Re: Input protection

From: Andreas Robinson (sleeper75se_at_yahoo.se)
Date: 2002-01-06 13:39:46


Hi Joerg,

--- In buildcheapeeg_at_yahoogroups.com, "Joerg Hansmann"
<info_at_jhansmann.de> wrote:

> But the HF gets demodulated at non-linear components
in the OP and
> thereby adds a seemingly unpredictable offset
voltage that
> really is music from the next AM-transmitter.

Ah! Then we can listen to music while we're measuring
EEG! :-)
Seriously though, I see your point...

> I have simulated a BAT 54 (30V, 0.3A) with
switchercad
> and can`t believe how bad the leak current/impedance
is compared with
> a npn (e.g 2n3391)

It seems we are limited to transistors then, but maybe
we don't need anything else. (Read on...)

> > > > active emergency-shutoff...

Is a certified PITA.

I've done some experiments (aka simulations) with a
FET-based current detector. Building a differential
amplifier is simple enough, but the gate-current for
the types of NMOS transistors I tried, got rather high
(1nA).

In other words: It would certainly measure
overcurrents but destroy the EEG signal at the same
time.

I went to Biosemi (where else?) and read about their
protection method once more:

http://www.biosemi.com/faq/limit_current.htm

If I understand it correctly, they only check for
short-circuits during power-up. If that's good enough
for them, why should it not be good enough for us?
Single faults are handled by the DRL, and double
faults are "very unlikely".

Here's an outline of a crude method for the
fault-checking:

We connect the amplifier input to the electrode
through a relay. The input can be switched to either
the electrode or a 1Mohm resistor connected to ground.
A window comparator measures the voltage over the
resistor and signals an error if it detects more than
+/- 5mV. The 5mV => 5uA through a 1K resistor.

When the EEG is turned on, it does a self test -
checking all amplifiers for errors before connecting
the patient.

I don't know what kind of errors relays introduce
(thermocouple voltage offsets?), but there are types
designed for low-level signals such as audio, that
cost about €3-€6 each (we would need at least four).

So, what do you think?

Regards,

Andreas

_____________________________________________________
Do You Yahoo!?
se.yahoo.com



This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:36 BST