From: Joerg Hansmann (info_at_jhansmann.de)
Date: 2002-02-01 13:22:08
Hi Andreas,
----- Original Message -----
From: Andreas Robinson <sleeper75se_at_yahoo.se>
To: <buildcheapeeg_at_yahoogroups.com>
Sent: Friday, February 01, 2002 2:05 AM
Subject: [buildcheapeeg] Re: Brainmaster input stage
> --- In buildcheapeeg_at_yahoogroups.com, "Joerg Hansmann"
> <info_at_jhansmann.de> wrote:
>
> Hi Joerg,
>
> > The 10Mohm has not been in series with the 100nF C.
> > It was the R-part of a highpass filter. So it can
> > not be removed.
>
> Oops. I meant "in parallel" of course... when the
> capacitor shuts of at low frequencies, you basically
> have a 1Mohm resistor
10 Mohm ? At least this was the value I have written
in the schematic...
> in series with the input,
> causing a lot of noise, perhaps because it is a real
> resistor, not a capacitor with resistance.
I was not talking of Johnson noise, that is only
produced by real resistors but of the effects
of the noise current from the inputs of the opamp that
flows into the input coupling network and that produces
noise voltages there regardless of the network beeing
ohmic, inductive or capacitve.
> My thought
> was that with a bigger capacitor, you could "push"
> that noise down below 0.1Hz, where it is out of the
> way.
But the same time the C grows bigger, the cutoff frequency
of the input hp will get even lower, enabling even more
LF-noise to enter the input amp.
So the LF noise could be eliminated only in a second HP
after the amp.
> > With these _very_ big capacitors Z would be OK and
> > noise also, however for these values you would need
> > to use electroytic (or tantalum) type Cs and they
>
> What? Are there no 100uF ceramics? ;o)
Have not seen any until now ;-)
> [snip]
>
> > It would not work because of the above reasons.
>
> Ok, just a thought.
>
> By the way, why don't you use an integrator plugged
> into the REF-input?
I have seen this sort of hp in the datasheets and bm schematic
and do not understand the point.
IMO it is effectively a hp-filter. However I could also get
the same characteristic with a simple RC hp. So why waste
an additional op-amp ?
> That would allow for higher
> instrumentation amplifier gain
I do not understand why this (admittedly cool looking hp that
fakes dc coupling at the first glance ...) should have to
do anything with the gain.
> and unchanged offset
> handling capabilities.
IMO it does not enhance offset handling capabilities if the amplifier
is used with high gain setting (e.g. 100 and above).
The last stage of the AD620 or IN114 is a difference circuit with G=1.
The integerator plugged into the REF-input will only compensate the
offset of this stage, however can not help the fist stage if it is
saturated by electrode DC or offset.
So why use an additional opamp for nothing ?
Or have I missed something ?
...
> I surfed the net a bit... The magic diode parameters
> are KF and AF. I calculated KF based on the
> current-noise properties of the LT1167 and made a
> diode model with only those two parameters. It appears
> to work.
Cool !
> In short: Add a diode to the schematic, name it
> whatever you want, such as NOISESRC, then write a
> SPICE-directive that says
> .model NOISESRC D(KF=1.31e-17 (or whatever), AF=1)
>
> The calculation and spice simulation is attached.
Thanks.
> I had to play around a bit with the DC-current of I1
> (60000 amperes, how does that sound? :-)
It sounds like only a theoretical diode would
survive this ;-)
> to get the
> right scale (1V/A), but you can probably improve from
> there on. Feel free to compare with the LT1167
> datasheet:
> http://www.linear.com/pdf/1167f.pdf
> page 9.
>
> > @ Andreas:
> > In another posting you have shown your excellent
> > skills in solving noise equations. Perhaps you can
> > comment on this...
>
> Thanks for your faith in my abilities. :-)
No problem. ;-)
> It's just
> "plug in equation A into equation B and give it to the
> supercomputer (My HP-48 calculator)".
The portable math-lab ... or even better ?
I do not have anything comparable on my PC,
so for future equations I fully rely on you
and your supercomputer ;-)
Regards,
Joerg
This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:37 BST