UniPhan. Or, easy to add 48V (phantom) to.... almost every possible pre!

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Igor

Well-known member
Joined
Jul 17, 2006
Messages
2,193
Location
Israel
Mm. So, universal solution to add phantom to almost every pre.
The story before the pictures not necessary to read, I marked it ----- -----

;)

----- -----

Cool. Mentioned I hate DC/DC converters. Most of them.

Some time ago, came my friend with horribly racked V672.
The prob was contacts in gain switch and cracky pot, fixed'em and almost forgot, however, it probably was the start for idea.
Now, try to follow me. (He!) :)
Poor-thing's +24V power supply was done on  2x12V, 0.5A chinese trafo, 1000Uf 25V cap after rectifier and 7824.
No ceramic bypass even. All breadboarded. Quiet recognizable hands which should be cut out on guillotine one day, he!
Audio is quite a religion, so, let's give it some respect.
Than, saw the thing I really became crazy. There was small 15->12+12 dc/dc converter hooked to top of +24V to get total 48V phantom ref to earth.
With small ceramic 0.1uf from "phantom" point to gnd.
Besides the work inside the unit was real crap, the idea of criminal monster gave some lulz to old Addams like me.
The choose of converter was absolutely wrong, the person probably looked the easy way to get money for drugs or cheap booze,
and the noise at +48 was about 20 mV measured with tester in AC mode.
Huh! Guys, again, if you read this: don't do it _this_ way, have a respect to audio! However, try to read it till the end.
Ok,  happens; I wished all the best to both-wrong-left-handed person who racked in hurry V672 and stopped to think about this...

Next, I became crazy on person which asked me if I use DC/DC converters in my units.
Well, for talkback combiner, it is OK to have less than 1mV AC of noise worst case, measured 10hZ-140kHz at power supply rails,
when inserting a weighting filter bla bla it becomes acceptable. The load is low (some milliamp's); so, this noise can be filtered.
For Microrack, the situ is different, say, 15W DC/DC converters aren't quiet in price acceptable range.
To express myself right, high quality DC/DC converters costs way more than we can afford.
I had some experience with 5->+/-15V 5W converters and have to say, finding converter with less than 1mV noise
at nominal load is not easy. I mean 1mV noise when I see it without weighting filters on analyser.

So, the prob of dc/dc converters is: cheap are wrong designed, work at 60 or 100k frequency, and NO GOOD for audio. Not kosher.

Well. 2 weeks later, one guy asked me if I can make universal phantom add-on which will work with
input voltage from +10 to +36V and bring 50-100 ma of clean phantom power at its output.

I reminded the DC/DC discussion thread, of course, you will find some lulz like "simple switcher" for audio use, he!,
http://www.groupdiy.com/index.php?topic=46702

AND after 3-4 hours the PCB design was ready. Not my first experience with PWM PS, so, gone smooth.
The last one successful PWM power supply experiment BTW was conversion of microamperes/kilovolts to volts/milliamperes,
to charge 24V accumulator from piezo-elements installed on road, for one really funny local startup company.


----- -----

OK...received the parts today, stuffed, and after some hours of testing....me happy!

UniPhan is: step-up converter, done audio kosher, (good as krishnait's food too), works at 300...400Khz,
has short circuit / overvoltage protection, stable, effective.
it requires: starground point, +10....36V input, and it's output is clean 48V at any load in 0...100 mA range.

Budget? For me it costed PCB proto and parts. In hundreds, should cost about
$25-30 in x100...200 batch production, if assembled in China, parts + labor only.

Proofed a claim about good DC/DC is not cheap and SHOULD be home brewed.

Well, let's see how good is it.

To be clear, if I mean noise in mV AC, next related to 10Hz-100kHz AUDIO range.
For comparison, normal console's supply gives a noise at 48V rail about 0.5...1 mv.
Inside pedigree pre, maybe it is possible to find shunt phantom regulator with less tha 0.2mV noise...

Well. Here is the baby. Say "Hello, Uniphan!" ;)

UniPhan.jpg


No load

noload.jpg


Test conitions: 15V input, 1k load (50 ma at 48V):

dut.jpg


drawcurr.jpg


50maload2.jpg


50maload1.jpg


100mA load:

100maload2.jpg


100maload1.jpg


Ah yes, to compare. We did a project with one company for CO2 laser power measurement
with Japanese made 5->2x15V/200ma converters. I had some in my closet and took one to workbench.
Today they cost about $25 per converter in hundereds.

