RELIC HAWK

[Max]

Updated...

It is becoming very dificult to increase nº of components...

-2X 100K resistor to - from U5 and U9
-1.6M disconnected from +
-U5 power rails fixed.
-Some minor stuff like more separation on some traces/pads.

Please someone verify !

regards,
Fred

Hi Fred,
I had a bad gnd connection at U7... due to solder on top layer... after fixed that device started working somehow but very bad...

1 eur coin at 20cm but sound is very bad... chatters always present ...it's unstable and noisy.

I followed signals but just relevant anomaly is that U7 has on output at pin7 something 9Vpp@50Hz , that's huge amount of sinusoidal noise!

Let me explain... I found that at U4 I can trace the introduction of about 20mV noise at 50Hz that then propagates and is amplified by following op amps along the chain... so I end up with 9Vpp at U7!

Now the problem is... the coils are shielded... and I'm using a very low ripple powersupply... so where's that dang 20mVpp noise enter the PCB ??? Where is it from ???

I haven't fluo-lamps there... I haven't anything that emits strong 50Hz radiations... but my dubt is device is so sensitive that source of noise is probably farther than I expect to be.

Another consideration is about decoupling caps... I still haven't them on circuit... so maybe the problem could be there...

Kind regards,
Max

[Fred]

Hi Max,
That´s strange.reminds me of some wrong ground reference at some place.

You didn´t forget the ground clip at you scope probe?
I will have to wait for the 27m4 and 24m2 Ic´s so i cant try it all yet.
I have populated it with resistor .Look great already
regards,
Fred.

[Max]

Hi Max,
That´s strange.reminds me of some wrong ground reference at some place.

You didn´t forget the ground clip at you scope probe?
I will have to wait for the 27m4 and 24m2 Ic´s so i cant try it all yet.
I have populated it with resistor .Look great already
regards,
Fred.

Hi,
yes I put the clip at GND level... and got such stuff... changed ic at U7 and got less "noise" : tested NE5532 that has more CMRR but problem is big and not related only to supply I think... there's something wrong maybe about GND tracks ? Ground loop ?

I don't know... will look for that but it's clear something is wrong if I get such enormous 9Vpp at U7!

The sampler works fine... at least seems work nice on the scope.

Kind regards,
Max

[Fred]

Max,
Did you try to short-circuit the RX coil?
Or the opamp inputs untill you find where it enters the circuit
Regards,
Fred.

[nakky]

Hi Max and Fred,look at U11 is it gnd or - on point 12 .Grt Nakky

[Fred]

Hi Max and Fred,look at U11 is it gnd or - on point 12 .Grt Nakky

Hi Nakky,
U11 is reversed, pin 2 is on the other corner , what you see is pin 9

Thanks
Regards,
Fred.

[Max]

Max,
Did you try to short-circuit the RX coil?
Or the opamp inputs untill you find where it enters the circuit
Regards,
Fred.

Hi,
yes, I did but find at U4 the "source" of sinusoidal noise... well... better say I can trace till there.

Maybe it's external noise (almost sure it is cause of the 50Hz frequency)... I still don't know... will try to find the hot spot if I can.

Kind regards,
Max

[Max]

Hi,
tested with battery to be sure it's not powersupply the cause of noise... and device works bad way also... the noise voltage increased of 1Vpp !

So it's not supply ripple there... but probably device gets external noise by the coils or some pcb track...

I think I found the source of noise: surface wave at 50Hz from buried powerlines

The strange is that with other detectors I couldn't see it so BIG!

Even with BandidoII , which I got lot of noise troubles I had low 50 ripple at op amps.

So I think is a bit strange that I get so high noise level now...

Kind regards,
Max

[Max]

Hi Fred and all,
last night I dreamt about a kind of "wild weasel" mission... maybe the effect of too many whiskey last evening... but but but... in the dream I was considering maybe other stuff...

So... I was flying over north vietnam (maybe) and approaching a viet SAM station making low altitude fly with something F4 (I think...could be some F4... dream was so confused...)

Ok ok... now you'll ask why I'm telling you that..., nice, let me explain...

during approaching at low altitude I made a long reconnaissance mission to find if there were other SAM stations all around in the area but nothing... then lastly I pointed straight to the mission SAM to drop the missile...then I wake up...a bit confused and went to TV watching BBC till I had the brilliant idea...

So what ? The all around reconnaissance is the key...

I made it today on the board... and found an interesting thing: do you see the two 220nF caps before U4 inputs ??? Now look at pin3 and pin5 of U4 ???

Don't you see anything ???

The GND points are very far one each other: GROUND LOOP

I connected the scope between gnd points and find out there's consistent (low magnitude of course) sinus signal at 50Hz! The signal was then amplified thousand times... and I got the 9Vpp at U7.

So, so, so do you understand me now ???

The GND path is so long and is also about like a "single turn" shaped I can get consistent amount of inductively coupled sinusoidal noise there.

Solution (?) ...well... I made this and things now are better than before:
cut the upper track that connects U4 pin5 to U8 gnd point; then I made a straight connection using rigid pvc covered wire to the 220nF caps I was talking above... and

I got a drop in noise at U7 pin7 from 9Vpp to 1.2Vpp 50Hz noise!

Now device seems get 1eur coin at 26cm about easy with the 137x255DD stuff and it's not so bad considering the noise is scaled down BUT still there.

