From: Joerg Hansmann (info_at_jhansmann.de)
Date: 2001-12-04 19:16:04
Hi,
----- Original Message -----
From: Doug Sutherland <wearable_at_earthlink.net>
To: <buildcheapeeg_at_yahoogroups.com>
Sent: Tuesday, December 04, 2001 3:27 AM
Subject: Re: error 403: Forbidden! Re: [buildcheapeeg] How to get the current design
> I took a look at all of the schematics on the OpenEEG site
> and made some notes so its easier to understand differences
> between all of these designs. Joerg: you have been busy!
Thanks :-)
> Thanks for sharing all of your work. Following is a summary
> of each design, and I have made a few comments below that.
>
> COMADC01 (schematic dated 07/28/2001)
> - 1 channel isolated 8+ bit ADC for RS232 PC interface
> - Uses no microcontroller, direct connect to PC
> - Uses 555 timers for clock generation and pulse width modulation
> - Uses no instrumentation amp, signals feed into PWM circuit
> - Sends pulses to PC over RS232 handshake lines (RTS and CTS)
> - Pulse range 1mS to 10 uS
> - PC code must run in DOS only mode (disabling intterupts)
DOS-mode is preferred, but WIN9X is possible, however
with ugly artifacts due to multitasking in the background.
Writing a Windows driver (similar to the gameport-driver)
could bee a solution. However I have doubts, that the expenditure
pays in the end.
...
[correct and detailed analysis of the various designs snipped]
...
>
> modEEGdigi01 (schematic dated 11/28/2001)
> - 2-channel isolated EEG amp with RS232 interface
> - Uses AT90S4433 microcontroller
> - Microcontroller does ADC and connects to PC
> - Sends data to PC over RS232 data lines (RX and TX)
> - Uses 6N139 Opto couplers
> - Uses differential uV amplifier
> - Uses INA114 instrumentation amps and TCL277 op-amps
> - Input stage modeled after Biosemi designs
Only partially. The URL on the schematic refers to the driven right
leg circuit.
The input stage with the odd 1st order HP is my invention:
It has the advantage that the common mode input impedance
of the INA114 is preserved.
This is essential for good CMRR with imbalanced electrode impedances
(see article about potential divider effect
http://www.biosemi.com/publications/artikel3.htm
"3.1 Influence of common mode voltage" )
The brainmaster input stage e.g. uses input HP filters, that
definitely decrease common mode input impedance for several magnitudes.
> - Amp on separate board, connects to ADC board w/uC
> - Not much documentation included yet
> - PC Cable connectiotion documented
> - Firmware source and rom not in zip file
The firmware is essentially the same as for RS232EEG01.
It should only be extended to generate the calibration
signal.
> - Should work with DigitalGuru software
>
>
> Here are some personal opinions/comments on these designs
>
> - ComADC and ComEEG could be problematic:
> - interrupts must be disabled for timing reasons
> - must run in DOS only for above reason
> - will not work with either windows or linux
> - you could boot a windows machine into DOS only
> mode and run the software, but cannot run any
> other software at the same time.
> - will not work with DigitalGuru software
>
> - Perhaps ComADC or ComEEG could be connected to a
> microcontroller to solve the above problems, but
> in my opinion it's better to use the new input
> stage as in ModularEEG.
>
> - The RS232EEG and MODEEG designs look much better.
> The use of uC for ADC and RX/TX data lines is a
> better approach. This allows easy integration
> with both traditional PCs and embedded PCs. I
> could easily take these designs and add on
> extra circuits for ECG, EMG, GSR, and temp if
> I wanted to, I like the flexibility. These also
> use a data format which (I think) will work
> with the DigitalGuru software.
>
> - There are no features in any of these designs
> for standalone operation, all require a PC.
> They could be modified slightly to add audio
> feedback and/or interfaces to small LCDs that
> can display text, bar graphs, and line graphs
> of the waves.
Yes. This could be interesting if the PC-Version works as
expected.
For excessive filtering / FFT / audio-feedback synthesis
however a real signal processor could be necessary...
> I have done a lot of work with
> this and I'm willing to add these features to
> the existing designs. All that is needed for
> small LCD is one output pin from the uC and
> one RS232 level converter (recommend DS275).
> Check out my jacket sleeve display, it would
> be REALLY cool to display EEG data here ...
> http://home.earthlink.net/~wearable/jacket-wearable.html
>
> My overall opinion is that the ComADC, ComEEG, and
> GamePortEEG designs are too low end and are too
> restricted and limited.
ACK.
> I think we need a design
> with microcontroller interface to do a good job on
> the PC software side of things. People will want
> to run Windows and/or Linux applications and will
> expect some good graphics. We cannot make a good
> GUI in DOS. People will not want to run machines
> in DOS only mode, they will want to be able to
> cut and paste EEG data into spreadsheets and such
> without having to reboot.
