From: Joerg Hansmann (info_at_jhansmann.de)
Date: 2002-02-04 14:03:27
Dear Jim:
----- Original Message -----
From: Jim Meissner <jpmeissner_at_mindspring.com>
To: <buildcheapeeg_at_yahoogroups.com>
Sent: Sunday, February 03, 2002 1:53 AM
Subject: Re: [buildcheapeeg] Re: Brainmaster input stage
>Dear Joerg:
>
>I used a Mind Mirror for years and did not have this time constant problem that you
>are anticipating. They used either 10 or 100 uf capacitors. Usually you put the
>probes on the head first, then plug in the connector.
During my experiments I often had the situation, that after some time
the electrodes got loose and had to be repositioned. Or that movement pulled
at the cables and gave artifacts.
So normally I have the electrodes plugged in and then the artifacts (or DC) appear.
>There should not be any voltage
>transient. At most it would take 10 seconds or so to settle down to within a few
>microvolts.
OK.
Lets caluculate this:
timeconstant * 5 = R * C
Factor 5 as a rule of thumb gives adequate settling (depends on what accuracy is
required).
C = 100uF ( 100e-6 Farad )
timeconstant = 10 seconds
To find the associated resistor for this given values the equation has to be
solved for R:
R = ?
R = (timeconstant * 5) / C = 500kOhms
If this 500k is really a part of the input network hp the input
impedance of the Mind Mirror is rather low (resulting in different
problems already discussed)
>By putting some back to back low leakage diodes ( like the 2N3904 BC diode ) across
>the 10 meg resistor, the settling time will be much faster.
In the attachment you can see the simulation result:
A 100uVp-p 1 Hz sinewave on top of a 1mV DC-pulse.
The high pass consists of 100uF and 10meg.
As you can see, the trace settles after about 4000 seconds
(more than 1 hour !!! )
The two BC-diodes are not of much use because at 1 mV DC they will not
conduct much current.
So I conclude that this input hp is rather useless.
10 seconds settling time is OK but not 4000 seconds.
Getting low noise requires a big C.
Getting good CMRR with imbalaced electodes requires high input impedance.
Usabiltiy requires DC rejection within 10 seconds (high pass).
These demands mutally exclude each other.
IMO therefore an input high pass is no good idea.
However I really would like to see the schematic of the
Mind Mirror you have mentioned, to see how they have
solved the above problem.
>he capacitor coupling eliminates the electrode offset voltage that you were worried
>about and also blocks any DC if the op amp were to fail to a rail. This also
>addresses the concerns the Mortz expressed.
That is right.
Regards,
Joerg
This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:38 BST