audioguru

The book is WRONG!
The old LM358 operates very poorly at ultrasonic frequencies. It is the first low-power opamp so it is very slow.
The circuit has its inputs connected backwards so it cannot amplify.

__________________
Uncle $crooge

waginia

thank for ur ultrasonic curcuit,i think it is useful to me!

AVR_Craig

Hi All. This is my first post here. I have been having a heck of a time with what I thought would be a fairly simple task. Unfortunately, I don't know much at all about analog stuff, filters, etc. Days of head scratching and Google-ing have landed me here.
I am a little off-topic because I am not using a transducer, but everything else suits this thread. The threads I found about buried wires had no useful info.

I am trying to do a variation on the buried wire thing.
I found this circuit a while back and breadboarded it. In the wire I am using a 40 khz square signal originally generated by a spare PWM on a AVR micro.



I substituted an LM358 for the 348 pictured as well as the resistor and cap to set the freq of the 567.

I am using a telephone pickup coil for the sensor, although I'd like to be able to use an SMD inductor if possible.

It did not work at all, I now think the amp section is tuned for a freq other than 40khz. I read up some and modified it to a simple 2 stage amp, but with much too high amplification. Despite this, it worked well on the breadboard. I did a PCB for it and shrunk it down to about 19mm square.

It still picks up the wire, but now also scissors, pliers, nails, etc. I think it is picking up the internal signal from the 567.

Can anyone share a +5 volt circuit that will detect the wire loop, and just as important, not detect it when it is not near.

Is it possible to tell from the schematic if the op-amp is tuned for a certain frequency? I don't need to use 40khz, anything over 8khz should be fine.
If it matters, it will be used in a remote control car.

Any ideas are greatly appreciated. This is one part of a much larger project and I can't spend much more time on it alone with my limited knowledge in this area.

Thanks in advance,
Craig

Nigel Goodwin

So why didn't you read Audioguru's post above which says the LM358 is useless at ultrasonic frequencies?.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

audioguru

AVR Craig's circuit is designed for the very odd biasing of an obsolete LM3900 quad "Norton" opamp.

Ordinary opamps are incorrectly biased with the (+) input connected to the +5V supply. So they cannot amplify anything.

The datasheet for the very old LM348 shows a minimum supply voltage of 10V and its response drops above only 2kHz.

The lousy old LM358 has a minimum supply voltage of 3V but it also must be biased correctly. It hardly has any output at a frequency as high as 40kHz.

The MC34071 single, MC34072 dual and MC34074 quad opamps have a minimum supply voltage of 3V and a bandwidth to 100kHz. They also must be biased correctly like all ordinary opamps.

__________________
Uncle $crooge

AVR_Craig

Hi Nigel,

I did read it - yesterday.
This all started two months or so ago. I was trying to paint a bigger picture to get all the details in.
Please read it again. The first 3/4 of the post is how I achieved failure.

In addition to reading Audioguru's post about the 358, I read your posts about filters in the amp stage. That is why one of my two questions was:

"Is it possible to tell from the schematic if the op-amp is tuned for a certain frequency?"

The schematic doesn't mention a frequency and has no values for the 567's resistor and cap to set frequency. I don't remember why I chose 40 khz, but I can use whatever is best, or I can start over with a known good circuit.

audioguru

Hi AVR Craig,
Your terrible circuit doesn't work.
It is designed for a different kind of opamp (LM3900).
Its ordinary old opamps are too slow and are incorrectly biased.
The opamps are not tuned, the NE567 is tuned.

__________________
Uncle $crooge

AVR_Craig

Thanks Audioguru.

That is not my circuit, I found it online and it never worked.

I ended up using the 358 as a 2 stage amp fed directly into the 567. According to my scope, it did amplify the 40khz signal. Maybe the ones I got are much better than spec.

It does detect the signal loop wire, but since I shrunk the board, it now also detects metal objects. I think it is detecting the internal signal in the 567.

If I use one of the op-amps you mention with reasonable gain and better power decoupling on the 567, should I do anything else for filtering, etc.

