Go Down

Topic: Too much noise when reading Amplified Analog Signals (Read 1 time) previous topic - next topic

GingerGiant

Dec 12, 2012, 03:09 am Last Edit: Dec 12, 2012, 03:28 am by GingerGiant Reason: 1
I've run into a bit of trouble with my latest circuit. The signal it generates is flippin' awesome, but I cannot read it into the Arduino without ridiculous amounts of noise. I think it is related to the differences in ground voltage between the arduino and my circuit. running a wire from the output directly to the pin I want just doesn't work.
Suggestions?

After some more testing, I've noticed that the data I'm reading from the arduino matches whatever happens to be displayed on my O-Scope. If I connect the scope's ground to different power pins on the arduino, the signal changes. I would love to be able to use the unmodified signal from my op amps, but I don't know how to get the arduino to comply.

MarkT

IC1, the one on the right has no feedback network and there is no DC bias to pin 3 (non-inverting input).  Both of those are
show-stoppers.
[ I won't respond to messages, use the forum please ]

retrolefty


I've run into a bit of trouble with my latest circuit. The signal it generates is flippin' awesome, but I cannot read it into the Arduino without ridiculous amounts of noise. I think it is related to the differences in ground voltage between the arduino and my circuit. running a wire from the output directly to the pin I want just doesn't work.
Suggestions?


Well there won't be a difference if you run a wire from an arduino ground pin to your circuit's ground, as you should. But as you didn't show your wiring to the arduino I won't assume you did. 741s are very old (for decades now) that are not known for their quite behaviour and should never be used these days of much better opamps at same or cheaper prices.

Lefty

GingerGiant


IC1, the one on the right has no feedback network and there is no DC bias to pin 3 (non-inverting input).  Both of those are
show-stoppers.


That's actually part of the design. I was looking up op amps on wikipedia when I stumbled across the zero level detector. When everything is cooperating, this circuit can generate periodic pulses with the same frequenncy as the input signal's  fundamental frequency. I don't actually know how it works, only that that's what it is doing. I'm building a guitar tuner.



Well there won't be a difference if you run a wire from an arduino ground pin to your circuit's ground, as you should. But as you didn't show your wiring to the arduino I won't assume you did. 741s are very old (for decades now) that are not known for their quite behaviour and should never be used these days of much better opamps at same or cheaper prices.


The arduino is not connected. I'm not quite sure how or where I should plug it in. Originally, I was trying to build a single-supply op amp powered by the arduino, but that wasn't going anywhere fast.
I'm using the 741 because that's what I happen to have lying around. I have a 071's, if you think those would work better.

retrolefty



IC1, the one on the right has no feedback network and there is no DC bias to pin 3 (non-inverting input).  Both of those are
show-stoppers.


That's actually part of the design. I was looking up op amps on wikipedia when I stumbled across the zero level detector. When everything is cooperating, this circuit can generate periodic pulses with the same frequenncy as the input signal's  fundamental frequency. I don't actually know how it works, only that that's what it is doing. I'm building a guitar tuner.

The circuit looks reasonable to me, first stage a simple gain stage that is capacity coupled to a second stage operating as a zero crossing comparator stage which will generate square wave output only.



Well there won't be a difference if you run a wire from an arduino ground pin to your circuit's ground, as you should. But as you didn't show your wiring to the arduino I won't assume you did. 741s are very old (for decades now) that are not known for their quite behaviour and should never be used these days of much better opamps at same or cheaper prices.


The arduino is not connected. I'm not quite sure how or where I should plug it in. Originally, I was trying to build a single-supply op amp powered by the arduino, but that wasn't going anywhere fast.
I'm using the 741 because that's what I happen to have lying around. I have a 071's, if you think those would work better.

Yes, 071s would be much better. There is a big problem wiring this circuit to an arduino. The second stage is operating as a simple comparator so it's output voltage will only ever be one of two values, either +15vdc or -15vdc (actually somewhat less as a 741 is not a rail to rail op=amp) and both those value are too high positive and negative, and will damage an AVR analog input pin, which can only handle voltage ranges from 0 to +5vdc.
Lefty 


GingerGiant

#5
Dec 12, 2012, 04:10 am Last Edit: Dec 12, 2012, 04:12 am by GingerGiant Reason: 1


Yes, 071s would be much better. There is a big problem wiring this circuit to an arduino. The second stage is operating as a simple comparator so it's output voltage will only ever be one of two values, either +15vdc or -15vdc (actually somewhat less as a 741 is not a rail to rail op=amp) and both those value are too high positive and negative, and will damage an AVR analog input pin, which can only handle voltage ranges from 0 to +5vdc.
Lefty 



That gives me an idea. What if I wired up the second op amp using the 5V and GND from the arduino for its +Vcc and -Vcc?

