phantom power noise

[volki]

tmbg - if your u87 version is the one with the battery space in the mic body, you definitely don't wanna run it at less than a few volt's off 48 - the capsule polarization voltage is directly taken off the phantom voltage. the u87a doesn't really care though, since it has a dc-dc converter, and the polarization is at 60v.

prr - maxwell's demon? what's that all about? prolly a totally different expression here in germany, the english version of which i just can't seem to relate to...? =)

 

[SSLtech]

[quote author="xvlk"][quote author="SSLtech"]Noise on the phantom supply will be cancelled to the degree that the mic preamp rejects common mode signals. [/quote]
Dear SSL and the others.
Somewhat of you answer can become unclear and to obey thermodynamic law.
There are two kinds of phantom - circuit noise:
1) The noise of phantom power supply unit
2) The noise of phantom circuit (resistors, etc).

...<snip happens>...

xvlk[/quote]

Understood. I was specifically adressing the question of noise on the phantom supply as per the original question. :grin:

Two uncorrelated noise sources -one on each leg- will of course lead to a non-rejected noise. :green:

Keith

 

[PRR]

> maxwell's demon? what's that all about? prolly a totally different expression here in germany, the english version of which i just can't seem to relate to...? =)

http://www.auburn.edu/~smith01/notes/maxdem.htm
http://en.wikipedia.org/wiki/Maxwell's_demon
http://www.maxwellian.demon.co.uk/name.html
http://english.ttu.edu/kairos/3.1/coverweb/porush/Demon.html
http://www.amazon.com/exec/obidos/tg/detail/-/0750307595
http://www.amazon.de/exec/obidos/ASIN/0750307595
http://buecher.compricer.de/0750307595

Although Max's demon was just a thought-experiment, there is a practical hot/cold separation device that has no moving parts, and appears to work like that demon. Hot and cold molecules are separated with centrifugal force and clever chambers and ports.

http://www.visi.com/~darus/hilsch/

If you measure its efficiency, it isn't working for "free". The amount of change in entropy and air pressure is less than the air-power you have to put in. It will not replace your home heat/cooling systems. But it has been found commercially useful in machine shops, where pressurized air is readily available and a small blast of cold air keeps drills and lathe-cutters cool under heavy cutting loads.

http://www.exair.com/vortextube/vt_theory.htm
http://www.exair.com/product_info/product_info_page.htm
http://www.exair.com/vortextube/vt_page.htm
http://www.exair.com/spotcooling_products/ct_page.htm

 

[volki]

whoah, what a flood of url's :shock:
actually some are digging deeper than i'd like to know. but i get the picture. thanks! :green:

 

[synthetic]

I know that this is basic Ohm's law stuff and that I'm an idiot, but what is the formula to calculate the series resistor to ground for a phantom power LED? I've looked at a few schematics, but many power supplies seem to send 15V or so as phantom power (like the Rane) so I don't want to follow that schematic. Also, if I remember my Ohm's Law correctly, don't I increase current when I add resistance? Or does the LED draw what it needs?

Sorry, I just don't want to fry my $5 blue LED. Maybe I should get a few $0.01 red LEDs to sacrifice to the blue smoke gods.

 

[PRR]

> I know that this is basic Ohm's law stuff..., but what is the formula to calculate the series resistor to ground for a phantom power LED?

R=V/I

What I (current) do you want? Most LEDs are rated 20mA (0.020A) max, but they all work on less, and most are pretty bright at 5mA.

What is your available voltage? 48V? The LED eats some of that: 1.7V for red LEDs and more for other colors, maybe over 3V for Blue? In a 5V circuit, that makes a difference, so you better check the specs. In a 48V circuit, 1.7V or 4V makes very little difference: 48V-1.7V= 46.3V, 48V-4V=44V, only 5% difference, and the current is not at all critical.

