Electronics Tutorial - Building a Low noise signal amplifier Part 3/3 - Testing it 

Let me know how it turns out! Cheers!

I saw this approach being used in other projects. The one that comes to mind is EEVblogs uCurrent project ( www.eevblog.com/projects/ucurrent/ ). This project also highlights another advantage - if you buy the PCB from some manufacturer, you can use the silk screen layer to mark the various inputs/output, add a logo or something or add any drawings or things necessary - this will make the device look far more professional than if you write things by hand.

I noticed this in approach in various products - even commercial ones. Its quite nice that you can chose the PCB color (the solder mask) to be the same as the rest of your box - so its not obvious that its a PCB. Also you can add text in the form of the silk screen layer - to describe connector and switch functionalities.

Aluminum cases are very inexpensive and thicker than metalized plastic. Maybe there is cast iron? In my experience these Al boxes do the trick.

Hello electrotechnica I googled it yes it's true what you're saying but aluminum is a bad shield for EMI on low frequency the amplifier in the video is designed to measure low frequency noise.

Hello! From what I could notice the LM833 does make a bit of noise, but it's canceled out by the other op-amps high PSRR (power supply rejection ratio). From what I remember, by swapping the LM833 with another AD8676 I did not notice a major difference; in general it will be best to have all op-amps really low noise, but the only one that is truly critical is the first amplifying stage.

The scope must also have some BW limiting in place and the number of samples and rate must match that limit. Having fewer samples will cause the noise to mix down and increase the displayed noise floor.

Of course not! I left a link to it in the description. One condition though, if you build it, let me know how it works, I'm really curious what your results are.

I think I needed an 100R resistor (ideally) and the only trimmer I had at my disposal was 1Kohm; The ends where interconnected intentionally to try to make it easier to adjust the value. With 1K connected the usual way the 100R point is somewhere around 10%; but with the ends interconnected, the same resistance can be obtained somewhere around 20%; If you have a 200R trimmer, than you can of course connect it the normal way

Hello. To be honest, I did not really notice a difference but that might be because I was working in an area that is quite EMI quiet. If there was more noise in the area than the difference might have been more noticeable. Regarding ground, connecting each circuit element to earth ground but also in between will of course create loops as you pointed out which is not a good thing. In my case the grounding was EG to oscilloscope; oscilloscope to Amplifier; Amplifier to enclosure.

The metal enclosure is critical and must be connected to the coaxial cable shield.

I can't say for sure what was my logic when I made the circuit; but the supply voltage seems to be manufacturer depended - OnSemiconductor and TexasInstruments LM833 require a +/-5V whereas ST has a minimum of 2.5+/- (see table 2 www.st.com/resource/en/datasheet/lm833.pdf ) I guess I was using the ST branded component so 9V was sufficient

@@FesZElectronics - Good job. I didn’t even check the ST datasheet. I just assumed all LM833’s would have the same Vcc range. Looks like I’ll be buying the ST version. Thanks for the great videos and helpful information!

> Since 78xx regulators are common devices, I wonder if something can be done to reduce their noise. Noise is measured over bandwidth, so reducing the bandwidth over which the output acts, reduces the measured noise. This means more filtering at the output reduces measured noise. >You used a battery for its low noise contribution. Could a linear regulator be used? As seen in the video linear regulators insert noise into the supply lines. This can couple directly into the measured signal (see next point). They can be used, but batteries are a convenient way to get very low noise, plus you dont need to be worried about EMI/RF coupling from outside sources. >The op-amp is supposed to be insensitive to supply noise. So how much would get through? If you look in the datasheet of the opamp, it usually has a quoted 'power supply rejection ratio' or PSRR. Worst case for this opamp is 106dB from +-5 to +-15V. this is the change of input offset voltage compared to PSU voltage. You can do the calculations yourself how much that would be for a noise signal of a certain bandwidth. It is important to note, that the PSRR is usually denoted for a limited frequency band, but it is actually frequency dependent!! If you look further in the datasheet, at fig 21, you can see its frequency dependency and note that it quickly drops off at higher frequencies. Thus the type of noise signal is important to be considered too.

I'm not sure what you are referring to... The symbols without a clear name are the RF connectors on the input and output - these have 4 ground pins; and I also have in top left corner a switch that also has 2 ground fixation pins.

Hello Irsa. The monte-carlo command will only work on a single parameter - it doesn't care what component you are using. To use the "tc" attribute, you need to use resistors, but the mc command doesn't know this. So to use both mc and tc, the syntax would be " {mc(10k,tol)} tc=0.0001 - this way the temperature coefficient is applied on the monte carlo final value. Let me know if this works out.

Okay sure.@@FesZElectronics

@@FesZElectronics yes i cjecked it and it works.

@@FesZElectronics I have another question. Can we introduce faults in components in LTSPICE, like we can do in MULTISIM. zone.ni.com/reference/en-XX/help/375482B-01/multisim/settingplacedcomponentsfaults/#wp245032 Please guide me. MULTISIM provides option for these kinds of faults for every places components. Is there a way to introduce faults in components in LTSPICE?

three-mile island?