15V phantom power

[Andy Peters]

New question.. Do you feel the 5V and 500ma capability are safe? As in safe to depend upon for future products. 5V is already getting long in the tooth for digital logic, so I would expect some pressure to move USB or a USB like interface to 3.3V or lower.

USB 3.0 allows 900 mA, still at a nominal 5V but VBUS is allowed to go down to 4V (and up to 5.25V).

Sure, pretty much nothing digital these days runs on 5V, but if USB changed to a 3.3V supply then they'd have to at least change the connector (much the same way that PCI slots are keyed to indicate 5V-, 5V and 3.3V or 3.3V-only capability). Also, since voltage drop across the cable length is a concern, a lower voltage could be a problem. Anyways, it's easy enough to put a switcher (or LDO) in the product which gives the needed, and regulated, voltages.

-a

 

[irod]

You know, the thing is most customers don't speak electronics. They say, "oh that mic pre doesn't sound right" or "why is my mic noisy?" For instance I know some engineers who absolutely hate Universal Audio 610 preamps, because they "don't sound right" and "overload easily" . Apparently quite a number of them have a faulty/weak P48 supply, and that's probably why. I bought an LA-610 a while ago, and had to send it back because the P48 did not deliver enough current and the voltage dropped ever so slowly (even unloaded). Got another unit that works and sounds pretty good.

If you look at various (studio) mic schematics, you'll find that the majority won't work well with less than real P48. That may be different for stage mics, because stage condensers tend to be electrets (actually, that's changing, too). Weren't the Peavey mics, electrets, too? Ten years ago, lots of entry level studio mics were electrets, but that has changed.

A voltage tripler may be a solution, at least if only a few channels require phantom power. And yes, the vast majority of new products has correct P48 (except USB audio interfaces that are bus powered). It's one of the things I always measure when I review gear.

Hi Rossi

I realize it's been over 10 years, but maybe you remember the context)
I mean UA 610 phantom power issue. I have such a one and see that max P48 current is limited to 3.4mA. When I apply 10kOm load to P48 circuit I get about 2.4mA and +15V phantom DC (33V droppage).
This doesn't seem to be normal because my Neumann TLM49 draws 3.2mA according to the spec and that's uncomfortably close to the M610 P48 limits.
So it looks like the same issue you've mentioned above.

Can you tell me pls - "Got another unit that works" means you've got just the same model (UA 610) and it was good, or you were speaking about some other preamp?

Regards
Yuri

 

[abbey road d enfer]

When I apply 10kOm load to P48 circuit I get about 2.4mA and +15V phantom DC (33V droppage).

Something's not quite right here.
2.4mA into 10k should read 24V.
Unfortunately, the schematics of the 610 and its numerous variations is rarer than hen's teeth, so I think you'll have to trace the schematic of this part.
I think there are only two likely possibilities: either they took the heater winding to a multiplier, or they used a resistive voltage divider to drop B+ to 48V.
In the former case, you would need to increase the capacitor values (maybe they are defective), in the latter, you may alter the divider values (you just need to make sure you don't overload the xfmr.

 

[JohnRoberts]

Something's not quite right here.
2.4mA into 10k should read 24V.
Unfortunately, the schematics of the 610 and its numerous variations is rarer than hen's teeth, so I think you'll have to trace the schematic of this part.
I think there are only two likely possibilities: either they took the heater winding to a multiplier, or they used a resistive voltage divider to drop B+ to 48V.
In the former case, you would need to increase the capacitor values (maybe they are defective), in the latter, you may alter the divider values (you just need to make sure you don't overload the xfmr.

good point, old cap doubler/triplers could easily have tired electrolytic capacitors that are not pumping enough charge.

JR

 

[irod]

Something's not quite right here.

Right) I was wrong where - not 2.4 but 1.2mA. Load is 10k and there are two current-limiting resistors inside M610 that feed the signal lines. Looks like there're two resistors about 50-60k each.

Droppage on these resistors is 60*1.2/2 = 36v. Droppage on 10k load is 12v.

I suppose such a schematic

So when I apply a load to hot, having cold unloaded, current flows through both the resistors and the xfmr primary coil (it's DC resistance is quite low, about 50 Ohms I suppose). When I measure the current near load I get the total current - 1.2mA on 10K load. Or 1,6mA on near-zero load - 48/(60K/2) = 1.6mA
And really, if I apply load to hot, I get a droppage from +48v to +15v on hot and cold both.

So the situation is much worse than I imagined first - maximum total P48 current is only 1.6mA. Very bad for my Neumann)

I believe there are no problems with power supply, since the output is stable +48V - regardless of the load.

>ou may alter the divider values (you just need to make sure you don't overload the xfmr).

That's the point))
Standard says the P48 feeding resistors should be 6.81K. Not 60K. And max P48 current is 48/(6.81/2) = 14mA. Not 1.7mA.
The only reason I can think of is that the UA developers took care of the health of the transformer... just for the case when hot/cold is shortened to the gnd.

Can you tell me the critical DC current for UTC-O1 xfmr primary coil?))
http://bitsavers.trailing-edge.com/components/utc/1963_UTC_Catalog.pdf - page 18.. DC resistance on primary = 52 Ohms

Regards
Y.

 

[ccaudle]

Looks like there're two resistors about 50-60 kOm each.