The strange is that I get this in ALL-Metal but in disc stuff is different... also about audio signal ...it's like cracking sound, not clear tone. I get clear tone, beep at 1.25Khz just nearing the coin around at 17-18cm now.

I think the noise reduction was good but NOT enough for proper operations... so maybe something else is needed now... I need dream something good this night!

What do you think ???

Kind regards,
Max

[Aziz]

Hi Max,

try to decouple all IC's with 100 nF ceramic capacitors on the power supply pins. Particularly fast digital switching IC's should be decoupled allways, to reduce the ground loop current flows. The capacitors should be placed as close as possible to power supply pins of the IC.

Fast operating operational amplifiers (high dU/dt) should also be decoupled. Dual voltage supplied Opamps need two decoupling capacitors to Gnd.

This will suppress long track current flows on power supply signals.
Aziz

Update:

Your ground has a high DC impedance. Do a simple voltage divider with R-R network and do a voltage follower with an Opamp to produce a lower impedance ground.
;-)

[Max]

Hi Max,

try to decouple all IC's with 100 nF ceramic capacitors on the power supply pins. Particularly fast digital switching IC's should be decoupled allways, to reduce the ground loop current flows. The capacitors should be placed as close as possible to power supply pins of the IC.

Fast operating operational amplifiers (high dU/dt) should also be decoupled. Dual voltage supplied Opamps need two decoupling capacitors to Gnd.

This will suppress long track current flows on power supply signals.
Aziz

Update:

Your ground has a high DC impedance. Do a simple voltage divider with R-R network and do a voltage follower with an Opamp to produce a lower impedance ground.
;-)

Hi,
thanks for hints... capacitors are on the road... I had to mount them already but had no time... and I have to do on the bottom side of pcb cause there's no more room for this stuff on top side.

About GND I'm following original schematic, in RH the GND level depends on the zener diode voltage... the doubled zener voltage is GND level respect to the - rail, so I cannot simply make the R-R stuff here... but there's sure room for improvements on this design!

Kind regards,
Max

[Fred]

Hi Max,

This is very interesting, thanks !
Nice that you have found it.It is not so easy.
I made a new modifications to the PCB , according to your founds : a ground plane.
This bring several vantages ,for the decoupling caps in particular.
I dont think we can consider anymore a single layer as an easy option anyway, but it still should be possible.
With this groundplane it will give more work to do the clearance hole for components, but many of them will be grounded on each side.I also could have removed some "copper side" traces, as they were not needed, but i didn´t , this was not really necessary and will still allow a single side design.

Let me know what you think about this,it looks confusing at first because of the different colours but if you remove layers you can see it better.Also print preview works well.

Your results looks good, 26cm with a not completely optimised design is nice.

Please continue the good work!
Keep dreaming

Regards,
Fred.

PS: updated for better GP

[amtech2005]

RELIC HAWK

I am just a little upset today !!! So Thanks Fred and Max for your hard work and Dedication Here !!!! When this project is completed it will be better than the original !!! That is why this Relic Hawk is discontinued because it did have problems in the field. Remember what another member said here about a month ago , about the Relic Hawk working good only with only certain ground or surface condition's ??? Sometimes companies throw out models too fast , one after the other , without field testing enough !!!!!!!! And then you read about problems in the field . And that's why this " FORUM" has modification posting's all the time. And also people are alway's asking how to make their detector better !!!! And they are the one's that overpaid several hundred dollars for their detector and got very dissapointed when they were out in the field !!So all the metal detector manufacturers around the world "with their 10 models or more for each company" being manufactured .Why so many models ??? So the Detector manufacturers sould not Complain Here ever but "LEARN FROM US " !!!! . I am sick and tired of Detector companies that mislead Everyone and say "TOP OF THE LINE". Top of the line in what catagory?? I am not here to slander anyone . So Carl is welcome to delete this post if needed . Many people work hard for a living and when they go out and spend $600.00 or $700.00 for that advertised "Best Detector" and get only 60% depth of a "Real Top Of The Line Unit " , Yes, Something is definitely Wrong with the detector industry !!!!!!!!!!! This post is MY Opinion , Thanks for reading , .....................So to me "THE MAIN PURPOSE OF THIS FORUM IS TO BUILD A GREAT DETECTOR AT A REASONABLE PRICE OR FIX THE DETECTOR THAT WAS OVER PRICED by SEVERAL HUNDRED DOLLARS"!!!!!!!!!.............Eugene.

[nakky]

Good morning Max,tell us what did you dreaming last night?

[Max]

Good morning Max,tell us what did you dreaming last night?

Hi,
well... nothing special... some videogame stuff resident evil (I played RE4 last night for a while...) nothing important about the development here...

Today I added the 100nF caps on bottom side... seems strange stuff now there but works: 0.35Vpp of noise at pin7 of U7

So, yes, caps do their job excelent way... with other mods.

Now device is much more stable, though not perfect behaviour... I have good things in all-metal but bad stuff in disc mode: seems doesn't work when in disc mode, don't know why.

I think there's still something wrong but I cannot say where at now...

The advantage is now I get "clean" beeps with the coin and other stuff not cracking sound each time... that was due to instability and still too much noise.

All this using powersupply... don't know with batteries... but I think now will be much quiter than last time I tested (hell of noise)

Kind regards,
Max