Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.
Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.
I have attached pdfs of schematic and board. Could you check and let me know if I am doing something that would effect the signal? I tried my best to reduce the trace lengths between input and output.
Hi there. Overall, that's a fairly good layout, I think. A lot depends on the performance you expect to achieve, and other physical constraints such as connector locations that I can't discern by looking at this one board in isolation from the rest of the system. Everything is a compromise. I offer the following comments to optimize performance of this one board, ignoring any other physical constraints.
1. It looks like the input traces and output traces are on the same plane as the ground plane. That is okay, but there will be some capacitive coupling to the ground plane from those traces, and those are the worst signals you want to have stray capacitance on. If they must be on the same side, increase the space between those traces and the ground plane to reduce the capacitance. Alternatively you can put them on the other side and remove the copper on the ground plane that runs under/over those traces.
2. I assume as a practical matter that the traces are a short as possible, but the lengths for inputs and outputs are too long for my comfort, adding inductance. A lot depends on the performance you really need, but that's the kind of stuff that can cause ringing. The length of those traces are most important.
3. Your input and output connectors are all right next to each other. Ideally, the input connectors are on one edge and the output connectors are on the opposite edge. That's why the in/out pins are on opposite side of the op amp package. Since the gain of your amplifier is low, you will most probably be okay, but why temp fate?
4. Since this is an all analog board, I would try to put the power connections/power filters in the middle-top of the board rather on the left side to better approximate a single point ground. Power current will tend to flow in a straight line from each amp to the power connection points. A fine point, admittedly, since the bypass capacitors near each amp will mitigate a minor ground resistance issue.
1. It looks like the input traces and output traces are on the same plane as the ground plane. That is okay, but there will be some capacitive coupling to the ground plane from those traces, and those are the worst signals you want to have stray capacitance on. If they must be on the same side, increase the space between those traces and the ground plane to reduce the capacitance. Alternatively you can put them on the other side and remove the copper on the ground plane that runs under/over those traces.
I did thought about it, but I tried to eliminate those input and output traces running close to power traces. Could I run the power traces on the bottom and reroute those input and output traces on the top layer?
2. I assume as a practical matter that the traces are a short as possible, but the lengths for inputs and outputs are too long for my comfort, adding inductance. A lot depends on the performance you really need, but that's the kind of stuff that can cause ringing. The length of those traces are most important.
I was unable to reduce those lengths any further as I have to use a 8-coax d-sub connector for input and output for 4 channels. Not sure on what else I can do on this part.
3. Your input and output connectors are all right next to each other. Ideally, the input connectors are on one edge and the output connectors are on the opposite edge. That's why the in/out pins are on opposite side of the op amp package. Since the gain of your amplifier is low, you will most probably be okay, but why temp fate?
Thats true, I have seen that setup on some off-the-shelf op-amp circuits, but here as I am using that coax d-sub, I ended up having inputs and outputs on the same connector.
4. Since this is an all analog board, I would try to put the power connections/power filters in the middle-top of the board rather on the left side to better approximate a single point ground. Power current will tend to flow in a straight line from each amp to the power connection points. A fine point, admittedly, since the bypass capacitors near each amp will mitigate a minor ground resistance issue.
For now, this would be an all analog board, in future I would like to consolidate my digital board onto this board and have a single board for all. But, I guess I can move the power supply connector to the top of the board towards the middle.
Thanks for all your comments and suggestions. I am really learning something new.
You're welcome. I understand the connector constraint. You can only do what you can do. So, that leaves only the in/out traces on the same side as the ground plane. Just keep a gap between those in/out traces and the ground plane, say 1/4 inch, or so. You could put them on the other side, too, but then you probably have to reroute everything.
I tested the circuit with LM6181 along with AD811 and AD811 looks like having low noise on the output compared to LM6181. One more thing I noticed is the output ringing noise is less in magnitude using dual 5V vs dual 15V. Does this mean using dual 5V is ideal in these kind of amplifier circuits? My input signal is 1V and output will be no more than gain of 2.
Could someone enlighten me on the advantages of connecting resistors to ground at non-inverting input of an inverting op-amp as shown in the schematic?
It´s not too important.
The resistors are for compensation of the small dc offset voltage, which is created by the dc bias current flowing through the feedback resistors.
These additional resistors generate about the same dc voltage at the pos. input - and, thus, can compensate this small unwanted dc voltage at the inv. input.
I think, in your case with a gain of approx. unity you do not need these resistors.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
