Generic problem with microphones,phantom powering and op-amps?

[N Cook]

Actually a repair problem for a low noise microphone pre-amp but seems to be
a general design flaw.
Pre-amp uses a Burr Brown INA103 very low noise instrumentation op-amp.
In this M-Audio Omni i/o preamp and an outline design application in the
Burr Brown book show much the same circuitry.
The 48Volt phantom supply to the mike is protected by 6.8K limiter
resistors.
But to block the 48V DC to the op-amp there is a 10uF/100V electrolytic in
each line directly to the inv & non-inv i/p of the op-amp .

If , as seems in this case, a balanced line microphone with a short to
ground is connected to such a system
then the +48V / 0V across the elecrolytic will instantneously go to 0V
/ -48V with -48V
directly connected to the op-amp i/p powered from +-15V rails and according
to the databook
can be taken to only +-12V.

Blown input to this op-amp due to just the owner connecting a microphone.
Anyone familiar with this, adding limiting diode pair at each input ?
I see no point in replacing this 15 GBP/ 25 USD IC until this design flaw
is attended to or it will happen again should a fault to ground develop in a
mike or lead while the 48V power is on.

 

[John Woodgate]

Anyone familiar with this, adding limiting diode pair at each input ?

Yes, diodes to each +/-15 V rail.

Part of the price you pay for not using a transformer.

 

[Phil Allison]

"N Cook"

Actually a repair problem for a low noise microphone pre-amp but seems to
be
a general design flaw.
Pre-amp uses a Burr Brown INA103 very low noise instrumentation op-amp.
In this M-Audio Omni i/o preamp and an outline design application in the
Burr Brown book show much the same circuitry.
The 48Volt phantom supply to the mike is protected by 6.8K limiter
resistors.
But to block the 48V DC to the op-amp there is a 10uF/100V electrolytic in
each line directly to the inv & non-inv i/p of the op-amp .

If , as seems in this case, a balanced line microphone with a short to
ground is connected to such a system
then the +48V / 0V across the elecrolytic will instantneously go to 0V
/ -48V with -48V
directly connected to the op-amp i/p powered from +-15V rails and
according
to the databook
can be taken to only +-12V.

Blown input to this op-amp due to just the owner connecting a microphone.
Anyone familiar with this, adding limiting diode pair at each input ?
I see no point in replacing this 15 GBP/ 25 USD IC until this design flaw
is attended to or it will happen again should a fault to ground develop in
a
mike or lead while the 48V power is on.

** The use of back to back zeners to protect sensitive input devices is
standard practice with phantom powered mic pre-amps. See figure 23 on this
pdf for how it is commonly done:

http://pdf.alldatasheet.co.kr/datasheet-pdf/view/49084/AD/SSM2017.html


Hard to believe any commercial product has no protection against the common
phantom accident.



.............. Phil

 

[N Cook]

"N Cook"



** The use of back to back zeners to protect sensitive input devices is
standard practice with phantom powered mic pre-amps. See figure 23 on this
pdf for how it is commonly done:

http://pdf.alldatasheet.co.kr/datasheet-pdf/view/49084/AD/SSM2017.html


Hard to believe any commercial product has no protection against the common
phantom accident.



............. Phil

Many thanks for the link. Not only no protection in this 300 USD as new,
year 2000
made, bit of kit but no mention in Burr Brown data book book either.
Previous repair job concerning problem with phantom, as it was never used ,
except when someone accidently engaged it, I totally disabled that function
on that large mixer/amp

 

[John Woodgate]

Many thanks for the link. Not only no protection in this 300 USD as
new, year 2000 made, bit of kit but no mention in Burr Brown data book
book either. Previous repair job concerning problem with phantom, as it
was never used , except when someone accidently engaged it, I totally
disabled that function on that large mixer/amp

There are plans to produce an Audio Engineering Society standard on
input protection, but progress is slow.

