API 312 - Phantom Issue

[sircletus]

If the circuit is anything like the above CAPI schematic, put a 10K from the unused lug of the 48V switch (SW1) to ground. That way, it only drains when switched out so you can drain quick with 10K and it saves every minutia of juice when 48V is on. If it's not the same circuit, just put a 100K across the LED. An extra 500uA isn't going to make a lick of difference.

API didn't provide schematics, so I'll pull one of the pres and take a look at the switch.  Otherwise, putting a resistor across the LED would certainly be easiest to do in a pinch.

 

[Bo Deadly]

API didn't provide schematics, so I'll pull one of the pres and take a look at the switch.  Otherwise, putting a resistor across the LED would certainly be easiest to do in a pinch.

Actually if you find a spot where the LED has through hole leads into a PCB, the easiest solution might be to find an SMD resistor that is just the right size to tack on to the back of the PCB across the LED. That could be really clean.

 

[Audio1Man]

As squarewave suggested use SW1 to discharge the phantom power by adding 100 -1k  ½ watt resistor to ground. R5 must be ½ watt or larger, don’t use a small SMD type.

I did a design of a mixer for Mackie and the production engineers made a change of my 100 ohm ½ watt axial resistor (R5) to 100 ohm 0805 SMD, the phantom power turned on some time once before opening. SMD don’t have any mass and the incoming power pulse burned them open.
Duke

 

[pucho812]

API didn't provide schematics, so I'll pull one of the pres and take a look at the switch.  Otherwise, putting a resistor across the LED would certainly be easiest to do in a pinch.

What kind of manufacture does not provide schematics for a large format console that probably cost an arm and a leg

 

[sircletus]

Actually if you find a spot where the LED has through hole leads into a PCB, the easiest solution might be to find an SMD resistor that is just the right size to tack on to the back of the PCB across the LED. That could be really clean.

The LED's on these are actually soldered to flat pads; no through-holes, but yeah, I had thought of doing EXACTLY what you said.  But I found the "off" switch poles and, as Duke suggested, I'm going to try a 1/2W 10k between those and ground and see how that works.  The full 48V discharging through a 10k winds up dissipating just about a 1/4W, and I don't want to push my luck.  I'll use a spare this weekend and see what happens.

And pucho, I feel you.  The Vision cost several arms and multiple legs.  Fortunately, none of them were mine!  But it's fallen on me to do most of the maintenance.  I will say, though, that three years in we've only had a few hiccups and API has repaired/replaced with no questions asked.  Thank goodness for their 5 year warranty!

 

[Bo Deadly]

As squarewave suggested use SW1 to discharge the phantom power by adding 100 -1k  ½ watt resistor to ground. R5 must be ½ watt or larger, don’t use a small SMD type.

100R? That's hardly necessary. With a 10K it would take less than a second to discharge the cap. With a 100R, that's 48/(100+200)*48 = 7.7 watts! Even though it's for an instant R5 and your 100R might need to be more than a 0.5 watt. With 100K across the LED, it would be 21 mW so SMD would be just fine. And it would drain down to < 2V quick because of the LED. Then the 100K would drain off the rest over a few seconds.

 

[abbey road d enfer]

Hey, Duke.  As a matter of fact, this API-lack-of-bleeder-resistor issue has been destroying the FET buffers of my active AEA microphones.  I've measured around a volt and a half sitting on the 212L inputs 15 minutes after turning off phantom.

That is very dubious. The active stage is designed to operate with P48 permanently engaged. I can't conceive how sumitting it to a lesser voltage would destroy it.
In the particular case of the 212, the input xfmr ensures there is no differential voltage either.

  See here:

https://groupdiy.com/index.php?topic=77263.0

This thread results in the conclusion that the destructive mechanism is shorting powered lines when patching with TT jacks.
So the actual full explanation is that, although the phantom is switched off, there is still a significant resisdual, and that patching the mic should be done only when the voltage is enough discharged to be unharmful.
Indeed, adding a bleeder resistor would make the discharge time more controlled.


[/quote] Anyway, this is an issue across our entire Vision console, and I'd like to spend the summer modifying all of the mic pres with a proper bleeder resistor.  Any idea where the best place to put one might be?  I was thinking of putting a large value resistor across the 10µ reservoir cap...
[/quote] Not the best choice. If you do that, this resistor would constantly draw wasted current.
A 10k resistor would result a discharge time of about 2 seconds for reaching a non-critical voltage (6V), but it would draw 4.8mA.
It is much better to use the until then unused NC pole os the switch to connect to a 10k resistor.
See attachment.

 

[sircletus]

That is very dubious. The active stage is designed to operate with P48 permanently engaged. I can't conceive how sumitting it to a lesser voltage would destroy it.
In the particular case of the 212, the input xfmr ensures there is no differential voltage either.
This thread results in the conclusion that the destructive mechanism is shorting powered lines when patching with TT jacks.
So the actual full explanation is that, although the phantom is switched off, there is still a significant resisdual, and that patching the mic should be done only when the voltage is enough discharged to be unharmful.
Indeed, adding a bleeder resistor would make the discharge time more controlled.


Anyway, this is an issue across our entire Vision console, and I'd like to spend the summer modifying all of the mic pres with a proper bleeder resistor.  Any idea where the best place to put one might be?  I was thinking of putting a large value resistor across the 10µ reservoir cap...
Not the best choice. If you do that, this resistor would constantly draw wasted current.
A 10k resistor would result a discharge time of about 2 seconds for reaching a non-critical voltage (6V), but it would draw 4.8mA.
It is much better to use the until then unused NC pole os the switch to connect to a 10k resistor.
See attachment.