25buk_dcdc.jpg


 
Pipe, aye! The unit measures 3 pipes width and 2 pipes height. OK, will do in a free time!
P2P: reason I used SMD components is to get rock solid ground layout, and shortest traces.
p2P seems not the best way here.
However, if say your tube pre-amp, which is TRUE and P2P :) have 12.6 volts DC heater and power transformer have no 24 or 48v winding,
the easy way to add phantom is here.
 
Hey Igor,

What happened? Any news on production of the Uni-Phan? I've been looking forward to ordering a couple of those, did it turn out to be too expensive?

Best,
Dan

 
He's a pretty busy guy... He has multiple support threads to watch over and I'm sure he's in the middle of 3 or 4 projects as we speak.

 
MagnetoSound said:
Hey Igor,
`
What happened? Any news on production of the Uni-Phan? I've been looking forward to ordering a couple of those, did it turn out to be too expensive?

Best,
Dan


Ok... Please check the docs. Everything is here.

http://ij-audio.com/downloads/UniPhan_48.rar

Unfortunately, there is an error on pcb, R2 SHOULD BE 10 OHM INSTEAD OF 10K!!!

There is a file errata.txt and you will find here some Farnell part nr's as well.

The rest is fine. I used r3 = 12k, it can be in range 9...18k, and sets lowest possible input voltage for converter.
Below this voltage, the converter will enter in sleep mode.

The estimate cost for these PCB's is USD16. They will appear at my webstore (ij-audio.com/store) soon.

The estimate cost for the rest of parts is USD40-55 (Farnell), however, most of SMD components should be ordered in 100+ qty's.

I can make full kits in $80 price range, which is reasonable and definitely less than buying parts in hundreds qty's.

Cool than... post interest if interested!


rickc said:
good work there.

nice to see so much USA made test gear on your bench...

Hehe! I bought most of equipment at surplus store, where it came from Israeli army. Most of stuff even have "raphael" calibration tags :)
 
that looks like a good part, with other apps possible.
questions: Is there a minimum load needed?
What is the efficiency like at low load currents?  I'm thinking of this in the context of something that is battery operated.

Please do post when kits are ready for purchase. Might as well pay you for your effort!

thanks,
--rickc
 
rickc said:
that looks like a good part, with other apps possible.
questions: Is there a minimum load needed?
What is the efficiency like at low load currents?  I'm thinking of this in the context of something that is battery operated.

Please do post when kits are ready for purchase. Might as well pay you for your effort!

thanks,
--rickc

Min load: no need for this.
Efficiency is not bed, about 91%. For minimal required current, just dig the DS supplied with docs.
For battery operation, dig something like microphone's capsule converter like in Neumann U87 etc,
it for sure have less min req'd current (=more battery life), but will supply one, maximum 2 mics.

Regarding the kits - no prob, I just have to figure out how to pack all these smd things, most of ceramic caps
have no marking,  parts are very small, etc.
 
Igor said:
MagnetoSound said:
Hey Igor,
`
What happened? Any news on production of the Uni-Phan? I've been looking forward to ordering a couple of those, did it turn out to be too expensive?

Best,
Dan


Ok... Please check the docs. Everything is here.

http://ij-audio.com/downloads/UniPhan_48.rar

Unfortunately, there is an error on pcb, R2 SHOULD BE 10 OHM INSTEAD OF 10K!!!

There is a file errata.txt and you will find here some Farnell part nr's as well.

The rest is fine. I used r3 = 12k, it can be in range 9...18k, and sets lowest possible input voltage for converter.
Below this voltage, the converter will enter in sleep mode.

The estimate cost for these PCB's is USD16. They will appear at my webstore (ij-audio.com/store) soon.

The estimate cost for the rest of parts is USD40-55 (Farnell), however, most of SMD components should be ordered in 100+ qty's.

I can make full kits in $80 price range, which is reasonable and definitely less than buying parts in hundreds qty's.

Cool than... post interest if interested!


rickc said:
good work there.

nice to see so much USA made test gear on your bench...

Hehe! I bought most of equipment at surplus store, where it came from Israeli army. Most of stuff even have "raphael" calibration tags :)

I'm in for a few when they go up!! please notify me if you remember :)
 
Back
Top