Re: [buildcheapeeg]Patient Safety and Input protection

From: Joerg Hansmann (info_at_jhansmann.de)
Date: 2002-01-10 22:38:18


Dear Jim:

----- Original Message -----
From: Jim Meissner <jpmeissner_at_mindspring.com>
To: <buildcheapeeg_at_yahoogroups.com>
Sent: Tuesday, January 08, 2002 3:56 AM
Subject: Re: [buildcheapeeg]Patient Safety and Input protection
...
>> What do you think of the TMA0505S DCDC converter ? It has
>> low coupling C and high isolation R)
>I don't know anything about this DC/DC converter. Would you give me a
>reference to your schematic

http://www.jhansmann.de/eeg/modEEGdigital01.pdf

> so I can look where you are using it. I would not trust it for patient
> safety isolation no matter what the spec said. Opto isolation is safer. ( in
> my opinion )
> my website if you are interested in "my" solution. = "Iso01.gif" )

I have downloaded it.

>> Can it handle speeds of 56 kbaud ?
> All I needed was 9600, never tried higher.

> you can use Brian's circuit using QSE158-ND ($1.42, qty
> 1)and QEE113-ND ($.74, qty 1) with 1 or 2 inch spacing between the
> LED and photodetector on separate PC boards. ( I really like
> this. )
>> Can it handle 56 kbaud too ?
> We need to ask Brian. The T/on T/off is 100 nsec. for the receiver. What
> do you think?

If the transmitter is as fast as the receiver it can handle even 112000baud
easily.

> A floating ground requires some expert consideration of impedances at various
> frequencies.
>
>> The ground (exactly Vref/2) of the modularEEG is not "floating".
>> It is connected to the low-impedance output of an op-amp and
>> heavily bypassed to +5V and 0V.
>
>The low-impedance of the op amp is "not" very low!

Lets assume an open loop output impedance of the Vref/2 buffering OP
of 50 Ohms and an open loop gain of 100000, then for low frequencies
the output impedance will be 50 Ohms / 100000 = 0.5 mOhms.

For higher frequencies the bypassing Cs hit in.

>Any "real" capacitor is
>only a low impedance over a relatively small frequency range. Electrolytic is
>good at low frequencies, then you need a solid tantalum,

For low frequencies the closed loop output impedance should be good enough.

>then you need a
>monolithic ceramic, etc. I would like to see your simulation

For what circuit exactly ?
(BTW. I have done a simulation for the current TinyEEG analog
power filtering - (in another posting) )

> using "real"
> capacitor impedances over frequency with Xc series R series L. I have had to
> fight this grounding battle in the past. It will never be as good a ground
> as the chassis or PC board ground plane.

For HF ceramic Cs at each IC should be sufficient.
For low-frequencies I would rely more on the low closed loop
output impedance of a regulator or buffer.

No one prevents you from defining the buffered Vref/2 as the ground plane.
In fact I have done that in the modularEEGdigital pcb by copper pouring.
With a 2-layer pcb I would use a real separate Vref/2 plane below the analog circuitry.

>Something always oscillates. This
>would be a good application of a DC/DC converter to generate a +/- buss for
>the op amps. A simple clean solution. You may even need a separate power
>supply or batteries to decouple the input stage.

Better not. I do not like batteries. They are always empty when I need them,
have memory effect (NiCad), self discharge, need much volume, have great weight (compared to a little DCDC converter).

Regards,

Joerg



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