So given about 48V-2V= 46V, and say 5mA (0.005A) current, 46V/0.005A= 9,200Ω. Sort of an odd value. I find a 10K on the floor, will that work? 46V/10,000= 0.004,6A or 4.6mA. This is safely below the 20mA maximum (even if the 10K 10% resistor is really 9K or 11K) and reasonably close to the 5mA operating current that we picked out of thin air.

Or hey: we have a baggie of 6.8K resistors left over from Phantom wiring. 46V/6K8= 6.76mA, more than our thin-air 5mA figure but lower than the typical 20mA rating of an LED. Try it. If too bright, try two 6K8 in series for 46V/13K6= 3.4mA.

Ah, you learn that your particular LED is 4V, not 1.7V or my rounded 2V. So 48V-4V is 44V. 44V/6K8 is 6.47mA, a fine current. In general, if the supply voltage is much more than the LED voltage, and you are not desperate for all the bright you can get, these calculations can be very rough.

You need to check heat. When things are well, this will be just volts times amps. Say 46V*0.0067A= 0.31 Watts. A 1/4 Watt resistor will run VERY hot, drift in value, and generally be no good. Use a 1/2 Watt resistor and let it get plenty of air.

As an extreme case: you work in the sunlight (ugh!), you need all the bright you can get. You want to run the LED up to 19.9999mA. 46V/20mA= 2K3 resistor. This is an odd value, but you can't use the nearest 2K2 value: 46V/2K2= 20.91mA. The LED might live forever with 5% abuse, but it seems like a bad plan. In 10% standard values, you are stuck with 2K7. 46V/2K7= 17mA. Less than max, but you will hardly notice a 15% change in brightness (the eye is logarithmic like the ear). Now we have 46V*0.017A= 0.78 watts, and even a 1 Watt resistor will run hot.

 

[PRR]

> if we afix a piece of bread to a cat's feet with the butter side facing the cat's belly, and throw the cat out of a 13th story window....what happens?

Murphy is quite clear: the unbreakable tape/glue/nails used to afix the bread to the cat WILL break, the bread will fall free, the cat will land on its feet and the bread WILL fall on its butter.

This despite the fact that cats are not subject to Murphy's Law. (The cat knocks over the dish that can't be knocked over, the cat sees nothing wrong with that. Murphy got YOU, not the cat.)

There is an explanation why bread falls butter-to-floor, based on the relative heights of tables and heads. It holds true over a wide range of hypothetical bread-eaters who use tables, not just human-height.

 

[synthetic]

[quote author="PRR"]What I (current) do you want? Most LEDs are rated 20mA (0.020A) max, but they all work on less, and most are pretty bright at 5mA...[/quote]

Thanks for the thorough response! My LED has a big chart on the back that I can't make sense out of. "6V, 91 ohms @ 30MA, 120 ohms @45MA, 1/2W resistor." Then, on another line, "1.5VDC: 100 ohms/2MA, 100 ohms/2.5MA. 3 VDC: 620 ohms/2MA, 470 ohms/2.5MA. 6VDC..."

Should I try a 10k resistor? I believe my supply produces 46V, so 46V-6V (?) = 40V. 40V/0.002A = 20000. 40V/0.005A = 8000. I'm leaning towards 10k's in parallel to rate at 1/2W.

I'll post pictures if it doesn't explode. Hell, even if it does.

 

[tmbg]

[quote author="volki"]tmbg - if your u87 version is the one with the battery space in the mic body, you definitely don't wanna run it at less than a few volt's off 48 - the capsule polarization voltage is directly taken off the phantom voltage. the u87a doesn't really care though, since it has a dc-dc converter, and the polarization is at 60v.[/quote]

Hm, that might explain a strange phenomenon I've been experiencing with this mic... it seems like at moderately high SPLs (like a vocal plosive) the U87 cuts out, almost like it 'resets', loses its phantom for a second... perhaps it's running on a slightly below spec phantom, and a cap is keeping it going or something?