>
> I assume that the new ModularEEG replaces the old
> EEG_RS232 with improvements in the input stage. I
> assume then that the old EEG_RS232 is obsoleted.
ACK.
> I am interested in making a prototype of the new
> ModularEEG design. I have a few questions for
> Joerg on this new design:
>
> 1) Is there a working prototype of ModularEEG?
No. It is "work in progress".
However I am confident that the digital part of modEEG
should work, because it is more or less copyed from
the working RS232EEG.
The analog part is new and only tested partially in
a MICROSIM EV8.0 simulation. (MICROSIM EV8.0 is a
restricted freeware version of a pspice clone with
a nice GUI)
> 2) Has this PCB board design been tested?
No.
> 3) Do you consider this to be a final version?
At this moment: Yes.
> 4) Can I get the firmware sources from you?
It is on my EEG-page (http://www.jhansmann.de/eeg/eeg.html)
in the file:
"RS232EEG_release010226.zip"
in the sub-zip-file
"firmware001214.zip"
in the file
"EEG02.c"
(I have put this file in the attachment)
> 5) Can this be used with the DigitalGuru
> binary source code?
It can be used with Rob Sacks "ElectricGuru"
> 6) Do you have any specs on the frequency response
> band width, sampling rate, CMMR, noise, etc?
>
Yes. Here they are:
(better looking as RTF-file in the attachment)
Specifications of modularEEG
digital part:
isolated power supply:
? input voltage: 7..12V dc, or 5V dc stabilized (from PC or mains adaptor)
? isolated with DCDC-converter(TMA0505S):
Isolation voltage Input/Output 1'000 VDC
Isolation capacity Input/Output 60 pF typ.
Isolation resistance Input/Output >1'000 Mohm
isolated RS232 Data-lines (opto-isolated)
? 2x 6N139 (RxD, TxD)
Viso > 2500 Vrms (Input-Output Momentary Withstand Voltage)
Resistance (Input-Output) Rio=10^12 ohm
Capacitance (Input-Output) Cio=0.6 pF
Microcontroller:
? Atmel AVR AT90S4433 at 7.3728Mhz
with 6 channel (multiplexed) 10 bit resolution ADC, internal Sample and Hold Amplifier
analog part:
? EEG-amplifier:
2 differential mode (bipolar) channels
? 512uVp-p input range
? 0.5uV resolution
? lower -3dB corner frequency: 0.25Hz, 2nd order
? upper -3dB corner frequency: 75Hz, 5-pole Butterworth
? driven right leg circuit (DRL)
? shield drivers for shielded electrodes
? noise: about 1uVp-p over above specified bandwidth
? common-mode rejection: typ. 115dB
? Impedance, Differential 10^6 ohm || 1nF
? Impedance, Common-Mode 10^10ohm || 60pF
misc.:
? 256Hz samplefrequency, (optional 200 Hz)
? 4 unused ADC-channels (input range +-2V)
? 4 TTL-inputs (for switches e.g. for marking events by the user)
? 4 LED-driver outputs (e.g. for light goggles)
? 1 PWM output
? hardware UART (RS232 transmission
? parameters are: 1 startbit, 8 data bits, 1 stopbit, no parity, 57600 baud)
? on board calibration generator: 50uVpp 10Hz rectangle wave
? easy in system programmable uC-EEPROM/Flashrom
? uC-firmware written in AVR-GNU-C
And here snipped from an older mail:
Description of the various "blocks" in the design:
(Hopefully the references are still correct...)
1) C1,R1 forms an input side differential mode HP with fc=0.16Hz to
prevent IC1 from saturation by electrode offset DC-voltages.
2) C2,R2 forms an impedance symmetry compensation network.
3) An additional purpose of of R1,C1,C2,R2 is to protect the user from
worst case circuit failures of IC1 (power lines shorted internally to input lines)
4) IC2A is a shield driver that reduces the effective (shielded) electrode
cable capacity by the open loop gain of IC2A
5) IC2B is part of a servo circuit, that reduces common mode voltages at the
amplifier inputs (IC1) by amplifying (only) the common mode voltage, inverting it
and feeding it back into the users right leg. (right leg driver circuit). By this
The common mode suppression of IC1 is extended.
6) C35,R41 is a HP (fc=0.16Hz) that suppresses the amplified offset voltge of the INA114
(max 50uV RTI * 1000)
7)P2 adjusts the overall gain to 7812.5 so that the full scale input range of 512uVpp is scaled
to the 0..4V input range of the AT90S4433 ADC-part.
8)IC3A, IC3B forms a 5-pole Sallen Key Butterworth Low Pass with 75Hz cutoff frequency.
(5th pole is on the digital board)
9) IC8A buffers the uref/2 voltage
Regards,
Joerg
This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:32 BST