TGSL Tuning & troubleshooting

[Dennis the Mennis]

Your noise please

To everyone with a TGSL, both good and bad, please help with the following tests:

1a. With coil fully attached, please try to measure the noise at LM308 (U104) output pin 6 (using oscilloscope) for peak-to-peak measurement. I know this is difficult because it varies, make a good estimate.

1b. Please also measure the same noise with AC voltmeter and also say what frequency range your voltmeter is accurate for.

2. Please remove RX coil at PCB connection and measure DC offset voltage at LM308 pin 6, both channels.

3. Please report the air depth you measure with 1 Euro and/or US quarter. Measure for both minimum and maximum sensitivity pot setting. If too much chatter at maximum sensitivity, choose the most sensitive setting you can and describe the setting.

-----------------

Thanks. I want to do everything possible to figure out why some TGSL get poor depth, and how to fix. Your answers will help. I am thinking about the following concept:

Maybe noise level measured at LM308 (U104) pin 6 is cause of poor depth.

Noise breaks up the detected signal, and the small pieces cannot get through the low pass filter that follows.

We need to compare the background noise and DC offset of the LM308 U104 output pin 6 using the data you measure and post to this thread.

Regards,

-SB

Hi Simon ,

During some noise tests I found out that my LM308’s were not that good .
I’ve replaced them by a set of TL071 ( and for test purpose a set of TL081 ) and the results for my MD were great .
I’ve never had such an air depth !! ( on test bench in workshop )
28-29 cm on a 1euro coin ( with some hums and cracks ), the Tx coil is shielded and the Rx is not shielded ,so here has to be more experimenting to be done .

The strangest thing is that the noise levels for all 3 sets of chips ( LM308 / TL071 /TL081) were almost the same , so what the problem was ….

Anyway : the noise level on U104 and U105 pin 6 with the Rx coil attached is about 20-30 mV for both channels. ( measured with scoop , didn’t get any reading on my DVM )

The DC offset voltage with the Rx coil not attached is I think +2 - +4 mV ( also measured with ancient scope ) for both channels .
This is very hard to see due to the residual voltage left ( 5-10mV ) even with pcb in a shielded box.
So perhaps here’s something to gain , lots of experimenting to done .

With an air depth of 28-29cm and sens setting at 100% I did get however some chatters .
Reducing the sens to 75 % the chatters disappeared and the air depth was about 22 cm.

I did not tried DFBowers mod yet , first I’ll try to get my TGS ( not the TGSL ) stabile .

Regards ,

Dennis the Mennis

[simonbaker]

Hi Simon ,

During some noise tests I found out that my LM308’s were not that good .
I’ve replaced them by a set of TL071 ( and for test purpose a set of TL081 ) and the results for my MD were great .
I’ve never had such an air depth !! ( on test bench in workshop )
28-29 cm on a 1euro coin ( with some hums and cracks ), the Tx coil is shielded and the Rx is not shielded ,so here has to be more experimenting to be done .

The strangest thing is that the noise levels for all 3 sets of chips ( LM308 / TL071 /TL081) were almost the same , so what the problem was ….

Anyway : the noise level on U104 and U105 pin 6 with the Rx coil attached is about 20-30 mV for both channels. ( measured with scoop , didn’t get any reading on my DVM )

The DC offset voltage with the Rx coil not attached is I think +2 - +4 mV ( also measured with ancient scope ) for both channels .
This is very hard to see due to the residual voltage left ( 5-10mV ) even with pcb in a shielded box.
So perhaps here’s something to gain , lots of experimenting to done .

With an air depth of 28-29cm and sens setting at 100% I did get however some chatters .
Reducing the sens to 75 % the chatters disappeared and the air depth was about 22 cm.

I did not tried DFBowers mod yet , first I’ll try to get my TGS ( not the TGSL ) stabile .

Regards ,

Dennis the Mennis

Thanks for posting results.