 

[martin griffith]

Many thanks for the link. Not only no protection in this 300 USD as new,
year 2000
made, bit of kit but no mention in Burr Brown data book book either.
Previous repair job concerning problem with phantom, as it was never used ,
except when someone accidently engaged it, I totally disabled that function
on that large mixer/amp

also look at the other "typical " method on P12 of TI's PGA2500


martin

 

[Pooh Bear]

Actually a repair problem for a low noise microphone pre-amp but seems to be
a general design flaw.
Pre-amp uses a Burr Brown INA103 very low noise instrumentation op-amp.
In this M-Audio Omni i/o preamp and an outline design application in the
Burr Brown book show much the same circuitry.
The 48Volt phantom supply to the mike is protected by 6.8K limiter
resistors.
But to block the 48V DC to the op-amp there is a 10uF/100V electrolytic in
each line directly to the inv & non-inv i/p of the op-amp .

If , as seems in this case, a balanced line microphone with a short to
ground is connected to such a system
then the +48V / 0V across the elecrolytic will instantneously go to 0V
/ -48V with -48V
directly connected to the op-amp i/p powered from +-15V rails and according
to the databook
can be taken to only +-12V.

Blown input to this op-amp due to just the owner connecting a microphone.
Anyone familiar with this, adding limiting diode pair at each input ?
I see no point in replacing this 15 GBP/ 25 USD IC until this design flaw
is attended to or it will happen again should a fault to ground develop in a
mike or lead while the 48V power is on.

This is classic. I suppose Burr Brown missed the 'usual method' of protecting
the inputs.

Add 4R7s in series with the INA inputs and fit clamp diodes to both supplies
supply at the INA inputs. Use decent size parts. It'll blown an 0805 R for sure
!

I'm sure one of the other TI mic amp parts hows this on the schematic ( see
PGA2500 for example ) and I'm sure the SSM parts have it too.

Graham

 

[Pooh Bear]

"N Cook"

** The use of back to back zeners to protect sensitive input devices is
standard practice with phantom powered mic pre-amps. See figure 23 on this
pdf for how it is commonly done:

http://pdf.alldatasheet.co.kr/datasheet-pdf/view/49084/AD/SSM2017.html

Hard to believe any commercial product has no protection against the common
phantom accident.

............. Phil

Quite possibly because it requires some beefy parts to be effective.

Graham

 

[Phil Allison]

"Pooh Bear"

Quite possibly because it requires some beefy parts to be effective.

** Hardly a rational explanation.

No surprise from Stevenson the Fake.



............. Phil

 

[Pooh Bear]

"Pooh Bear"

** Hardly a rational explanation.

No surprise from Stevenson the Fake.

............ Phil

I've seen 0805 Rs fail open in that type of protection. The component needs
to have some ability to dissipate the energy. Check out the peak current
too.

Graham

 

[Phil Allison]

"Pooh Bear"

I've seen 0805 Rs fail open in that type of protection.

** Hardly a rational explanation.

Only a total idiot would use that as an excuse for no fitting protection.

Pooh just did.




........... Phil

 

[Winfield Hill]

martin griffith wrote...

also look at the other "typical " method on P12 of TI's PGA2500

Ahem. "It is recommended that the Schottky diode chosen for this
application be specified for at least a 10A surge current."
Here Burr-Brown recommends ON Semi's monster mbra120 diode,
http://www.onsemi.com/site/products/summary/0,4450,MBRA120ET3,00.html

But sheesh, that beefy leaky high-capacitance diode is spec'd to
handle 40A for 8ms! The surge current is much less, depending on
the series resistor used. For example, 2x 27 ohms (0.9nV of noise)
limits the 48V surge to 2A lasting less than 1.5ms (if Cin = 47uF).
A nice little SD103 can handle that with oodles of room to spare.

 

[Fred Bloggs]

martin griffith wrote...




Ahem. "It is recommended that the Schottky diode chosen for this
application be specified for at least a 10A surge current."
Here Burr-Brown recommends ON Semi's monster mbra120 diode,
http://www.onsemi.com/site/products/summary/0,4450,MBRA120ET3,00.html

But sheesh, that beefy leaky high-capacitance diode is spec'd to
handle 40A for 8ms! The surge current is much less, depending on
the series resistor used. For example, 2x 27 ohms (0.9nV of noise)
limits the 48V surge to 2A lasting less than 1.5ms (if Cin = 47uF).
A nice little SD103 can handle that with oodles of room to spare.

THAT recommends 4.7 ohms series and 1N5819 bridge rectifier...