Methods to design op amp ciruits?

[jose_peeterson]

Dear all,

I am using "TINA ti" to design a shock sensor circuit. The sensor is represented by a "ac source with a capacitor in series and a resistor in parallel". The aim is to design the circuit with signal conditioning to allow only signals in the frequency range 300 to 20khz.

I think the amplitude of the "ac source" represents the "g forces of shock". so the device should be able to measure 5G to 500G without clipping the op-amps rails.

First, I did not use any method but just randomly insert cap and resistor values to meet the frequency criteria. the reference circuit is here https://www.datasheets.org.uk/PKGS-00NB-R_PKGS-00NB-R/Datasheet-077/DSAE0057787.html

Now the transient analysis is not as expected. for even 1mv i get clipping. I have to change back the cap and resistor values which in turn affect my frequency response.

Can somebody please help me how to build the 3 stages separately and put them together.


please view my circuit and the values to comment on this.


Thanks,

SS_3_STAGES_COMPLETE.JPG

 

[barry]

I didn't look that closely, but I question the fact that you've got inverting opamps with the negative supply connected to ground. This means that all positive inputs will be clipped to zero. That's ok, as long as your input goes negative

 

[jose_peeterson]

you mean, I connected the positive terminals of the inverting op amp to ground. Why would connecting positive terminals to ground cause + inputs to go to zero? Actually i dont want to clip to zero, so to how much voltage should i connect the + terminals?

I am a newbie so i ask a lot of questions. sorry. pls explain?

THANKS in advance.

(As you said clipping did occur), the graphs for the 1st two stages are shown below. Third stage is still not proper.

https://obrazki.elektroda.pl/6537612800_1359038020.jpg
https://obrazki.elektroda.pl/7879118300_1359038021.jpg


Also, can you pls explain why a negative db Freq response amplifies my sine (green) signal(should it not attenuate it?)? and why is the output shape not a perfect sine wave.


Thanks.

 

[barry]

Okay, you've got your first opamp configured as an inverting amp with a gain of approximately Vo= -Vi*4.

Any positive input will try and drive the output negative, but because your negative supply to the opamp is zero(ground), the output can't go negative. Still, if your input goes negative, the output will happily go positive.

When you say '1mv causes clipping' are you looking at the output of the first opamp? Is it clipping at the positive rail?

 

[LvW]

Also, can you pls explain why a negative db Freq response amplifies my sine (green) signal(should it not attenuate it?)? and why is the output shape not a perfect sine wave.
Thanks.

jose,
because it was not mentioned explicitely by barry: Use +/- supply voltages and everything will be OK.

 

[jose_peeterson]

thanks, Lvw.

can you please look at the circuit diagram and the resultant waveform and explain to me why the output (in red) does not have a smooth turning point (sine curve).

https://obrazki.elektroda.pl/9001376500_1359058631.jpg
https://obrazki.elektroda.pl/3817652500_1359058632.jpg

Thanks.

I set 1v as the input ac source.

 

[barry]

You are exceeding the specs of the part. It's MAXIMUM power supply voltage is 5.5V, youve got 10V.

 

[jose_peeterson]

hey guys,

My final output after some RANDOM settings is shown below. the schematic, transient response and the frequency response.

https://obrazki.elektroda.pl/9481178000_1359061834.jpg
https://obrazki.elektroda.pl/2406758500_1359061836.jpg
https://obrazki.elektroda.pl/9481178000_1359061834.jpg
https://obrazki.elektroda.pl/2406758500_1359061836.jpg


guys, can you please briefly explain the circuit operation.

Thnx.

 

[mspagon]

https://obrazki.elektroda.pl/3817652500_1359058632.jpg

Isn't the negative power supply pin on the op amp still connected to "ground"? I might be wrong but I though the negative terminal of the battery was at a 0 potential?

 

[jose_peeterson]

The positive terminal of the battery is connected to ground and the negative terminal is used as -5v. actually its supposed to be -2.75v only. My stupid mistake! now I need to reduce the gains to prevent clipping and also maintain flat bandwidth. damn shit! And I have no systematic way of doing this!

SS_3_STAGES_COMPLETE.JPG (50.8 KB)
tinadiag.jpg (54.6 KB)
freq_response.jpg (91.0 KB)

hey guys,
Please see my final circuit. May i know why my negative voltage of transient response has more peak amplitude than positive voltage?

Thanks.

 

[barry]

hey guys,

My final output after some RANDOM settings is shown below. the schematic, transient response and the frequency response.

https://obrazki.elektroda.pl/9481178000_1359061834.jpg
https://obrazki.elektroda.pl/2406758500_1359061836.jpg
https://obrazki.elektroda.pl/9481178000_1359061834.jpg
https://obrazki.elektroda.pl/2406758500_1359061836.jpg


guys, can you please briefly explain the circuit operation.

Thnx.

Can you identify what the signals are? All you're showing us is a bunch of sine waves.

 

[helpmejerry]

A fast glance suggest that the red one is the output and the green is the input. With out going into much detail its looks like a cascade of an amplifier, a high pass filter and a low pass filter. From the ac simulation it looks like there are two poles and a zero and small gain. For reference 20 dB equals to a gain of 10 if we assume the y scale is 20 dB.

I would try to lower the input amplitude of the signal to say 100 mV and check if the antisymmetry is still present. I assume that the voltage signal produced by the shock sensor is in the range of mV.


GL

 

[jose_peeterson]

helpmejerry is right!

The shock sensor output generally gives a half sine wave. can you explain to me why there is a phase lead or lag in the transient response of signals that are outside the flat band region? In the flat band region the phase shift is close to zero.

 

[FvM]

can you explain to me why there is a phase lead or lag in the transient response of signals that are outside the flat band region? In the flat band region the phase shift is close to zero

.
It's simply by nature of the filters. They have the minimum phase system property with a respective one-to-one relation of magnitude and phase.

 

[jose_peeterson]

I think we are missing the main objective.

When it comes to SIGNAL conditioning. What is the METHODOLOGY to build the circuits? I mean we could use the standard LPF, HPF and BPF. The filters in this case are unusual. I dont understand what each of the RC networks connected to the first TWO stages are doing. I know that the last stage is the 2nd order sallen key LPF.

THX.

 

[barry]

I think we are missing the main objective.

When it comes to SIGNAL conditioning. What is the METHODOLOGY to build the circuits? I mean we could use the standard LPF, HPF and BPF. The filters in this case are unusual. I dont understand what each of the RC networks connected to the first TWO stages are doing. I know that the last stage is the 2nd order sallen key LPF.

THX.

HUH???

Your original post asked NOTHING about what those RCs were for. All you said was that you were randomly sticking components in, and the output was saturating. Maybe if you actually posted THE MAIN OBJECTIVE we could answer it.

 

[FvM]

The circuits aren't particularly unusual, but possibly not serving the intended purpose.

Besides two corner frequencies, you didn't tell about the intended frequency characteristic.

 

[jose_peeterson]

HUH???

Your original post asked NOTHING about what those RCs were for. All you said was that you were randomly sticking components in, and the output was saturating. Maybe if you actually posted THE MAIN OBJECTIVE we could answer it.

so may i know what each rc network is doing in each stage.

thx

 

[FvM]

 

[jose_peeterson]

Thanks fvm.