It sounds like we're barking up the right tree here.

Yes, why the chips make a difference? I would think it is the output offset voltages are different. The LM308s maybe had weird offsets.

I would expect the noise to be the same, because I don't think practically any of the noise comes from the chips themselves. It's mostly from the coil and stuff near the front end. So that is a good finding.

If you do put in dfbowers mod, you might find the old LM308 chips work better at some pot setting. Or not. A good experiment to try.

Also get the MD outside away from EMI noise and see what happens.

At least you have a nicer MD now, and more hope something similar will work for others. Nice work.

Cheers,

-SB

[dfbowers]

Hi Simon ,

During some noise tests I found out that my LM308’s were not that good .
I’ve replaced them by a set of TL071 ( and for test purpose a set of TL081 ) and the results for my MD were great .
I’ve never had such an air depth !! ( on test bench in workshop )
28-29 cm on a 1euro coin ( with some hums and cracks ), the Tx coil is shielded and the Rx is not shielded ,so here has to be more experimenting to be done .

The strangest thing is that the noise levels for all 3 sets of chips ( LM308 / TL071 /TL081) were almost the same , so what the problem was ….

Anyway : the noise level on U104 and U105 pin 6 with the Rx coil attached is about 20-30 mV for both channels. ( measured with scoop , didn’t get any reading on my DVM )

The DC offset voltage with the Rx coil not attached is I think +2 - +4 mV ( also measured with ancient scope ) for both channels .
This is very hard to see due to the residual voltage left ( 5-10mV ) even with pcb in a shielded box.
So perhaps here’s something to gain , lots of experimenting to done .

With an air depth of 28-29cm and sens setting at 100% I did get however some chatters .
Reducing the sens to 75 % the chatters disappeared and the air depth was about 22 cm.

I did not tried DFBowers mod yet , first I’ll try to get my TGS ( not the TGSL ) stabile .

Regards ,

Dennis the Mennis

That's very good news! It sounds like you are at least in the ballpark.

[dfbowers]

That's very good news! It sounds like you are at least in the ballpark.

BTW.. who is the manufacturer of your LM308s and TL071s? I honestly cannot tell the difference between TL071acp (TI) and LM308n (ST) performance wise..

[Vladimir]

My big apology

My big apology, because of poor visible marks on the capacitor, I have the wrong value.
The correct value would be 560pF to 720pF that is necessary to determine, depending on the type of IC which is used as the U105. Once again I apologize to everyone.
Greeting

[mikebg]

Another apology

Another apology. It relates to my posting # 167, which has no attached circuit diagram. I use some other circuit diagram posted by Ivconic, which seems quite different from what shows Stefano in posting # 2 and Vladimir in postings # 171 and # 185. My postings # 167 refers to the following circuit:

Can someone post me the whole circuit diagram, which discusses Vladimir?

[dfbowers]

My big apology, because of poor visible marks on the capacitor, I have the wrong value.
The correct value would be 560pF to 720pF that is necessary to determine, depending on the type of IC which is used as the U105. Once again I apologize to everyone.
Greeting

Hi,

Are you suggesting that this mod be applied ONLY to pin 5?

Don

[Vladimir]

Can someone post me the whole circuit diagram, which discusses Vladimir?

This is a circuit diagram with the changes about which I write.

[Vladimir]

Hi,

Are you suggesting that this mod be applied ONLY to pin 5?

Don

Hi
Yes

[simonbaker]

BTW.. who is the manufacturer of your LM308s and TL071s? I honestly cannot tell the difference between TL071acp (TI) and LM308n (ST) performance wise..

I originally wondered why U104 and U105 were "special" precision chips when the others were more ordinary.

Now I think it is just to try to control the output voltage offset, since it is so critical to the detection logic.

