do input protection diodes add noise?
Alan Fang
I have part of the amp built up, and was doing some noise measurements
today. The problem is that the noise seems to be higher than what is
expected from thermal noise of resistors and input noise of op-amp. The
"extra" noise seem to be pretty much flat across the entire frequency
spectrum.
One thing I suspect is the input protection diodes at the input of the
op-amp. The manufacturer suggested I put two diodes (in parallel and
opposite directions) between the + and - terminals of the op-amp. I'm
not sure what kind of noise a diode (1n4148) generates when it has no
bias voltage across it. Can anybody help out?
The other suggestion from the manufacturer was to use JFET's (something
like a 2N5457) in place of the diodes. Will that be quieter?
Any other suggestions? Thanks, people
ACF
bsch...@ionet.net
Does it make any difference if the lights are on or off? The PN diode
junction is sensitive to light.
Also, have you tried a 1:1 isolation transformer in combination with the
diodes to the input of the op-amp? If you have any length of wire prior
to the diodes you might be creating a de-tuned AM radio. The isolation
transformer should help reduce this effect.
John Woodgate
_not_spam_me.com> inimitably wrote:
>One thing I suspect is the input protection diodes at the input of the
>op-amp. The manufacturer suggested I put two diodes (in parallel and
>opposite directions) between the + and - terminals of the op-amp. I'm
>not sure what kind of noise a diode (1n4148) generates when it has no
>bias voltage across it. Can anybody help out?
Disconnect them. Does the noise go down? (I doubt it.)
--
Regards, John Woodgate, OOO - Own Opinions Only.
Phone +44 (0)1268 747839 Fax +44 (0)1268 777124.
Did you hear about the hungry genetic engineer who made a pig of himself?
PLEASE DO ****NOT**** MAIL COPIES OF NEWSGROUP POSTS TO ME!!!!
Alan Fang
The way it's set up now, it'd be a real pain to remove them at this
point. I guess what I was trying to get at was a more theoretical
answer.
ACF
Jonathan Bromley
>
> I was hoping somebody wouldn't mention the "obvious" way to find out.
> The way it's set up now, it'd be a real pain to remove them at this
> point. I guess what I was trying to get at was a more theoretical
> answer.
I don't know, and I honestly can't be bothered going to a solid-state
textbook to find out - you can do that just as well. But here's a
guess; at least it'll provoke someone into telling me I'm wrong :-)
At zero bias voltage, there is of course no current flow in the diode.
But at the junction there are in effect two opposing current flows
which cancel: both are equal to the reverse leakage current.
Maybe you will see the shot noise in these currents? But the
reverse leakage current of a 1N4148 is tiny.....
But I have to admit that I would prefer to do the experiment.
Jonathan Bromley
Ing. Franz Glaser
>
> I was hoping somebody wouldn't mention the "obvious" way to find out.
> The way it's set up now, it'd be a real pain to remove them at this
> point. I guess what I was trying to get at was a more theoretical
> answer.
As a first reaction I would say that it is not the clamping diodes
which generate the noise, especially as you mentioned that it is
a wideband noise (thinking of "white" noise). This should not
occur, when no current is flowing. From my shortwave experience
(many years ago) I cannot remember any NOISE from anything else than
from current flowing, and each theory I heared of spoke about
current noise.
Unfortunately today the word noise seems to be used also for all
the coupling and BCI interference. But this would not be white.
Regards,
--
Franz Glaser, Glasau 3, A-4191 Vorderweissenbach Austria +43-7219-7035-0
Muehlviertler Elektronik Glaser. Industrial control and instrumentation
http://members.eunet.at/meg-glaser/ http://members.xoom.com/f_glaser/
http://www.geocities.com/~franzglaser/ http://start.at/bedarf
Roy McCammon
>
> Hi, people, it's me, the High Voltage amplifier guy.
>
> I have part of the amp built up, and was doing some noise measurements
> today. The problem is that the noise seems to be higher than what is
> expected from thermal noise of resistors and input noise of op-amp. The
> "extra" noise seem to be pretty much flat across the entire frequency
> spectrum.
>
> op-amp. The manufacturer suggested I put two diodes (in parallel and
> opposite directions) between the + and - terminals of the op-amp. I'm
> not sure what kind of noise a diode (1n4148) generates when it has no
> bias voltage across it. Can anybody help out?
>
> like a 2N5457) in place of the diodes. Will that be quieter?
I don't know if the diodes have noise themselves,
other than thermal. But, it the amp has voltage
noise, this noise will appear across the diodes
and slightly forward bias them which will then
produce a small current noise.
