I would like to know about what is the process to design a guard ring.
I’m using a Op Amp to amplify an analog signal but I’m facing interference issues probably current leakage from the input pins to the output pin which is giving me an unexpected output.
I was advised to use a guard ring to eliminate the noise.
I read online and found this image:
However, I still have some doubts. Do we trace connections on the schematic to make the ring and then make traces when doing the layout?
What type of signals are you amplifying?
The reason I ask is - guard rings are only necessary for extremely high input impedance circuits. Normal amplifiers, even very high gain ones, don’t require it.
There are no special functions in KiCad (yet) for designing a guard ring, and as a result you will have to improvise a bit.
First, for the schematic, there is nothing special at all. You can draw some indication that some nets have to have a guard ring, but that is just visual. The guardring itself is always an already existing signal, although in some cases, it’s may be necessary to create a net for your guard ring, such as using an opamp as a buffer. There are also no real standards for what a guardring looks like in a schematic, although often it is drawn with a dotted line.
For the PCB, Guard rings surround some sensitive part, and usually surround it in close proximity. To do this, you have to disable the Push 'n Shove mode of the Interactive router. Walkaround mode may be of use here. Another common property of a guard ring is that the track is exposed. You do not want the guard ring itself to be covered with solder mask. There is no good way to do this in KiCad. Copying tracks from a copper layer to the solder mask layer (and increasing the width) is not easy in KiCad. Maybe KiCad V7 is getting some function for this I am not aware of yet though. It’s probably the simplest to lock the copper tracks of the guard ring, so they don’t move anymore, and then draw over them on the solder mask layer with graphical lines. A somewhat coarse grid will make this easier.
However, before you start drawing guard rings, optimize your PCB layout first. Make the area you want to guard with your guard ring as small as possible. This often means putting resistors physically close to the input of opamps.
But I’m not even convinced you need a guard ring. You should start with passivizing the 2nd opamp of the TL082. Opamp inputs should never be left open. This is a quite common cause of weird stuff, such as oscillations. R7 seems to be a 10M resistor, but it’s also connected to a 100k resistor. (I do not have much interest in completely reverse engineering your schematic from a bad screenshot).
In addition, you can thoroughly clean the PCB. If this does not change anything, then it’s likely the behavior you see has some other origin and a guard ring probably won’t help much either.
There are opamps that have better pin out for guard rings, with internally unconnected pins around the non-inverting input. Alternatively the 5 pin SMD package makes it easier than trying to squeeze the guard ring around pin 2 of your design
There is plenty of room to route signals in between pins of a DIP package. If you reduce the pad size, then even routing 2 pins in between the pads is doable.
Alternatively, you can also “air wire” the sensitive part. If the sensitive net does not touch the PCB, then there is also no need to air wire it. But possible leakage over the parts themselves remains of course.
I’ll try making a board without the guard and see how it works.
I thought about the guard ring because the output pin is very close to input pins which could cause current leakage.
Giovanne: This is the third topic you have started about this same circuit. All of the info you needed was in replies in your first topic. It is very frustrating for anyone trying to help you, since you never seem to learn what people are telling you – you keep changing your circuit and layout willy-nilly throughout your posts, and you never seem to incorporate the fixes people have been pointing out.
For the sixth time or so: don’t leave the spare opamp open
(short neg-in to out, and tie pos-in to gnd)
For the fourth time or so: you still think ground is battery neg when it is the center of the 1k resistive divider. This is obvious if you redraw the circuit as we told you before.
You added C1 and C3 on the opamp inputs, perhaps intending to bypass power pins to ground (though you still don’t know where ground is), but you killed all signal.
Ground plane is your friend. Of course, you need to know where to tie it.
But the most interesting issue is C2. Wow. You really need to learn some basic electronics.
There have been many, many posts over several threads by many very patient and well intentioned members regarding the design of this circuit, none of which having anything to do with the actual use of Kicad.
@giovannelucas, it is time you took this circuit to a more suitable forum elsewhere.