Otherwise, wouldn't it seem any old garden variety op amp could handle the signal? The bandwidth is tiny, maybe 10 Hz. Heck, even 100 Hz is tiny. The gain is controlled by feedback, so again most op amps have plenty. I don't think that noise from this chip is significant, it comes more from earlier places. I think it's all about stability of the output offset voltage.

For kicks, I'd like to see if cheapo LM358 or such would work there if you added individual offset control, like a dual dfbowers mod. However, if the LM308 has more offset stability over time, that would be an important reason for its use also.

-SB

[dfbowers]

This is my method for solving the problem of noise.
Try, it is very effective.
Greeting

Another observation.. depending upon which LM308 this mod is applied to, it can turn the LM308 into a high frequency oscillator. (I tried it.) Maybe feedback to pin 3 of the LM308?

[Vladimir]

Another observation.. depending upon which LM308 this mod is applied to, it can turn the LM308 into a high frequency oscillator. (I tried it.) Maybe feedback to pin 3 of the LM308?

I agree, but for me it works perfectly. Perhaps the reason the previous modifications that I used ...

[mikebg]

ATTN: capacitive load

Vladimir, your capacitive load of opamp U105 creates instability. In theory it is avoided by a resistor as shown.

[simonbaker]

TGSL standard vs. Vladimir comparison, LTSpice

This is a circuit diagram with the changes about which I write.

I did a crude LTSpice comparison of the TGSL basic filtering, comparing the standard TGSL with Vladimir's modifications.

1. The top pic shows the Frequency response (Bode) plot comparing TGSL standard and Vladimir.

The green trace is TGSL standard.
The blue trace is Vladimir mod.

Note that the standard TGSL band center is about 8 to 9 Hz, Vladimir center is about 4 or 5 Hz. Vladimir gain is slightly lower, -1.7 db.

2. The middle pic is the response to a target (I made up the target waveform, may not be realistic).

The green trace is the TGSL standard.
The blue trace is Vladimir mod.
The red trace is the raw target signal.

Detection occurs when the response is above 0 volts.

Note that there is more delay (25 milliseconds) with the Vladimir mod, as expected with a slower response circuit. This would affect pinpointing. The detection pulse with Vladimir mod is wider (102 milliseconds vs. 70 ms), which is nice, makes longer audio beep.

Of course my reference target signal is probably not realistic, but it gives you a comparison.


3. The bottom pic shows the frequency response of each filter stage for Vladimir mod.

The green trace is the low-pass filter of R26/R25 - C15/C12.
The blue trace is the LM358 band-pass filter.
The red trace is the LM308 band-pass filter.
The cyan trace is all the stages combined.

I would note that not all the stages have the same center as the final stage. This may mean the combination is not quite optimally designed. It seems that the first stage low-pass filter is a little agressive, maybe that is the reason. However, if you look at the breakdown of the standard TGSL op amp stages (http://www.geotech1.com/forums/showp...&postcount=250), you see the centers are not well matched there. So Vladimir's mod is not any worse from that regard I think.

---------------

Conclusion: Vladimir mod looks like it achieves a slower responding TGSL, which may be better for deep targets, but more difficult pinpointing for shallow targets. The standard TGSL stages may also not be balanced as well as they should. Vladimir mod should give a slightly fatter audio beep. I would expect Vladimir mod may have less noise chatter, but can't really tell from this. Vladimir's gain is slightly lower (-1.7 db) than standard TGSL.

However, this is just theory. The MD must work well in practice. Only building and testing will say if a particular mod is useful. And of course different mods have different uses. Vladimir's mod may be what I want for dry sand beach hunting where objects go deep. Standard TGSL better for hard, mineralized soil, shallower targets, better pinpointing.

-SB

[simonbaker]

Vladimir, your capacitive load of opamp U105 creates instability. In theory it is avoided by a resistor as shown.

I think that leading resistor is recommended also to make the circuit consistent for different chips. What values of R and C are best, is the question. Also, is this the best place to put this filter? And there should be a similar filter on the disc channel to keep them balanced, I think.

-SB