low leakage input protection...diodes or ???

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

Freq Band

Well-known member
GDIY Supporter
Joined
Jan 5, 2006
Messages
608
Location
Electra City
BACKGROUND....
Regarding circuits (or opamps) that prize their low bias current....

Adding input protection in the form of back-biased small signal diodes, often kills the circuit's desired low bias current, from excessive leakage.

As mentioned for example, in the OPA627 datasheet (FIGURE 7)....
http://focus.ti.com/lit/ds/symlink/opa627.pdf
Robert Pease (among others) has also written about using the C and B (E open) pins on a typical transistor say, a 2n3906...or the G and D+S pins of the best N-FET for the job...2n4117.
And TI has an app note... http://focus.ti.com/lit/an/sboa058/sboa058.pdf

QUESTION...
I am wondering if there are disadvantages, of these schemes, (besides the bias/leakage issue discussed above), rather than the typical 1n419, 1n4148, 1n547 small signal diodes, if very-low leakage will not be considered as essential.  


BTW, in my tests, I was able to find lower-than-average leakage in the "A" versions of the 1n457, 458, 459 series, over the typical 1n914/4148 types.

Here is my test layout.
Of course, these diodes need to be shielded from light during the tests.


Various protection diodes...(I am referring here, to D5, D6 input protection diodes)


=FB=
 
Leakage is not an issue at D5 and D6, because there is no significant voltage drop  when opamp is in linear operation.

I suspect with JFET gate diodes the capacitance could be an issue for some JFETS (perhaps mitigated with capacitance across feedback network). 

JR 

 
Does using 2 diodes in series do anything ?
.....e.g. two D5's in series, and two D6's in series. ?

--->|--->|---
---|<---|<---

..... In my particular opamp circuit, I'm using a single-ended input, not differential. It's a power opamp, and has a bi-polar input stage.
It's wired non-inverted, gain of 15, and  a passive volume pot and DC blocking cap on the input.
http://www.cirrus.com/en/pubs/proDatasheet/PA12U_S.pdf
They are expensive to replace, so I want adequate protection, all around.
~~~~~~~~~~~~~~~~~~~~~~~~~~~
(another note on input protection)
http://www.analog.com/static/imported-files/tutorials/MT-036.pdf

=FB=
 
It's a power opamp, and has a bi-polar input stage.

So why do you bother about diode leakage..? The part you use has input currents well into the nA range.

They are expensive to replace, so I want adequate protection, all around.

Look at the max. ratings: diff. input voltage 37 V, cm input voltage equal to the supply. In other words: once powered up you can't kill this beast except by plugging the input to another power amp or stright into the power outlet... Without power applied it is more easily to cause damage because cm voltage is exceeded. Use a pair of 1N914 diodes to the supply rails after a 100 Ohm series thermistor (PT).

Samuel
 
.....Use a pair of 1N914 diodes to the supply rails after a 100 Ohm series thermistor (PT)
I was wondering if this suggestion by Samuel is the same theory and basic layout as the image I posted.


Without power applied it is more easily to cause damage because cm voltage is exceeded.
I killed one opamp device already during testing...this may have been the cause.

This amp will be in a user-unfriendly environment, and I want it protected.....and learn along the way.



...and it's PSU...


 
Put some resistance in series with the module's input pin. 22K ought to protect against almost anything found on a stage.
 
Back
Top