I've been trying to test the circuit, but my power supply does not appreciate something I'm doing. After several seconds of activity, the output signals dies. I can get it working again if I turn the power supply on and offf. Recommendations?

retrolefty




Yes, 071s would be much better. There is a big problem wiring this circuit to an arduino. The second stage is operating as a simple comparator so it's output voltage will only ever be one of two values, either +15vdc or -15vdc (actually somewhat less as a 741 is not a rail to rail op=amp) and both those value are too high positive and negative, and will damage an AVR analog input pin, which can only handle voltage ranges from 0 to +5vdc.
Lefty  



That gives me an idea. What if I wired up the second op amp using the 5V and GND from the arduino for its +Vcc and -Vcc?

That would solve the over voltage problem for connecting to an arduino.

I've been trying to test the circuit, but my power supply does not appreciate something I'm doing. After several seconds of activity, the output signals dies. I can get it working again if I turn the power supply on and offf. Recommendations?

Sure, troubleshoot your circuit, compnents, power supply, or wiring until it doesn't do that.
Lefty


liudr

Before trying different amplifiers, try and reduce the resistors for IC2 by a factor of 10. I've used 741 before and found that at even 1,000 Hz frequency the large resistors make the amplified signal noisy. You have a scope so you can compare how your signal from the first stage looks like with two sets of resistors. If you do 47K and 1K then you will get the same amplification but hopefully less noise.

MarkT



IC1, the one on the right has no feedback network and there is no DC bias to pin 3 (non-inverting input).  Both of those are
show-stoppers.


That's actually part of the design. I was looking up op amps on wikipedia when I stumbled across the zero level detector.


Where's the link to this zero-level detector?  It looks suspect to me - certainly the lack of DC bias to the input is a no-no.

The 741 isn't used very much these days, it walks with a zimmer-frame compared to modern op-amps!
[ I won't respond to messages, use the forum please ]

retrolefty

#9
Dec 12, 2012, 06:27 pm Last Edit: Dec 12, 2012, 06:30 pm by retrolefty Reason: 1



IC1, the one on the right has no feedback network and there is no DC bias to pin 3 (non-inverting input).  Both of those are
show-stoppers.


That's actually part of the design. I was looking up op amps on wikipedia when I stumbled across the zero level detector.


Where's the link to this zero-level detector?  It looks suspect to me - certainly the lack of DC bias to the input is a no-no.

No, it's fine as grounding the pin two (it's - input) sets it's operating point. Any positive voltage to the + input pin (even just a few millivolts) will cause the open loop amp to saturate to full positive output and any negative voltage (again just millivolts needed) will switch to full negative voltage output. An op-amp working as a simple comparator is operating as a logic switching device not as a linear amplifier.
http://4.bp.blogspot.com/-LTsyaMaEpz4/UB4u0NZfhUI/AAAAAAAABWo/Jn7U8xDBkJ0/s320/Circuit%2Bdiagram%2Bzero%2Bcrossing%2Bdetector-744414.jpg





The 741 isn't used very much these days, it walks with a zimmer-frame compared to modern op-amps!

Agreed, but the OP said he also has available 071 op-amps and while still pretty old they are heads and shoulders better then a 741.
Lefty


dc42

#10
Dec 12, 2012, 08:27 pm Last Edit: Dec 12, 2012, 08:30 pm by dc42 Reason: 1
At the very least, you need to add a resistor between pin 3 of IC1 and ground to supply its input bias current. Try 1K. The blogspot article doesn't allow for Vin being capacitively coupled.
Formal verification of safety-critical software, software development, and electronic design and prototyping. See http://www.eschertech.com. Please do not ask for unpaid help via PM, use the forum.

oric_dan


Before trying different amplifiers, try and reduce the resistors for IC2 by a factor of 10. I've used 741 before and found that at even 1,000 Hz frequency the large resistors make the amplified signal noisy. You have a scope so you can compare how your signal from the first stage looks like with two sets of resistors. If you do 47K and 1K then you will get the same amplification but hopefully less noise.

Yeah, this is the first thing I noticed too.

Also, the gain of the 1st stage = 48, which seems pretty high, especially when OP has given no
indication of what sort of signals are coming from the guitar ckt. If the gain is too high, the
output of the 1st stage will be saturating, and even noise will be triggering the 2nd stage.

Also, the 1st stage is DC-coupled, so any DC-offset out of the guitar ckt will be multipled by 48.
Probably needs to be AC-coupled.

This sort of thing needs a systems-level design, meaning amplifier design needs to be selected
in conjunction with actual [measured] signal levels.

Grumpy_Mike

Quote
When everything is cooperating, this circuit can generate periodic pulses with the same frequenncy as the input signal's  fundamental frequency.

No it will not do that. It will pick out the strongest harmonic in the waveform not the fundamental.
The thing is that a guitar has some of the higher harmonics louder than the fundamental for part of its waveform.

Go Up