From: yaniv_vi_at_yahoo.com
Date: 2001-01-04 02:50:52
--- In buildcheapeeg_at_egroups.com, "Joerg Hansmann" <jhansmann_at_g...>
wrote:
o.k. i get your point about the interface - you did alot of good work
for this project and i really appreciate youre willingness to work
and share the design , so sorry about the irda stuff .
so now you're current problem is 10UA current .
i remember in design of the brainmaster the input stage is the
instrumentation amplifier . and it's suppose to be safe enough .
(to collura who designed it does it for his living)
when i look at your first design - i guess the only 2 things
that can make a leakage current are :
1.the diodes (esd protection )
2.the amplifier
my guess is the diodes will make the noise .
and i don't remeber i came across diodes in any eeg design
that i saw .
maybe the design will be good also with out them ?
i'll start checking this on the net .
can you please tell what's the leakage current of input amp ?
by the way - how's you re desig change for 10UA ?is it working ?
what the preformance changes ?
and to another subject :
are you sure the in mass production we only do sample tresting for
iec ?
because on every chip i know there's a disclaimer agains medical
warranty - so i don't know if what enough for plain electronic device
(sample testing) is enough for medical device .
and are first quantities will be small - 100 max i think .
and another thing - one of the ways to market and sell is by third
part kit assembly - this is done by hand assembly , doesn't it matter
for sample testing ?
i'll try to get some info about the question here , and also please
send me your opinion .
and again sorry for the irda issue .
sincrerly yaniv v.
> Hi,
>
> ----- Original Message -----
> From: <yaniv_vi_at_yahoo.com>
> To: <buildcheapeeg_at_egroups.com>
> Sent: Sunday, December 31, 2000 7:09 AM
> Subject: [buildcheapeeg] question about safety testing
>
>
> > --- In buildcheapeeg_at_egroups.com, "Joerg Hansmann"
<jhansmann_at_g...>
> > wrote:
> > hi jeorg
> > if i understand correctly you have to test every unit that comes
out
> > of the factory with tests according to the iec601 ?
>
> If IEC601-1 is applicable , I would say: Yes. In series production
> probably random samples will be enough after an initial phase.
>
> > one thing about irda versus cable -
> > we can use a very simple not irda - infrared comm , that we'll
> > specially design with full duplex
>
> For full duplex the forward and the backward channel have
> to be designed in a way that they will not interfere with each
other.
>
> This could be done with multiplexing (e.g. time-multiplex,
> frequency-multiplex etc.) or physically isolated channels
> (fibre optics, etc.)
>
> IRDA transceivers simply suffer from scattered light that goes
> from the IR-emitter to the IR-receiver in the same device.
> IRDA transceivers block their input while transmitting in order
> to not receive their own data and to keep the AGC of the
> receiver adjusted to the remote device.
> Scattered light could be avoided with tubes and lenses and
> precise alignment of the two devices.
> However the cost-benefit comparison seems unfavourable to me.
>
>
> > , 115000 baud , and because we have
> > to build reciever card - we can build it to connect to printer
port
> > throught long 5meter cable . this could be very simple little
card ,
> > that can be put near the device , and he'll get he's power from
> > the printer port .
>
> Why printer port ? What advantages do you see over com-port ?
>
> The com-port seems better to me, because
> it has windows driver support and is supported by
> existing software like Rob's ElectricGuru.
>
> > i think this interface design could be very simple one .
>
> > it's seems o.k. with you ?
>
> Hm, yes... as long as I am not the one, who has to build it ;-)
>
> My current approach to solve the isolation problem
> is two optocouplers _inside_ the RS232EEG that are
> built of discrete IR-Emitters and Phototransistors and
> have an isolation distance of 2cm (that will give protection
> of about 20kV ).
> For such short distances and fixed alignment there should
> not be greater problems and the costs and development
> time will be sufficiently low.
>
> Regards,
>
> Joerg
This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:29 BST