What is the purpose of R3 and C4 in the schematic I posted?
Now I have them directly connected, could that also be part of my problem?

Should the 2 leads of the pickup coil attach to the input of the op-amp like the other circuit here with the transducer?

Thanks again for your help. I am way out of my normal realm here.

Craig

audioguru

Maybe it originally used an odd LM3900 opamp which is obsolete and sombody tried an ordinary LM348 opamp instead.

The datasheet shows that its max output is typically only 3V p-p at 40kHz and its max gain is only about 25.
But the NE567 is pretty sensistive if it is connected properly.

Attach the schematic of your 2-stage amplifier.

R3 is not needed and a direct connection between the opamp and the DC blocking capacitor is fine.
C4 is needed. All circuit examples in the datasheet of the NE567 show an input DC blocking capacitor.

It depends on your circuit. Probably not because a piezo transducer does not conduct. Your coil is a DC dead short.

__________________
Uncle $crooge

AVR_Craig

I didn't have a schematic for it. I did the PCB layout directly off the circuit on the breadboard. I was sketching it out to scan and post when I found a missing trace on the board. It is in the gerber file, but never made it to the board.

Anyway, it works now, as well as it did on breadboard. I do see the 358 clipping the signal at around 3 volts. I will order some MC34072 and a good selection of resistors and caps and see if I can improve it with some further research.

What value and type should I use for C4, the DC blocking cap between the op-amp and 567? Is there a formula for this based on freq? The datasheet shows it on every example, but offers no help on selecting it.

Thanks again for all your help. I am satisfied with the performance for now, it should only get better with the faster op-amp. Once I have this project done and I get more time, I will revisit this so I can actually understand all this better.
For now it is back to the land of 1's and 0's where everything is so much simpler.

Again, thanks.
Craig

audioguru

The lousy old LM358 limits its max output at 40kHz to only 3V because its output becomes a triangle waveform. It can't swing quick enough to go any higher before it must change direction.

The calculation for the blocking cap that feeds the input of the NE567 is calculated with the frequency where the response is reduced 3dB and the input resistance which is about 20k ohms (in the dataheet). To avoid loss then make the calculation at a frequeny that is divided by 5 (8kHz). So a 0.001uF (1nF) coupling capacitor would cause 8kHz to be down 3dB but 40kHz would be almost flat with hardly any loss. C= 1 divided by (2 pi fR).

Your 0.47uF capacitor was 470 times bigger than necessary.

__________________
Uncle $crooge

AVR_Craig

My scope shows this exactly.

Reading that makes me painfully aware of how little I know about analog electronics.
Thanks again for your help, I will post my results when I get it all straightened out.

Lesh

The LM358 has a GBP / Unity gain bandwidth of 700khz, so configured with a gain of 100 it should do 7khz. Configured with a gain of 10 it should to 70khz. With a gain of 1000 it will only do 700hz. A lot depends on the circuit doesn't it?

Edited to add, good point on the slew rate.

Edited again to add: A wise man said, "In theory, theory and practice are the same. In practice, they are not."

audioguru

Nope.
The bottom frequency scale is logarithmic so half a division is 0.316 times and not 0.5 times.
Therefore the GBP is 400kHz when its supply voltage is only from 10V to 15V as shown on a graph on its datasheet:
1) With a gain of 100 (40db) its response drops above about 4kHz.
2) With a gain of 10 (20dB) its response drops above about 40khz.
3) With a gain of 1000 (60dB) its response drops above about 400Hz.

The poor slew rate is shown on another graph on its datasheet. Its output is 13.5V p-p up to only 5kHz. At 40kHz its max output is only 3.0v p-p.

The better TL07x or MC3407x opamps have a gain of about 100 at 40kHz and their outputs have a max output of 27V at up to 100kHz.

__________________
Uncle $crooge

Lesh

Which graph? I think we are looking at different data sheets. TI and ST data sheets list a spec of 700khz min. ST does have a GBP to temperature graph.
Which manufactures data sheet are you looking at?