ECG Lead I configuration

[aspired]

I would like to measure the ECG LEAD I configuration.
I have the following hardware ( recommended by NI engineers)
SCXI 1125 ( it only takes differential inputs since the signals are floating).
SCXI 1305 - BNC conector inputs.
DAQ PCI 6052E
Labview 7.1 ( using DAQmx)

Lead I is generally measured by putting RA and LA signals in positive and negative inputs (respectively) of the differentaial amplifier which is referenced to RL ( which is grounded).
Since they dont have any customised hardware for ECG, I was told that I could do aquire the signals with the 1125 and apply the addition/subtraction in software.
Basically, if I connect the RA to the BNC Signal of one cable and the RL to the BNC GND ( or -input) of the same cable,
I will get measurement RA referenced to RL ( half of Lead I).
Similary put the LA to BNC signal of another cable and the RL to the BNC GND and that will give measuremnt LA referenced to RL ( Other half of Lead I).
Connect these two cables to the channel 0 and channel 1 of the SCXI 1305 and 1125 . Apply gain and filter. and subtract the signals in software to get the LEAD I.

However I am not gettting any signal even closely resembling the ECG and I am really frustrated.
MY hardware filer is 10Khz and gain is automaticlaly adjusted by giving input signal as +-3.5mV.I have applied a software filter of 0.01 - 400hz.
I have never used NI hardware for ECG before and I am really confused especially since the RL is floating and not really grounded or driven.
Any suggestions as to why this is not working would be very very very very appreciated.
I have attached a rough hardware wiring diagram of my system.
If Lead I is not possible then is any of the other 6 limb leads possible
?
Thanks a ton!

[Bill B]

Hello,

Thank you for contacting National Instruments.

I found your application to be very exciting and your drawing was very helpful. It looks like you are doing everything correct. However, I have noticed one problem with your connections from person to the SCXI-1305. The signals that you are trying to measure are floating. Not only are they floating, but they are extremely low voltages. More than likely, these signals are immersed in noise that is amplified by a gain of 1000 giving you nothing but a noisy signal as a result.

What you will need to do to improve your signal quality is ground your negative inputs, which are both referenced to the right-leg to the chassis ground of your SCXI-1305 terminal block. This is not the easiest task since yo
u have purchased a nice SCXI-1305 terminal block which has BNC connectors.

You can follow the instructions in this Knowledgebase:
http://digital.ni.com/public.nsf/websearch/6783C1F75D83349486256DE5008108A2?OpenDocument

Or you can simply solder a wire from the negative input (RL) to one of the grounded screws attached to the chassis inside the SCXI-1305 terminal block.

Please give this a try, this should improve your signal to noise ratio (SNR) and your signal should appear. You may need to adjust your filter frequency in software, a 100Hz low-pass filter should work fine, but you can test multiple cutoff frequencies to get the best response.

Please let me know what happens. I am interested to help you solve this problem.

Here is a link to the SCXI-1125 and SCXI-1305 User Manuals:

http://digital.ni.com/manuals.nsf/websearch/85C7726245BE7391862567BE0056BAFD?OpenDocument&node=132100_US (page 2-2)

http://digital.ni.com/manuals.nsf/websearch/226874A5F8AB7097862569B30
073B186?OpenDocument&node=132100_US

Regards,
Bill B
Applications Engineer
National Instruments

[aspired]

Hello Bill,

First let me thank you for your reply. It was very useful.

However, since I am not familiar wiht the NI hardware, I have some ( stupid!!) questions.

You have suggested 2 techniques.

1) "Solder a wire from the negative input (RL) to one of the grounded screws attached to the chassis inside the SCXI-1305 terminal block"

Is the screw you refer to, the same as the thumbscrew on the 1305 ( exterior to it) which is used to affix the 1305 on the 1125 ?.

If no, and the screw you are reffering to is 'inside' the 1305, then please can you elaborate on the location. I had a look at the manual and the board but I cant see any screw.

Once I do this, should the swtich be on F or G ?

2) The NI knowledgebase was also very useful. They s
ay:

"attach a jumper wire from the ground side of the built-in bias resistor to one of the chassis scews on the interior of the SCXI-1305"

Again, I am a little confused which screw they are reffering to since I cannot see any screws in the interior of the 1305. Also I would appreciate if you could clarify to me which is the bias resistor from the manual. Is it R3 (ch0), R6 (Ch1), R9 (Ch2)...
And is it is R3 (ch0)then I need to connect the lower leg of R3 to the screw . Is my assumtion correct ?

Thank a ton for your guidance!
Shubha

[johnsold]

Please be careful! The reason for using isolation is to protect the patient from ground currents. If you connect one of the patient leads to ground, you may be putting the patient at risk of electrocution. As little as 10s of microamperes at power line frequencies can be fatal if applied directly to the heart. Currents in the milliampere range from one limb to another are dangerous.

What you probably need is true differential amplifier (with isolation). You may not be getting a good simulation of that with your set up. The specified bandwidth of the NI isolation amplifier is 10 kHz. You are sampling at 1 kHz and then software filtering to 400 Hz. Any noise components between 501 Hz and 10 kHz will be aliased by your sampling and become im
possible to eliminate. In addition the samples from each channel are not taken simultaneously. With a 1 kilosample per second per channel rate and two channels, the samples for different channels could be 500 microseconds apart. Thus when you subtract, the signals you are subtracting are not exactly at the same phase. This is not a problem for your low frequency signals, where the phase changes very slowly, but any noise component higher than the Nyquist frequency (half the sample frequency) will not be cancelled.

I do not see a straightforward solution using the hardware which you have. It would probably work if the noise levels were low, but it has some fundamental problems in the presence of noise or strong interference.

Lynn

[aspired]

thank you for your response. However I am not using this machine on humans. so isolation is not a problem. Hence I would really appreicate any information about grounding the 1125 negative terminal ( questions I have asked Bill).
I guess i'll have to deal with the noise situation by more rigorous filtering or increasing sampling rate.

[Bill B]

Hello Shubha,

The Knowledgebase refers to making a connection for the F and G switch to connect to ground. This is not neccessarily what you need to do. You simply need to connect the negative input to any piece of metal that connect to the terminal block. Any screw inside or outside the terminal block will most likely work. The terminal block itself attaches with screws to the chassis which is grounded. Making this connection inside the terminal block will look cleaner. You simply need to locate the negative input of each channel and then use a wire to connect to any screw that is mounted to the terminal block. You may want to try this first on a single channel to see if there is improvement in your signal to noise ratio.

Regards,
Bill
B