingly delivering a full 48 volts, and having decent build quality, the noise was outrageously high. So high, I thought the first one I tried was defective!

After the replacement proved to be just as noisy, I tried adding a simple RC filter (4.7k/47uF) just before the 6.8k phantom feed resistors, and the thing is now dead quiet. I just took a wild grab into my spare parts bin; I'm sure the values are far from optimized. The voltage was dropped to 30 volts, which for what I bought these for (electrets requiring no more than 12 volts) is no problem. I haven't experimented with what the minimum value resistor is needed to lower the noise to acceptable level.

After this simple mod these are a great deal. I'm going to see how long they last running on 4 AA rechargeables.

https://www.amazon.com/dp/B07W53324Y?psc=1&ref=ppx_yo2ov_dt_b_product_details

 

[Gus]

4.7K will cause a big voltage drop if the microphone current at the higher side of the phantom spec.
Have you tried a choke for the R it should have lower DC resistnace

 

[Tubetec]

This one has a rechargeable battery built in , its similar to the Alctron but 1/3 the price .
https://www.ebay.com/itm/362804797626

 

[Bo Deadly]

I'm not surprised. Stepping up from 5V to 48V is just a bridge too far.

If it has a battery then a possible better way to do it would have been to step up to 15V, use that to charge a 12V battery and then use that to step up again to 48V. If the battery has decently low impedance, it will act like a filter for the second stage. And stepping up from 5V to 15V and from 12V to 48V is a lot easier for a DC converter.

But clearly this is a low-budget device so of course they're not going to put two DC converters and a high quality battery in there. Certainly not for $12.

 

[k brown]

This one has a rechargeable battery built in , its similar to the Alctron but 1/3 the price .
https://www.ebay.com/itm/362804797626

That looks like a great unit.

Says internal power source is only 3.7v! Hmmm . . .

 

[k brown]

I'm not surprised. Stepping up from 5V to 48V is just a bridge too far.

If it has a battery then a possible better way to do it would have been to step up to 15V, use that to charge a 12V battery and then use that to step up again to 48V. If the battery has decently low impedance, it will act like a filter for the second stage. And stepping up from 5V to 15V and from 12V to 48V is a lot easier for a DC converter.

But clearly this is a low-budget device so of course they're not going to put two DC converters and a high quality battery in there. Certainly not for $12.

No, no battery inside.

 

[k brown]

4.7K will cause a big voltage drop if the microphone current at the higher side of the phantom spec.
Have you tried a choke for the R it should have lower DC resistnace

No, no chokes on hand.

The majority of the mics I'd use this with are fine with as liittle as 9v, and draw very little current.

But, yeah I should try to see how quiet it is with something like 1k/100 uF.

 

[k brown]

With 1k/68uF and mics i'd usually use with this connected, the voltage at the mic jack is 39 volts. Still quiet.

 

[Khron]

I'm not surprised. Stepping up from 5V to 48V is just a bridge too far.

That sort of presumes that all bus-powered USB audio interfaces are doomed to have noisy phantom power..?

Sure, this is a bottom-of-the-barrel unit, but still. 10-15mA @ 48v would only require ~150mA @ 5v (assuming a modicum of efficiency), sub-1W. Just saying

But yes, i'd also go with some great big honking inductor & cap filter on the phantom rail

That looks like a great unit.

Says internal power source is only 3.7v! Hmmm . . .

3.7v is the nominal voltage for a lithium-polymer cell...

 

[gyraf]

..when getting noise from any type of p48 source, I immediately suspect the two 6K8 resistors' degree of matching...

/Jakob E.

 

[k brown]

..when getting noise from any type of p48 source, I immediately suspect the two 6K8 resistors' degree of matching...

/Jakob E.

The noise was pure HF hiss, which after adding the simple RC filter is all gone.

 

[Tubetec]

There is a few phantom power boxes that have a 50v DC input from a wall wart Smps .
I have been thinking of modifying a phantom power box with a transformer , stepping up to bridging impedence ,48v supplied via a centre tapped primary instead of the usual 2x6.8k ,so less voltage drop .
As an example ,how would my Oktava MK012 work transformer coupled ? say 600:10k loaded by the usual 25k balanced line input .

