MCIA(2)C2 TEST OSCILLATOR 1/16/04




Theory

The square wave oscillator is on a separate circuit board mounted in an additional card-edge connector in the MCIA(2)C2 Cabinet. For convenience, a blank amplifier board was used to build the one Op Amp oscillator. Existing pads for components and the IC allowed for fairly easy construction.



The square wave frequency is 15Hz ±20% and output amplitude is 39 uVolts P-P ±10%. This results in a 1st harmonic sine amplitude of 50uVolts P-P and a 3rd harmonic 45Hz sine amplitude of 17 uVolts P-P. These two components "straddle" the 30Hz 3db corner frequency of the 5th order anti-alias low pass filters of the MCIA channels. The filter responses at 15Hz are about 0db and at 45Hz are about -20db making the ratio of 1st to 3rd about 30 to 1 or 30db after filtering.



This approach allows both gain and filter roll-off to be checked. If both are OK, the MCIA(2)C2 output after filtering should be a fairly low distortion sine wave with amplitude of 500 Milli-volts P-P since the MCIA(2)C2 Gains are 10,000. After the 40Hz Corner Freq. Upgrade the sine wave will be more distorted.

Test

The Test Oscillator output is wired to the CH1/Both/CH0 switch so that it is applied to the OPPOSITE Channel when either 1 or 0 is selected, or neither Channel when Both is selected.



The TEST button on the MCIA(2)C2 front panel must be pushed in and HELD in to continuously apply the test signal to the Channel Opposite that of where the Channel Select Switch is set. e.g. Channel 0's INPUT will be fed the test signal when Channel 1 is selected by the Chan. Select Switch. Push/Release will cause Test Signal Present/Absent.



There should be NO electrode cables plugged into the input{CH0 or CH1} when the osc.
test is being run on THAT channel. Electrodes CAN BE CONNECTED TO THE OTHER CHANNEL to allow CH0/CH1 quick compare.



Note that the software associated with the MCIA Amp based EEG System has a Scaling Anomaly such that 500 milli-volts P-P [corresponding to a 50 uV P-P input to the Amp] applied to the A/D card in the Computer will result in a reported amplitude of about 60 uV P-P by the software. i.e. the overall system reports EEG amplitudes which are about 20% higher than those produced by the subject.



During NORMAL AMP operation, meaningful signal outputs cannot be obtained if electrode cables are plugged into the inputs BUT are NOT connected to a Subject nor to a recommended "Electrode Paste Strip Termination [PST]". A PST should result in an Almost Flat Line [less than 3 uV p-p of interpreted noise] display. A PST can be made by "buttering" one popsicle stick with about a 1.5" strip of paste using a second popsicle stick then JUST filling two actives and one reference electrode and then applying the 3 sensors to the buttered stick making sure to not have any of the metal sensors touching one another.



If all is OK, artifact pickup should be able to be introduced, and seen, by having the System Operator pinch either of the two active sensor insulated wires during the PST Test. To CAUSE even HIGHER artifact the Operator could place his opposite hand very close to a source of artifact [a lamp or wall socket] while continuing to pinch an active electrode wire OR remove and replace one active electrode from the paste strip.

[This message was edited by Kurt on January 17, 2004 at 04:12 PM.]

here is another note from John Picchiottino from Biofeedback systems, about some of the troubles he sees when testing amps:

"We did an MCIA(2)C2 Calib. IsoTest and repair for Dr. S*** **** yesterday and found that he had an intermittent broken wire in his ear-clip electrode lead. The connection would make and break and sometimes give him OK pickup of EEG and other times not. So you may wish to pass that along to users reporting similar trouble and have them do thorough leadwire continuity tests with an ohmmeter and have them wiggle some of the connections to look for intermittent opens.

"We don't supply the electrode leads and on a number of occassions customers have sent Amps only in to us, reporting problems, and all checks out OK with the Amps, but the problems may be in the input leads. Operationally, they should be able to do lead tests at their end, and own an inexpensive $15 Radio Shack Volt/Ohm/Milli Ammeter for those sorts of purposes.

John P."

[This message was edited by Kurt on January 17, 2004 at 04:14 PM.]

And don't forget, all NC amps should go in for yearly Isolation Test and Calibration. The cost is usually $75 plus shipping, and the place to send them is:

John Picchiottino
Biofeedback Systems Inc
2736 47th St.
Boulder, CO 80301
(303)444-1411

[This message was edited by Kurt on January 17, 2004 at 04:06 PM.]