From: Joerg Hansmann (info_at_jhansmann.de)
Date: 2001-12-29 17:24:44
Hi Andreas,
----- Original Message -----
From: Andreas Robinson <sleeper75se_at_yahoo.se>
To: <buildcheapeeg_at_yahoogroups.com>
Sent: Thursday, December 27, 2001 9:51 PM
Subject: [buildcheapeeg] Input protection
> Hello folks,
>
> here's my first and hopefully last design of the input
> protection stage.
I am not so sure ... ;-)
> If you have the know-how,
Don't know if I have it ... ;-)
> *please* take a look at it
> and tell me what you think.
>
> A short description is in order. I've attached a .gif
> of the circuit, so go take a look at it.
>
> The protection is based on Jim Meissners input
> protection, the transistors are really used as diodes.
If I understand Jim Meissners circuit right, it is a
pure ESD input protection, using antiparallel BC-Diodes of
two 2N3904 Si-NPN transistors.
> I'm exploiting the fact that below 0.7V they don't
> conduct much current. Around 0.1V, where the input
> stage will saturate anyway, the leakage is around 6pA.
> I hope this is an accurate figure, and that it is low
> enough.
>
> Now, if anyone applies a couple of thousand volts to
> the left side of the circuit, as when petting the cat
> with a glass rod and then touching the electrodes with
> it, the current generated is shunted through to the
> opamp output, where perhaps a pair of diodes can pass
> it through to the power rails. I'm not clear on what
> to do there yet.
Perhaps 2 shottkey diodes to the
power rails and a resistor in series to the output of the opamp ?
> If two or more amplifiers fail so that one amplifier
> input is connected to + and the other to -, the
> transistors will again conduct the current, this time
> from one power rail to the other, protecting the user.
Because you use the outputs of the input OP-amps too, you
have certainly taken into account, that these outputs can
also fail, by beeing internally shorted to the power rails,
and have iterated all possible combinations ?
> The resistances to the left are calculated so that the
> human is safe and the ones one the right so that the
> transistors don't get fried.
>
> Some specs:
>
> For four channels, if three leads gets shorted to +6V
> and two to -6V,
> which is the worst case, the current
> through the user is 90uA with an electrode impedance
> of 1K. The transistors get 9mA each.
If the channel 4 op-amp outputs gets internally shorted to -6V
all other outputs to +6V and all inputs to +6V, Q4 gets
somewhat below 76mA into its base and ca. 25mA into its collector.
(0.7V drop per diode neglected)
> For FDA
> approval, just change the 16K resistors to 33K.
> Hopefully the noise levels are be ok with LT1012
> amplifiers.
>
> For two channels, you can use half the resistance for
> the resistors on the left.
Another problem:
The protection circuit proposed by Jim Meissner depends on
one buffering (G=1) amplifier per channel, however the
www.biosemi.com/publications/artikel7.htm
circuit has only one common guard driver.
Do you want to give each channel a separate guard driver ?
Further you use one reference electrode. Do you think this could be
exchanged with the DRL output ?
BTW: Does the circuit really need +-6V or could it be operated with
-2/3V as the modularEEG ?
Regards,
Joerg
This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:36 BST