I could have stated that far better.  The lesser voltage is not destroying them.  The ~1.5V (the forward drop voltage of the phantom indicator LED) lingering on the inputs to the 212Ls after disengaging phantom and un-patching them is destroying them.  I have measured that voltage lingering for more than 15 minutes after disengaging phantom power.  Thus the need for bleeder resistor.

I ended up putting a 56k 1/4W between unused pole and ground.  Discharge time is acceptable, and the smaller package allows me to scab it onto the bottom of the PCB without worrying about making contact with adjacent modules.

In the 212, btw, the cap is 10µF instead of 100µF.

EDIT:  For anyone curious, I just measured 1.2 VDC at the input of a 212L 30 minutes after turning off phantom.  512s were no different, so I guess I'll be modifying those as well.

In contrast, after 30 seconds, my Millennias, BAEs, Vintechs, and Chandlers all read 5mV or less. 

 

[Audio1Man]

In the circuit R5 is 200 ohms and C7 is 120uf. R5 has not failed, then by adding a GROUND connection to the OPEN pin of SW1 the RC time ON & OFF are the same.
Duke

 

[sircletus]

In the circuit R5 is 200 ohms and C7 is 120uf. R5 has not failed, then by adding a GROUND connection to the OPEN pin of SW1 the RC time ON & OFF are the same.
Duke

Hey, Duke.  The 212 is a little different from the CAPI.  Image attached, if you're curious.

 

[Bo Deadly]

Hey, Duke.  The 212 is a little different from the CAPI.  Image attached, if you're curious.

475R? Hrm. That's a little high. At 10mA that would drop 4.75V. Presumably the 48V source is more like 52V?

 

[Khron]

At 10mA it would also drop 35v across the pair of 6.98k resistors, so that point is almost moot..?

But sure, there's a good chance they also bumped the "raw" voltage up a bit to compensate, as well. Not that it's a tight-tolerance thing to begin with, though...

 

[Bo Deadly]

At 10mA it would also drop 35v across the pair of 6.98k resistors, so that point is almost moot..?

Ehhh. The 475R is in series with both 6.8K in parallel so that equates to a difference of 3.9K vs 3.4K which at 10mA would leave only a little more than 9V at the mic as opposed to 14V. A circuit designed for 14V might not work nearly as well on 9V.

But I admit 10mA DC is probably not realistic. I was just thinking worst-case-scenario.

 

[sircletus]

Ehhh. The 475R is in series with both 6.8K in parallel so that equates to a difference of 3.9K vs 3.4K which at 10mA would leave only a little more than 9V at the mic as opposed to 14V. A circuit designed for 14V might not work nearly as well on 9V.

But I admit 10mA DC is probably not realistic. I was just thinking worst-case-scenario.

I measured 47.7V at the L200 power supply.  It's sitting right at 47V on the 212 mic inputs.

 

[Bo Deadly]

I measured 47.7V at the L200 power supply.  It's sitting right at 47V on the 212 mic inputs.

That's without load. You would have to hook up some mics or DIs or things that use phantom.

 

[AlexK_NL]

Hey, Duke. As a matter of fact, this API-lack-of-bleeder-resistor issue has been destroying the FET buffers of my active AEA microphones. I've measured around a volt and a half sitting on the 212L inputs 15 minutes after turning off phantom. See here:

AEA active mics aren't playing well with API 212Ls. SOLVED

Anyway, this is an issue across our entire Vision console, and I'd like to spend the summer modifying all of the mic pres with a proper bleeder resistor. Any idea where the best place to put one might be? I was thinking of putting a large value resistor across the 10µ reservoir cap...

dear sir,

just curious, did you fixed the issue on your vision console?
i am working on one and looking for the schematic of the 550L 4band EQ, are you able to assist?

regards, Alex

 

[pinebox]

Glad its not just me. I am encountering the same issue on a 2488 with a 48v switch mod. After this thread, I will add a 56k on the modules in parallel with C16? (Highlighted) Or should I use a higher value for less current draw since there is no LED indicator circuit to worry about with benefit of less waste?

 

[abbey road d enfer]

Glad its not just me. I am encountering the same issue on a 2488 with a 48v switch mod. After this thread, I will add a 56k on the modules in parallel with C16? (Highlighted) Or should I use a higher value for less current draw since there is no LED indicator circuit to worry about with benefit of less waste?

There's an additional problem with the 2488.
Resistor R23 (10k) should actually be 100r.
This resistor is there to minimize the current surge that happens when the phantom switch is engaged, since it charges the 100uF capacitor.
With the wrong value of 10k, the max current is limited to 3.5mA. Many mics simply don't work in that case, or in starved mode. A U47FET or U87 work fine because they draw less than 1mA.
I experienced this problem a long time ago, when I designed my first DI box at Barclay Studios, wondering why it worked perfectly with the Neve and not with the API.
I don't know the origin of this mistake but IMO it should be corrected on existing mixers.

 

[pinebox]

There's an additional problem with the 2488.
Resistor R23 (10k) should actually be 100r.
This resistor is there to minimize the current surge that happens when the phantom switch is engaged, since it charges the 100uF capacitor.
With the wrong value of 10k, the max current is limited to 3.5mA. Many mics simply don't work in that case, or in starved mode. A U47FET or U87 work fine because they draw less than 1mA.
I experienced this problem a long time ago, when I designed my first DI box at Barclay Studios, wondering why it worked perfectly with the Neve and not with the API.
I don't know the origin of this mistake but IMO it should be corrected on existing mixers.

Thanks for sharing your experience, I appreciate it. I will swap that out when I do the bleed resistors.