It is indeed the old one with the battery space. Is there anything in these mics that might need to be replaced due to its age? Any special mess I can do to keep it in good condition?

 

[Gus]

Go here

http://www.groupdiy.com/index.php?topic=256

Look in the external link section.

Also take a look at the current old U87 thread. Could you post the Fet used and the soure R selected?

 

[tmbg]

I'll take a look when I get home... at work at the moment. Sadly I dont have teh intarweb at home, so I'll have to write down what I find and post in the morning, SO

post me everything about the mic I should look for, and I'll also try to take some pics.

 

[Gus]

In the u87 I have seen inside I have seen k170 fets with 2.2k source and a 2n3819 with a 10K source all looked like done at the factory.. The source R is adjusted for about 10V at the drain. I have an Idea of the sweet spot source R for input and ouput headroom and distortion.

Take the body tube off the fet is at the top right side of the board. Source R is the resistor soldered to to posts for selecting.

 

[volki]

i heard that certain versions of the u87 come with a non-ideal bias point, resulting in unequal distortion thresholds for both half-waves and thus lower headroom. not sure if that rather applies to u87a, though. if the latter is true, that would explain why people like the non-a version better. any comment, gus?

 

[volki]

... it seems like at moderately high SPLs (like a vocal plosive) the U87 cuts out, almost like it 'resets', loses its phantom for a second... perhaps it's running on a slightly below spec phantom, and a cap is keeping it going or something?

that rather sounds like there's something wrong with the capsule, like perforation of membrane due to age, or maybe just other deficiencies in the capsule isolation. in this case, if you aspire into the mic, the moisture in your breath settling on the capsule can provide paths for eddy currents, in extreme cases to the point of shorting out the polarization. after evaporation of the moisture, the effect vanishes.

 

[tmbg]

Can cleaning the diaphragm fix this? If so, I know how to remove the capsule from the body, but how do you go about removing the grille to get at the diaphragms?

 

[volki]

definitely check the external links, they provide the info you need i guess. in particular:

http://www.josephson.com/clean.txt

or for info on u87 and tons of other mic's

http://recforums.prosoundweb.com/index.php/f/33/0

 

[Gus]

I left a hint in one of my posts. The problem described is the dirt at the edge of the capsule that bleeds the charge from the capsule problem. Breath having mosture forms a conductive path along the dirt discharging the skin when it dries the capsule can hold a charge again.

The fix is to clean or replace the capsule. Remember a K87 capsule costs some money if you mess it up. I practiced on china capsules before I even tried to clean a neumann or brauner. When capsules have not been cleaned for some time the gold can flake off with the most careful cleaning. For a k87 type with the film in bad shape just clean the clear section.

The flaking off of the gold film is something I am trying to understand, I think the gold being a thin film has small pathways to the PET and the smoke,dirt,spit etc move thur with time because of water vapor etc. Some kind of chemisty(ever look at the mess you cleanoff the capsule) takes place between the gold film and the PET film that allow the gold to PET bond to be broken. I would like to see an electron microscope view of the surface of the PET to Gold contact.

When the capsule is new the gold to PET has a strong bond.

Any Chemists here that can help out with this puzzle?

 

[tmbg]

yeah I read all about cleaning them, I just don't know how to get at the capsule, how to take the grille off.

 

[StephenGiles]

I use a Rode NT 1000 with BEHRINGER MXB1002 mic amps - result - silence! Am I doing something wrong????
Stephen

 

[usekgb]

I know this is an ancient thread, but I just have one question about this power supply. How big is the PCB that this layout goes on to? I have found the .GIF file, but I have no idea what size to scale it to. I need to build up a few supplies just like this one, but I have absolutly no experience in designing PCB's. I can etch boards from existing layouts just fine, so this is no problem. Has anyone made a PDF of this layout that maintains the correct size when it is printed? Thanks for the help.

Cheers,
Zach