i would like to design very low frequency(0.0001Hz to 50KHZ) preamplifier.
Thank you for your interest of using ADI products. I need more information about your applications so that I can give you the part that is best for you. Is your input signal will be coming from the sensor?If yes, what sensor and what is your expected input voltage range? What is your desired output voltage? What is your preferred supply voltage? Also feel free to tell more about your applications.
Thank you very much for your reply. I cannot give you all information regarding my project because it is confidential. But I will tell you regarding preamplifier.
My signal come from sensor, signal is in terms of microvolt and very low frequency like 0.001Hz. and I want output in terms of volts. And supply voltage is 12V
For all practical purposes the signal is from DC to 50KHz. With signals in the microvolt range you must use a precision op-amp. The exact type of opamp depends on your source impedance. The more modern ADI precision opamps have very low bias drift as well as low 1/f noise.
If you are talking about "a few microvolts" you may have a problem, even with the best opamp ADI has to offer.
Since the signal is virtually DC you will have temperature bias drift, 1/f noise... . In fact the choice of opamp will not determine the performance of your circuit. If it were that easy, everybody could be an analog circuit designer.
We really need to know, is it 1uV, 10uV or 100uV? What is the source impedance? Can the amplifier be located close to the sensor? Are there any power supply restrains? What is the required dynamic range of the signal? (this is maximum sensor amplitude divided by sensor noise... there is always the thermal noise)
Thank you very much for your reply.
My sensor is like hundreds of turns coil which has 8H inductance, 200ohm resistance and capacitance in terms of nF and my output of this sensor is in terms of micro Volts like 1 or 10uV. and amplifier is very close to the sensor and my power supply is within a limit.
so please suggest me how can i design my preamplifier.
Obviously the AD797.
This opamp is around for quite a while and still the best there is for this application. ADI has designed a few similar opamps like the AD859, but the differences are subtle and inconsequential.
I have seen a similar requirement in the offshore seismic, a EM (electromagnetic) sensor amplifier. I do not remember the frequency as low as in your application.
Whatever the final performance of your design will be, the AD797 will not limit your performance, your circuit layout however will.
Absolutely no switching power supplies anywhere.
Any digital electronic... put it far away from the input, not even in the same housing if possible.
Total attention to wiring, especially signal return and power ground (this is where most of the mistakes are made)
Avoid temperature changes and temperature differences in your sensor head. This will cause thermocouple effects which will affect your low frequencies.
Amplifier as close as possible (the first 100 * gain) on a separate PCB hardwired to the sensor as close as possible.
Screening the amplifier as well as all other electronics in MU-Metal (high permeability magnetic screening)
but only if it does not affect the performance of the sensor.... I don't know... try it out if you have noise problems.
if the offset drift is a problem, you may have to trim this out. I would not put a potentiometer on the sensor amplifier PCB, but rather have LPF with a frequency of 1/1000Hz later in the circuit and feed back a tiny DC voltage back to the front end to correct the offset.
You need some protection for the opamp. Even waving a screw-driver in front of the sensor coil will make your circuit jump. Maybe some BAV99 diodes to the supply rails.
You also may want to keep high frequencies (out of band) away from your amplifier. Allow for a feedback capacitor (a few pF) on the first opamp.
Thank you very much for your guidance.
Can i use instrumentation amplifier at the beginning of my circuit than i amplify my signal.
The instrumentation amplifier does not do anything for you unless you have long leads to the sensor and you have a lot of common mode noise.
In your case the amplifier is right at the sensor and the sensor is a large coil which is not connected to anything else but the frontend amplifier.
Depending on the gain you require, you need several gain stages obviously. I suggest to keep the gain in the first amp low maybe 10 only. After that you can do almost anything you want. Prototype the circuit fast and often until you get it right. It is not the amplifier or PCB which cost you the money, it is the time you spend contemplating this or that. You may want to SPICE simulate the circuit, just to catch the most obvious mistakes, but do not put to much trust into the simulation.
If you have a gain of 120 dB you will have a serious problem removing the offset, because you requirements are almost DC.
Thank you very much for your suggestions.
I am done with my design but there is some problem in phase plot. My magnitude plot is very well till 0.1Hz but phase plot is not good.
so can you please give me some suggestion to correct my phase plot.
How can you have a phase problem? This is a virtual DC amplifier. You will have noise problems and you will have bias drift problems.... Whatever you think is a phase problem is something else. I am sorry but it is virtually impossible to troubleshoot a problem with the kind of sparse info you are willing to provide.
I suspect the amplifier is drifting all over the place (1uV DC amplifier... 120dB gain???) and you think whatever you see is signal... it is not.
Put the sensor as well as amplifier into an enclosure. Power it up... let it run for hours to allow the temperature to stabilize.
Redo the measurement
Thank you for your reply.
I use ad797 amplifier and i can see my signal at the output and noise too. how can i reduce drift effect at output. and what kind of other information you need.
Signal is DC to 50KHz. Source is a very large inductor (8H, 200R) expected signal amplitude is 1-10uV. Output signal is a few volt to an ADC (I guess).
The source impedance will determine your noise. In this case it is about 2nV/rtHz at room temperature. Replace the sensor with a 200R film resistor for testing. The opamp is as good as it gets. However you cannot amplify by 120dB and expect the thing to do 50KHz. Even this opamp cannot do this. You need at least 2-3 stages of amplification. The first one is the most critical. Because of the very high gain and the DC path, bias drift is a problem. The entire circuit (especially the front end) must be at the same temperature. (an electrical connection made from dissimilar metals which are at different temperatures will generate an electrical voltage).
To trim out the bias you connect a low-pass-filter (1/1000Hz) at the output and feedback this signal back to the input.
Again, replace the sensor with a 200R resistor. Do not rely on SPICE, you must build it for real on a copper-clad board.
I have a faint idea what this circuit is for (a geophysical sensor). Based on the exact application of this circuit a lot of problems will show up. You mentioned before, that the precise application is confidential...
I noticed, that on this web-site the analog designers are all "hung-up" on the type of opamp they use. In most cases it does not matter, but everything else does matter. A good analog designer can do this circuit with a 30 year old OP07.
Retrieving data ...