TESORO GOLDEN SABRE

[simonbaker]

Hey,
1.The LM308 as I understand reading the whole topic probably twice is that there isn't any other good alternative (except the LT1008 which I now have)
so I don't think it's a good idea;
2. The BiPolar - definately yes I am removing those since I got bipolars by Nichicon, also the the 2 pairs of resistors could be replaced with 220K
3. Yep the 1K is missing, Ill add that, also the new audio mod
4. Adding ground plane - thanks i will watch for not doing loops in the GND plane

Honestly said I am making the board bigger, my box is 160x100 and have lots of room, I had smaller board working on but after ordering those parts (and mainly the Wimas - huge footprints) - nope not smaller but bigger board I need.

New idea:
I already asked Max about it, but still.... look at this picture, see the similarity? Could it be used for this project (it's the audio from relic hawk)??

Are you using the battery check? If not, get rid of R48, R49, R50, R51 (includes back-to-back resistors) -- you won't need them, i.e. the 220K at all! That part of the circuit left over by mistake.

Can somebody comment on R47, R55?

I think there may be many ways to experiment with the audio, try it!

Cheers,

-SB

P.S. Thanks again to Ivconic for original excellent layouts -- very appreciated.

[eclipse]

Hey,
1.The LM308 as I understand reading the whole topic probably twice is that there isn't any other good alternative (except the LT1008 which I now have)
so I don't think it's a good idea;
2. The BiPolar - definately yes I am removing those since I got bipolars by Nichicon, also the the 2 pairs of resistors could be replaced with 220K
3. Yep the 1K is missing, Ill add that, also the new audio mod
4. Adding ground plane - thanks i will watch for not doing loops in the GND plane

Honestly said I am making the board bigger, my box is 160x100 and have lots of room, I had smaller board working on but after ordering those parts (and mainly the Wimas - huge footprints) - nope not smaller but bigger board I need.

New idea:
I already asked Max about it, but still.... look at this picture, see the similarity? Could it be used for this project (it's the audio from relic hawk)??

[simonbaker]

For the things Max says ( about old types of the components in TGS)
I ordered 3 weeks ago components worth $180 (and damn im still waiting for them) just to push up as much as I can the electronics. The components include all (except pf caps) Wima - series FKP2 and MKP2 and electrolytics caps: Nichicon muse fx. All resistors are ultra low noise 10uV CMF55 series. Transistors are standard onces: BF245, PN2222/PN2907 (plastic ones). The IC: LT1057 in the front stage, two LT1018 comparators, LT1008 insted the 308, and 2xOPA2277.
The board, I decided to use Ivconic's, but i had to change almost all capacitors with other bigger footprints for the Wima's. Also the box I ordered is large, 160x100 i can add some stuff.
The layer with the traces I decided to add ground plane, some of the jumpers will be gone cos of that and also will order for manifacturing, its about 30$ so not that high.
As soon as I am done with all that will post the results, hope I am not throwing money on the wind ...

If you can, it would be nice to have a "minimal" PCB layout available for TGSL based on Ivconic's design with small changes:

- Use a "dual" IC instead of two LM308

- Remove unused resistors R48, R49, R50, R51 (including the back-to-back 100K substituting for the 200K R48, R51)

- Replace back-to-back 10 uF capacitors with single 4.7 uF Bipolar for C14 and C17

- Add the missing R27 resistor if it is still missing

- R47, R55 -- where are they? They seem to be on schematic, not on PCB. Not needed?

A goal is to make the PCB small enough to fit in a CD Tray of an inkjet printer, it would open possibilities for printing directly on board. It almost fits now.

Other PCB layouts would also be useful with different part substitutions so we have alternatives depending on available parts.

Are you using the Sprint layout program, or something else?

Regards,

SB

[eclipse]

Ok good good news all the parts are here! Nice nice a can now start building
the TGSL! Will post the more interesting data here
Almost all parts (except Bourns pots which don't look so healthy for this price) are damn good.
Yep Ivconic Ive seen that trick on ML machines - foil cage attached to cable with paper clip (hah!)

[Max]

>>On component side try to void gnd connections on every component that goes to gnd.

What does that mean?

Regards,

SB

Avoid ground loops.

[simonbaker]

>>On component side try to void gnd connections on every component that goes to gnd.

What does that mean?

Regards,

SB

[ivconic]

For the things Max says ( about old types of the components in TGS)
I ordered 3 weeks ago components worth $180 (and damn im still waiting for them) just to push up as much as I can the electronics. The components include all (except pf caps) Wima - series FKP2 and MKP2 and electrolytics caps: Nichicon muse fx. All resistors are ultra low noise 10uV CMF55 series. Transistors are standard onces: BF245, PN2222/PN2907 (plastic ones). The IC: LT1057 in the front stage, two LT1018 comparators, LT1008 insted the 308, and 2xOPA2277.
The board, I decided to use Ivconic's, but i had to change almost all capacitors with other bigger footprints for the Wima's. Also the box I ordered is large, 160x100 i can add some stuff.
The layer with the traces I decided to add ground plane, some of the jumpers will be gone cos of that and also will order for manifacturing, its about 30$ so not that high.
As soon as I am done with all that will post the results, hope I am not throwing money on the wind ...

Ground plane is ok idea; will improve S/N ratio noticeable. I tried once and it was successfull. Reason i am not doing it anymore are new enclosures - lets me to add foil "cage" inside...
Yet...adding ground plane is good idea.
One more thing;
i suggest you not to connect ground plane on every through hole. Keep gnd traces connected at bottom layer as those are already and use just one through hole connection to gnd plane. Just one...not many.
On component side try to void gnd connections on every component that goes to gnd.
If oposite...gnd plane may turns to produce more noise!!! Ha,ha,ha!
Strange but true!
Already checked and experimented...
Cheers!