Phantom power standard is 6k8 Ohms each resistor.
[edit: sorry, was originally reading on my phone and did not see that you had already pointed that out farther into the post]

 

[irod]

if UTC-O1 max capacity is 6.3mW (acc to spec) -> max current = SQRT(6.3E-3/52) = 11mA... Hmm...

 

[abbey road d enfer]

if UTC-O1 max capacity is 6.3mW (acc to spec) -> max current = SQRT(6.3E-3/52) = 11mA... Hmm...

This is the AC power that the xfmr can handle without excessive distortion. DC is certainly much higher. Max DC current is dominated by wire gauge. I've never seen thinner than AWG38 for a mic input. Max DC for AWG38 is 130 mA.
Now there is the issue of remanent induction, which happens when a winding is submitted to DC current. It shifts the B-H curve and results in distortion. It can easily be fixed by submitting the xfmr to AC and reducing progressively the voltage, which is typically what happens when submitting the xfmr to a loud source. Alternatively, one can use a tape demag (which used to be a staple of recording studios when tape was the de facto medium).

 

[abbey road d enfer]

I suppose such a schematic
View attachment 101215

Which is clearly wrong...

The only reason I can think of is that the UA developers took care of the health of the transformer... just for the case when hot/cold is shortened to the gnd.

When I wrote "you just need to make sure you don't overload the xfmr", I was referring to the power xfmr.
I would vouch no half-decent xfmr can die of being shorted to ground with phantom on.
You need to find out how 48V is produced in the 610. I can understand that if they chose to take it from B+ via a voltage divider, they would have settled for values that don't result in excessive dissipation, at the cost of not being in conformity with the standard.
OTOH, it may be possible they have not considered the standard, but rather experiment with existing mics.
It may seem far-fetched, but I have personally experienced an API mixer where 10k resistors inadvertently replacing 100r's resulted in correct operation with Neumann U47FET, but silence with AKG451.

 

[irod]

I will definitely trace the power supply wiring. Although I'm almost sure it's all ok in this part, because +48v control point I've found on the power supply PCB - I mean the point where +48v takes its origin))), - is 100% stable and always = +48v despite of load on hot/cold pins.
There is some TO-220 chip on PCB, probably a DC regulator, I'll take a look. From the other hand it may be +12v supply for 12AX/AT heating.

Thank you
Y.

 

[abbey road d enfer]

There is some TO-220 chip on PCB, probably a DC regulator,

OK, if for phantom, very likely to be a 783 or a 317HV, although some reckless use a standard 317 in that position.

From the other hand it may be +12v supply for 12AX/AT heating.

Then more likely to be a 7812 or a standard 317.

 

[irod]

Now there is the issue of remanent induction, which happens when a winding is submitted to DC current. It shifts the B-H curve and results in distortion.

I believe, 2-resistors scheme of phantom power feeding ensures it's no DC current in normal operation. DC current in primary winding can only flow when a hot or cold short to ground occurs - by accident, by a broken cable, or by improper soldering.

 

[abbey road d enfer]

I believe, 2-resistors scheme of phantom power feeding ensures it's no DC current in normal operation. DC current in primary winding can only flow when a hot or cold short to ground occurs - by accident, by a broken cable, or by improper soldering.

Correct. Things happen, that's why some attention must be brought to it.

 

[irod]

LOL

 

[JohnRoberts]

View attachment 101250
LOL

LOL.... mics that required 48V phantom have been around for as long as I was paying attention but many budget value mics were designed to operate properly from lower, easily available voltages, like typical rail voltages 12-24v.

I recall at Peavey that most install amps with phantom power were using low voltage. We increased the phantom voltage available in even budget top box mixers to 48V in the late 80s/early 90s mostly as a marketing feature, not out of any compelling need.

JR

 

[irod]

Here is the P48 supply schematic for UA M610. Traced by myself) May put it to archives/FAQ.
The devil definitely is in 22k resistor.
That's ridiculous - on the one hand, they followed the standard and used 2x6k4 feeding resistors.
But on the other - inserted 22k before them))

 

[abbey road d enfer]

The devil definitely is in 22k resistor.
That's ridiculous - on the one hand, they followed the standard and used 2x6k4 feeding resistors.
But on the other - inserted 22k before them))

Looks very much like the mistake I mentioned in another thread, regarding the API console that had 10k resistors there, instead of the nominal 100r.
Well, fortunately, you have done the hard work, tracing the circuit; fixing the problem should not be too difficult, I hope.
It's worth reporting to UA, they may not be so arrogant when they figure out they're not so perfect.
Sorry for ranting, but a company that does not publish their schemos seems fishy to me.
Actually, a 220 ohm resistor there reflects as 440 ohms in each branch, resulting in a total of 6.84kohms, a good match for the standardized 6.8 (or 6.81)k.
Someone at UA must know what they do, whilst others let products go to the market with major issues.

 

[JohnRoberts]

Mistakes happen, tell their customer service and hopefully they will make corrections.

JR

 

[irod]

))) I commited a ticket to UA support ~ a week ago) no reply yet.
Yes, I was thinking to replace 22k with 100R, but your proposal about 220 ohm looks very interesting, thank you)

 

[irod]

it is enough to confuse just one color strip)