Tony Williams
Jonathan Bromley <jsebr...@brookes.ac.uk> wrote:
> I don't know, and I honestly can't be bothered going to a solid-state
> textbook to find out - you can do that just as well. But here's a
> guess; at least it'll provoke someone into telling me I'm wrong :-)
> At zero bias voltage, there is of course no current flow in the diode.
> But at the junction there are in effect two opposing current flows
> which cancel: both are equal to the reverse leakage current.
> Maybe you will see the shot noise in these currents? But the
> reverse leakage current of a 1N4148 is tiny.....
1N4148, glass package, photocurrent?
A small, but possibly significant, 'noise'.
--
Tony Williams.
Frank Miles
I've seen glass-encapsulated low-leakage diodes (with some bias voltage)
be fairly noisy when exposed to incandescent light. In the dark,
though, ... ? I haven't been bothered there, but maybe my resolution
wasn't great enough.
-frank
--
Al, N2NKB
absolutely do not add noise.
Even 1/f very low frequency noise, down to a few Hz.
Been there, done that.
Al
Alan Fang wrote in message <37CF6301...@yahoo.do_not_spam_me.com>...
>Hi, people, it's me, the High Voltage amplifier guy.
>
>I have part of the amp built up, and was doing some noise measurements
>today. The problem is that the noise seems to be higher than what is
>expected from thermal noise of resistors and input noise of op-amp. The
>"extra" noise seem to be pretty much flat across the entire frequency
>spectrum.
>
>One thing I suspect is the input protection diodes at the input of the
>op-amp. The manufacturer suggested I put two diodes (in parallel and
>opposite directions) between the + and - terminals of the op-amp. I'm
>not sure what kind of noise a diode (1n4148) generates when it has no
>bias voltage across it. Can anybody help out?
>
>The other suggestion from the manufacturer was to use JFET's (something
>like a 2N5457) in place of the diodes. Will that be quieter?
>
Alan Fang
power supply rails?
ACF
Paul Taylor
The notes I have for the 1N4148 show Ir < 25nA @20V, Crb 4pF max.
The www.siliconix.com JPAD5 datasheet quotes "Ir 5pA @ 20V (typ 1pA)"
It's in a DO-92 package, so probably isn't sensitive to light. Crb 2pF max.
It isn't very rugged (35V, 10mA, 0.35W), so will benefit from an input
signal limiting
resistor if the added noise isn't a problem.
Paul
Alan Fang <acf...@yahoo.do_not_spam_me.com> wrote in message
news:37CF6301...@yahoo.do_not_spam_me.com...
Bill sloman
Alan Fang wrote:
>
> Hi, people, it's me, the High Voltage amplifier guy.
>
> I have part of the amp built up, and was doing some noise measurements
> today. The problem is that the noise seems to be higher than what is
> expected from thermal noise of resistors and input noise of op-amp. The
> "extra" noise seem to be pretty much flat across the entire frequency
> spectrum.
>
> One thing I suspect is the input protection diodes at the input of the
> op-amp. The manufacturer suggested I put two diodes (in parallel and
> opposite directions) between the + and - terminals of the op-amp. I'm
> not sure what kind of noise a diode (1n4148) generates when it has no
> bias voltage across it. Can anybody help out?
>
> The other suggestion from the manufacturer was to use JFET's (something
> like a 2N5457) in place of the diodes. Will that be quieter?
I hope you have included both the voltage noise and the current noise of
the input to your op amp.
The 1N4148 isn't a low leakage diode, and for voltage differences less
than 26mV will look like a resistor with a "resistance" equal to 26mV
divided by the leakage current. If it is 25nA, as has been said
elsewhere
in this thread, then the "resistance" is about 1M. The noise on the
current
flowing through this resistance will be much higher than the Johnson
noise from a similar resistor, because you are looking at individual
holes and electrons, and need to treat the current as discrete
charge carrriers generating shot noise.
As was pointed out elsewhere the Siliconix PAD5 offers 5pA of leakage
current, and it comes in an opaque T0-18 metal can. Their JPAD50 only
offers 50pA lakage current and comes in a TO-92 plastic package, but
at half the price.
Hope this helps
Bill Sloman, Nijmegen
Jorgen Lund-Nielsen
RF energy that flyes around in the air.
that adds noise, too.
Alan Fang wrote:
>
> Hi, people, it's me, the High Voltage amplifier guy.
