TESORO GOLDEN SABRE

[waltr]

I think slew rate and band with at 11Hz is not so important. But as you said input bias current is higher and the voltage is lower, didn't even see that.
The LM308 is very expensive here, I prefer to use the TL071, good price and less chatters at max sense.

I use the TL081 for U104 & U105. The TL071 will work as well. I tried TL061,71 & 81. They all work but the TL081 is slightly better.
The LM308 is too expensive and hard to get even here in the USA.

Why 11Hz?

The center frequency of the bandpass filter.

Yes, the first differential stage (U103) acts as a Band pass filter at 11Hz.

The second differential stage, U104 &105, do need more bandwidth to respond to rising/falling outputs from U103.
Also U104 & 105 must have very low output offsets and high impedance inputs (TL0x1 have FET inputs).

Note: The 'sensitivity' adjustment Voltage into the LM393 comparators may need to go slightly negative if the TL0x1 outputs have a slightly negative offset. Measure these with a DVM.

[green]

What reply is the schematic on? Reply #20 doesn't appear to be the one your using.

[waltr]

What reply is the schematic on? Reply #20 doesn't appear to be the one your using.

Sorry, I referenced the TGSL EDU schematic.

Post #20's schematic does have Ref designators but circuits very much the same.

[green]

Sorry, I referenced the TGSL EDU schematic.

Post #20's schematic does have Ref designators but circuits very much the same.

Thanks, just trying to learn something. Wondering why the OP07 is weak. I get 7.2Hz instead of 11Hz, wondering if I'm calculating wrong. Like Nandor replied, BW probably not the reason. Offset volts is low, maybe offset current. With 1meg feedback, offset volts could be +-4mV. Don't know if 4mV could cause the problem. Maybe Nandor could measure output offset as is and with 1meg feedback jumpered. Any other thoughts why OP07 could be weak?

[Nandor]

My bad! I confused it with the first stage, it really has 7,23Hz. It's a high Q bandpass filter. On some of my pcb's I use 2 diodes in opposite directions as Compadre does, it cuts off the gain at higher voltages so the close range targets dont overdrive the filter, it's really fast this way.
Anyway 7.23Hz should be no problem for this opamp, it has 600khz bandwidth and in this filter it has only 22X gain. TL071 performs much better here.
I tried to measure the output of this filter but it's very noisy and I cant relocate my scope.
This opamp is used in guitar effects and I have read that it's not as good as LM308.

[green]

I tried to measure the output of this filter but it's very noisy and I cant relocate my scope.

You tried with a DVM? I would have thought a DVM would have averaged the noise, maybe not.

[Nandor]

At the output of the TL081 I have 8mV-12mV constant voltage, if it goes wild from noises it rises up to 120mV. The noise level is quite high here, I think that power cables and power supplies are the main sources.

[green]

At the output of the TL081 I have 8mV-12mV constant voltage, if it goes wild from noises it rises up to 120mV. The noise level is quite high here, I think that power cables and power supplies are the main sources.

Couple things I would try. Remove LM358(U103) if not soldered in or disconnect Rx and jumper Rx input. Check for noise and output offset with a DVM at U104 and U105. I know zero about the circuit but I'm wondering why a TL081 wouldn't be better for U103.

Duh, TL081 is single not dual.

[Nandor]

The Ic's are soldered, but I can try disconnecting the Rx cables and test the noise level, will do it soon. I would like to make small improvements if possible, I had many great finds with this detector and I see room for more. Noise is the limiting factor as Davor says, very interesting for me that at night time the detector has better range. Will try to find a noisless place at home to see the true sensibility of my new pcb. It runs from 5 volts+/- and has 33cm on a Euro coin with 10Vpp at the oscillator. Daytime same spot in room the detection range is 29-30cm only.
More work is needed to eliminate the noise, maybe it could reach more depth.

[waltr]

All of our detectors will pick up local EMI noise.
Many of us can not easily test detectors in our workshops due to this.

Any EMI noise will decrease the detection distance.

This is defiantly true of the TGSL I built and when my furnace kick on and testing is impossible.
I have also found some sites where the EMI noise levels are high.

There really is nothing that can be done if the environment has high EMI noise.

[Nandor]

Took some measurements today, Rx was disconnected from the coil. At the output of the preamp it had 12mVpp distorted line with a few peaks, not similar to any kind of sinewave. The frequency was unreadable, was jumping from 50hZ to several Khz. After that I measured the filters output, 50Hz mains distorted signal. I have tried many points but I got the same results.
Anyways OP07 does not like my pcb as NE5534, at least OP07 works, NE5534 gives continous sound.

[Nandor]

One solution to reduce noise could be a new pcb with a quad opamp for the filters so it would be much smaller and compact, like the IDX pcb.
What do you think?

[mostafalesnar91]

hi friends!
I made tgsl with two coils(10" and 6")
both coils do not sens more than 18cm for coins!
very low!!

[mostafalesnar91]

i nulled coil with multimeter on lowest possible value!!
All metal/disc part does not work!!

[Qiaozhi]

i nulled coil with multimeter on lowest possible value!!
All metal/disc part does not work!!

That's not surprising ... because you're supposed to null to a 20 degree offset between TX and RX, not to the minimum null.