 

[Bo Deadly]

That sort of presumes that all bus-powered USB audio interfaces are doomed to have noisy phantom power..?

True. But I bet they had to do extra work to get a clean output. A simple single step-up would be noisy. Think of it as like a mechanical lever. As you move the pivot point closer to one end, the other end becomes more difficult to keep stable. I suppose they could just step way up to like 70V and then there are filter / regulator options that will capture the noise.

 

[ccaudle]

A simple single step-up would be noisy

With just a single inductor boost converter maybe, but a push-pull transformer shouldn't be bad. I have no idea of the cost or minimum order quantities for a custom wound ferrite transformer, it isn't something I've ever had to price out. My guess is that the need for a custom transformer is the biggest reason you don't see it more often.

 

[cyrano]

That sort of presumes that all bus-powered USB audio interfaces are doomed to have noisy phantom power..?

I used to have a Sound Devices USB Pre. The first one. It was USB bus powered. It's the best one I ever encountered, yet it was still a bit noisy on bad computers. SD did a very good job at filtering phantom power, but USB can be a real dog.

I loved that thing for recording. Not for playback. I abandoned it because SD never provided updates for the driver, which was a real pity. I'll never buy SD interfaces again.

My AKG C451s were also noisy with this interface, but stabilised after a minute or so and were fine then. A few other phantom powered mics never worked with it. I'll never buy (expensive) bus powered interfaces again either.

 

[k brown]

In the case of this little unit, the issue was not noisy USB bus power - I was powering it from a battery.

 

[cyrano]

I understand, Mr. Brown.

But if SD can't do a 100% reliable step-up from 5V to 48V, I wouldn't expect some Chinese manufacturer to do better starting from 3.7V.

Of course, in the case of the SD interface, they have to power the rest of the electronics too. So, the Chinese have it a little easier.

 

[Tubetec]

A battery makes an almost noiseless supply , that doesnt mean what comes out the back end of a DC-DC switcher running off the battery is quiet .
There are a variety of external phantom power supplies , they run off wall warts with 5,12, 18 and around 50 volts DC , Some including the Mackie /Alctron/Knox have LI-PO backup , Often its the same boxes under different brand names by the looks ,

The main thing about these is its a nice strong alloy box with two XLR connectors and plenty of space for circuitry and even a transformer if you want it .
What ever garbage is under the hood can either be modified for low noise or scrapped , likewise they make a handy housing for any small audio project for minimum cost .

Theres loads of guys over on the Cosmos Discord channel ,they blunder onto the scene with their big ass gaming PC's , 650-1000W psu's ,fans , blue neon lights
they plug in Cosmos ADC, get a hopelessly bad S/N ratio, and scratch their heads in disbelief ,and then jump to the conclusion they got a faulty unit and get on the case of the designer and his helpers about it.

 

[Tubetec]

I had a dig around looking at DC-DC convertors and found a few interesting new pcbs
2-20v input, 1A out upto 28v max, 1.2mhz opperating frequency .
https://www.ebay.ie/itm/174123815317and this one outputs upto 40 volts 4A,400khz frequency
https://www.ebay.ie/itm/19177759759632v 400khz , 50mV noise
https://www.ebay.ie/itm/193522217299And finally one that goes upto 52V
https://www.ebay.ie/itm/255570151261The 1..2 Mhz board looks interesting but only goes upto 28volts , could two boards be stacked to get the required voltage ? would the oscillator frequency need to be synched on both boards so the garbage cancels better in the inductors? maybe Thor knows more .
Would be interesting to try them out on the usual op amp dual rail +/- setup ,

 

[cyrano]

This one goes up to 48V, from 5V:

https://envistiamall.com/products/dc-dc-6v-48v-step-up-boost-converter-module-xl6009
And it's only 3.49$!

50 mV ripple. Is it possible to filter this out?