contactless electrode with lmc660

why no body didn't answered to me..?

Hello Siyavash,

The LMC6442 is a micropower amplifier, and is much slower than the LMC660 (10KHz vs 1.4MHz).

The LMC6442 is also not unity gain stable. It must be used in a gain of two or more.

To see if the LMC6442 can replace the LMC660 will depend on the circuit details. You would have to describe your circuit in more detail.

The LMC6482 is the rail-rail input version of the LMC660, and should easily drop in place of the LMC660 without any problems.

In regards to the "figure 7" noise reducing circuit (also known as "paralleled 'n' Amplifiers"). While the noise adds in a sum-of-the-squares fashion to reduce the overall effect, input capacitance and DC errors do not. The cost of the lower noise of this circuit is that the DC bias currents average and the input capacitance of all the device inputs add up linearly. Nothing is free...

Also, "low noise" devices will have higher input capacitance, which may limit your ultimate bandwidth. So you must balance the tradeoffs of using the "lowest noise" amplifier with the effects of higher input capacitance and/or bias currents.

Keep in mind that getting the lowest noise device may not be the ideal device. If your source impedance is high (tens of megohms to gigohms), the noise will be dominated by the noise of the equivalent source resistance. So getting a 5nV/rtHz vs a 7nV/rtHz amplifier makes little difference if there is already several hundred nV/rtHz coming from the source. That 35nV/rtHz amp may actually give the same, or better overall performance than the 7nV/rtHz amp if your source impedance is high (due to lower input capacitence).

Regards,

Hello Paul, thank you .

 In my active electrode that can acquisition EEG or ECG the source impedance is very high (giga ohm or higher) that is higher than Rop-amp(Rop-amp=input referred voltage noise/input referred current noise). So input current noise is dominate there for my first feature for select op-amp is Ib. because I have the lmc660 and it's feature is close lmc6001 lmc6484 lmc6442 in current noise I think I can use it with 2 stage first stage is a deriven shield and unity gain to reduce noise and also change high impedance.

And in stage 2 i amplify it with G=10 and its output and Vcm will send to a difference amplifier like that ina326 and a right leg drive use too.

But I don’t use internal op3 &4 of lmc660. Is it better to use them for reduce noise and the other harmful?

Best regards,

Siyavash

I want to design two layer PCB in according to the datasheet I should use ring around inputs pin but the Soic-16 I think pin distance is very low and is it possible to do it?

and what is your idea for deriven shield in exact this package. is it good that parallel op-amp or series there?

The other problem that is painful is most of application not about PCB is about more than 2 layer.

Hello Siyavash,

Paralleling the amplifiers reduces the VOLTAGE noise by 1/sqrt(n), but also increases the CURRENT noise by sqrt(n). So if you are truly worried about current noise, then you do not want to parallel the amplifiers.

If you parallel three LMC660's, the total VOLTAGE noise will drop from 90nV/Hz to 52nV/Hz (using 10Hz numbers - since EE/CG's are low frequency applications).

But even a 10Mohm source impedance is 411nV/Hz by itself.  One amplifier (90nV/Hz) would be 420.7nV total, and paralleling the three would get you 414.3nV/Hz - a 6.4nV/Hz difference. But is that worth the slightly higher current noise and 3x larger input capacitance? As the input impedance goes up, that noise difference becomes less significant.

The LMC660 has about ~3pF of input capacitance. To reduce the input noise by 6nV, you now have 9pF of input capacitance (more like 11pF with additional wiring strays). Input capacitance kills your BW, since it causes peaking, which requires compensation, which further reduces the BW.

Every application is different. All I can say is "try it". Design your circuit so that you can implement the "n" amp if you wish.

You can use the extra amps as guard buffers, or, you can parallel them and use them as the leg driver. Just place 200ohms to 1Kohms in series with each output for capacitive isolation and short-circuit current limiting.

Regards,

thank you,

It is my first low leakage design so excuse me for my primary question. Now I am routing the PCB and soon I try it. but can you describe Guard buffer with an example in TINA-TI? 0317.Siyavash.TSC

best regards,

Siyavash

Hi,

I made the PCB of my work, I tested it by a function generator with 1mV p-p input + 2.5volt dc. it work property with a 200 Gain. but when I tested it for acquisition ECG it doesn't work even it doesn't show heart rate. I describe my circuit again, in active electrodes ,I buffer and then send it to second inner op-amp with a gain 2 and then I send it to main board that I have an instrument amplifier INA 326. I think my major problem is use single supply op amp like lmc660, I tried to solve my problem with a right leg circuit that I give 2.5 volt dc + common reject to body, but my problem still unsolved, I am really need your help,

Thanks & Regards,

Siyavash.

Please answer to my question as soon as possible,