>
> I have part of the amp built up, and was doing some noise measurements
> today. The problem is that the noise seems to be higher than what is
> expected from thermal noise of resistors and input noise of op-amp. The
> "extra" noise seem to be pretty much flat across the entire frequency
> spectrum.
>
> One thing I suspect is the input protection diodes at the input of the
> op-amp. The manufacturer suggested I put two diodes (in parallel and
> opposite directions) between the + and - terminals of the op-amp. I'm
> not sure what kind of noise a diode (1n4148) generates when it has no
> bias voltage across it. Can anybody help out?
>
> The other suggestion from the manufacturer was to use JFET's (something
> like a 2N5457) in place of the diodes. Will that be quieter?
>
mariusz randak
> > Alan Fang wrote:
...
> > The other suggestion from the manufacturer was to use JFET's (something
> > like a 2N5457) in place of the diodes. Will that be quieter?
Yes, the diode connected JFET is less noisy then pn junction.
Bill explained it a bit here.
The glass case of the 1N4148 is also a problem. When highlighted the
diode is generating photocurrent! Noisy and modulated by the light!
Other very noisy devices you should avoid building low noisy
amplifiers are Zeners.
...
> > As was pointed out elsewhere the Siliconix PAD5 offers 5pA of leakage
> > current, and it comes in an opaque T0-18 metal can. Their JPAD50 only
> > offers 50pA lakage current and comes in a TO-92 plastic package, but
> > at half the price.
> > Hope this helps
> > Bill Sloman, Nijmegen
The best low leakage diode is diode connected Siliconix 2N4117A JFET
- Ir far below 1pA.
Selecting the JFET you should check the noise figure!
Regards,
Marius Randak
Chuck Parsons
mariusz randak wrote:
> Bill sloman wrote:
>
> ? ? Alan Fang wrote:
> ...
> ? ? The other suggestion from the manufacturer was to use JFET's (something
> ? ? like a 2N5457) in place of the diodes. Will that be quieter?
>
> Yes, the diode connected JFET is less noisy then pn junction.
> Bill explained it a bit here.
> The glass case of the 1N4148 is also a problem. When highlighted the
> diode is generating photocurrent! Noisy and modulated by the light!
> Other very noisy devices you should avoid building low noisy
> amplifiers are Zeners.
>
> ...
> ? ? As was pointed out elsewhere the Siliconix PAD5 offers 5pA of leakage
> ? ? current, and it comes in an opaque T0-18 metal can. Their JPAD50 only
> ? ? offers 50pA lakage current and comes in a TO-92 plastic package, but
> ? ? at half the price.
> ? ? Hope this helps
> ? ? Bill Sloman, Nijmegen
>
> The best low leakage diode is diode connected Siliconix 2N4117A JFET
> - Ir far below 1pA.
> Selecting the JFET you should check the noise figure!
If you are interested in very sensitive charge integrators or high speed
operation
the capacitance of the diodes effectively adds noise. That is the input
voltage noise of the op-amp ( or perhaps the signal source) divided by the
input
capacitance causes current/charge noise.
This is easily seen on PIN diode preamps. Particularly with no bias the
capacitance of the protection diodes can be quite surprisingly large.
mariusz randak
>
>
> If you are interested in very sensitive charge integrators or high speed
> operation
> the capacitance of the diodes effectively adds noise. That is the input
> voltage noise of the op-amp ( or perhaps the signal source) divided by the
> input
> capacitance causes current/charge noise.
>
> This is easily seen on PIN diode preamps. Particularly with no bias the
> capacitance of the protection diodes can be quite surprisingly large.
Yes it adds noise, but you must use input-overload protection designing
sensitive charge integrator. The only alternative I know is a spark gap
also having comparable capacitance and firing voltage above 60V.
The JFET I mentioned has about 3pF of capacitance, it is not a problem
in most applications.
The surge protection diodes (TranSil, TranZorb) have very high
capacitance,
it is effect of very large area of PN junction suitable to withstand
a current peak in range of kAmps.
marius randak
John Woodgate
inimitably wrote:
>The best low leakage diode is diode connected Siliconix 2N4117A JFET
>- Ir far below 1pA.
>Selecting the JFET you should check the noise figure!
I have some JFETs that leak a lot less than that, and are not at all
noisy. But they are all ones I've used in experiments, and they don't
seem to be JFETs any more. The magic smoke has leaked out. (;-)
--
Regards, John Woodgate, OOO - Own Opinions Only.
Phone +44 (0)1268 747839 Fax +44 (0)1268 777124.
http://www/jmwa.demon.co.uk Did you hear about