Hello Eclipse

Transistors are standard onces: BF245, PN2222/PN2907 (plastic ones).

Look for the e.b.c. pins !

[Fred]

I noticed in Compadre thread that a mod is to drive TX oscillation directly from battery voltage instead of regulated supply to get higher voltage, more depth.

Anyone try that for TGS?

The Relic Hawk also has a unregulated power supply at oscillator...

[driscoll]

Thank you Max for this answer very explicit.

Bests regards

[eclipse]

For the things Max says ( about old types of the components in TGS)
I ordered 3 weeks ago components worth $180 (and damn im still waiting for them) just to push up as much as I can the electronics. The components include all (except pf caps) Wima - series FKP2 and MKP2 and electrolytics caps: Nichicon muse fx. All resistors are ultra low noise 10uV CMF55 series. Transistors are standard onces: BF245, PN2222/PN2907 (plastic ones). The IC: LT1057 in the front stage, two LT1018 comparators, LT1008 insted the 308, and 2xOPA2277.
The board, I decided to use Ivconic's, but i had to change almost all capacitors with other bigger footprints for the Wima's. Also the box I ordered is large, 160x100 i can add some stuff.
The layer with the traces I decided to add ground plane, some of the jumpers will be gone cos of that and also will order for manifacturing, its about 30$ so not that high.
As soon as I am done with all that will post the results, hope I am not throwing money on the wind ...

[Max]

Hi Max,
Hi Ivconic,

I read your articles and I must say that it is impressive. Golden Sabre would be more efficient than the bandido ?

Hi,
if we talk about standard, factory made, detectors... I think bandidoII is better than TGS, even if difference must be small on same coils.
BandidoII is newer than TGS... have better components by default.
Also have manual GEB and all-metal no-motion... TGS haven't in factory made : just GEB preset and motion mode.

If we talk, instead, of homemade clones... well... it's really difficault provide an answer: a properly made bandidoII tuned for max sens with an homemade coil it's a big beast... like homemade TGS with big DD coil.

Which clone is best ?

Much depends on fine tuning and coils: with extremely good tuning the bandidoII circuit is better than TGS circuit, no dubt. That's also cause bandidoII use by default lower noise components... but if you replace all them in TGS too you'll not notice such good performances.
The notch section of TGS is a source of noise... most of the time: in TGSL this is avoided cause you haven't notch, but I tryed with same bandidoII components and didn't get same good results.

Unfortunately, if you simply replicate bandidoII and use a standard coil you'll end up with a good detector but not as good as TGS with big custom coil is.
Also... I made a DD22cm that's really better of CC 9x8'' that, in theory, must be very good about depth. Inconsistency at coil manifacturing I think... some few you can buy are very good...but the majority are average coils. Or simply small mismatch with TX oscillator or RX preamp cap ???

So you need always custom, homemade coils for max performances.

BandidoII works really good on DD coils, any size...to 30cm I tested and pinpointing is very nice using no-motion feature, also on large coils. You can mod TGS for that feature too, of course, or use external pinpointer for elusive stuff.

The big advantage of TGS is that there are plenty of informations, details and data you can use for construction... where bandidoII lack many useful informations.

Another problem with bandidoII is that, it's generally more sensitive to steel stuff than TGS... so disc is very critical also, and works good when you set at around 4... in TGS your iron/steel disc is very good starting at 0... usually you use 2 as setting just to eliminate small foil stuff, iron is already all rejected... steel get only on e.g. large nails.

So, the answer isn't, cannot be definitve... and depends much on DIY abilities.

In my case resulted that bandidoII has little more sens than TGS... but this changing many things at circuit and components: it's not bandido anymore! On standard schematic/components I get about same detection as with TGS w/ same coil.

Kind regards,
Max

[driscoll]

Hi Max,
Hi Ivconic,

I read your articles and I must say that it is impressive. Golden Sabre would be more efficient than the bandido ?

[simonbaker]

Hi Simon,



I am interested on your results. Post it here as soon as you have some.

But I am very sceptical. It will cause slightly drifting due to charge decrement. The voltage of the cells will go down, even the amount is not much. This could cause much drifting problem.
Aziz

I'm sure you're right, but I'll test sometime. What do you think will drift? Even if frequency drifts a little, would the phase relationships change that much? In fact, we all have different frequencies due to differences in parts, coils etc. If drift very slow, motion-mode detector should not be bothered. But I am not expert.
-----------
I like the 9.6 v rechargeable battery pack for RC cars, etc. I plan to try using that with LM2941 low dropout regulator (instead of LM7808 ) with TGSL also as experiment. It seems like a waste to use 12 v to get 8v. LM2941 should make alkalines run longer also. I got sample of LM2941, don't know where to buy yet.

Regards,

-SB

[Aziz]

Hi Simon,

Ok, thanks.

I plan to experiment with 7 rechargeables = 7 * 1.2 = 8.4 volts, no regulation at all (leave out LM7808 ), just big capacitors and see how it works and any problems. I'm thinking original design was for alkaline and regulation was mainly for providing more hours of use as battery voltage dropped slowly from 12 v to about 10 v. As you say, rechargeables have flat discharge curve, maybe regulation no so important? I'll try it after build working MD, just slowly testing and building first.

I am interested on your results. Post it here as soon as you have some.

But I am very sceptical. It will cause slightly drifting due to charge decrement. The voltage of the cells will go down, even the amount is not much. This could cause much drifting problem.
Aziz