goldscan IV :is there an error/mistake in timing circuit ?

[GoldscanIV-addicted]

Hi all,
I'm just thinking that it is a mistake in the 555 section of the timing.
Let's explain: the first tlc555p ic is connected as astable using 47nF capacitor and 6k8 ohm resistor plus 47k resistor and 10 k trimmer, so calculating (using national datasheet formulas) timings I've obtained are:

T1= 220uS and a period T of 2.3 ms to 1.973 ms (from 434Hz to 507Hz)

where T1 is the "on" or "tx" pulse duration that is not equal to the 100us we can see written in the schematic under the tlc555p block.

I built a prototype on breadboard to be sure and measure a tx pulse of 228us. Frequency is in the range 440-520 Hz.

This will mean more tx pulse duration and more current drawn and so more power but more battery drain.

Anybody knows what's going on ? What are the real timings of goldscan iv ?
Thanks

[Guard]

PCB

Hi
I am interested to build goldscan IV .
Please send PCB to my Email:

[email protected]

[GoldscanIV-addicted]

Sorry , I haven't any pcb but some info (maybe useful)

Hi, I've just tested the timing circuit and front end on a breadboard with jumpers.

I get:

T1 = 225us (measured and close to calculated) and T as in the previous post
confirms the calculus using R1 = 47k+10k (that is for max period)
T1 = (R1+R2) * C * ln2 = (6800+57000) * 0.000000047 * 0.693147 = 0.00208 s

T2 = R2 * C * ln2 = 6800 * 0.000000047 * 0.693147 = 0.000221 s

Tmax = T1 + T2 = 0.0023 s ---> 434 Hz

then i have also (measured values):

delay = from 10 us to 90 us

sampling pulse duration = 17 us (calculus gives 15us) the same for the 4 pulses

sampling pulses separation (both) = (about) 200 us (calculus gives 200us)
(they are close becouse of low tollerance components used and matching
4538 caracteristics)

disc sw1-c pos 1,2,3 ---> trains separation = 45 us (no, all metal, sel 1)
disc sw1-c pos 4 ---> trains separation = 110 us (sel 2)
(i think that is for iron disc to have a huge separation in train of sampling pulses, but really don't know)

That confirms all my calculations but the TX pulse that is 225us instead of 100us declared in the schema.

I've started working on a pcb design but the schematic it's very complicated (for me) but I think I can do it in a month or so (in spare time) but I'm not sure if I can minimize crosstalks and noise becouse I'm not a specialist (like Eric Foster)

BUT :
I don't know if it's legal to post any pcb for a commercial design like this one in the forum and anyway I won't break any law, rights, patents. I'll be right to ask the owner before doing such a thing.


Best regards,
Max

[Reg]

I believe the original Goldscan IV had a pulse width of about 250 usec. The minimum delay was set at 15 usec and a 15 usec sampling time. Some of the settings were set via a pot in the circuitry.

Now, are the original settings the best, or should they be altered? That depends upon what you want to do with the detector. A long pulse is needed to hunt for silver and copper objects, while a shorter pulse length can be used for small gold objects.

A good compromise is to add a pot that will allow you to adjust the pulse length from something like 50 usec to 250 usec and then one can experiment to see what works best for them.

Reg

[GoldscanIV-addicted]

Thank you Reg : now I understand where is the error

Hi Reg,
thanks for the info on the original goldscan IV timings. If the tx pulse was 250us then I suppose that the label "100 us" in the schematic posted is a little mistake and so everything else will be ok in the timing section e.g. with all components values. I think that the consumption indicated (220ma) is right for 250us of tx pulse at minimum freq (about 430-450 Hz) with the coil data specified.

I've realized the front end using lm709 (the plastic type) , ua709 (metallic case - very hard to find for me) and the ne5534a (philips -very good amplifier).
With lm709 I had a lot of drift and valuable noise in the output ; with the ua709 had a good stability and less noise but the best is when using ne5534a --> really stable and less noise than with the others but I must eliminate compensation network and put the right offset nulling trimmer.

I think that this schematic is a very good engineering work.

Best regards,
Max

[Reg]

Hi Max,

Yes, the NE5534 is a better choice for the preamp.

Now, there area a couple of minor changes that should be made to the schematic. On sheet 2, there are two J113's that need a ground. Each one is below a 1/4 of IC 2 and have a 100nf connected to the D connection. The D and the bottom of the 100nf should be connected to ground on each of the two FET's. Also, a 470K should be connected across the D and S of the FET's. This creates the autotune. You can change the value of the resistor to larger or smaller depending upon how fast you want the circuit to retune. However, you should keep the two resistor values equal if you do change them.

I have attached part of the schematic showing the grounds and resistor. Sorry about the sloppy additions but the program I used doesn't work well.

Cheers,

Reg

[GoldscanIV-addicted]

Autotune mod

Hi Reg,
and thanks for the hints. I've noticed that there was something wrong, I mean the two fets connections that perform the reset of the middle stage earlier in other posts also but never seen before the autotune resistors mod. I see that the reset is simply obtained switching the 470nF capacitors to ground and so virtually eliminating any potential on the op amps inputs relative to gnd. The autotune feature (thank you) is really interesting for me : that mean one haven't to manually retune or reset the circuit when a peak occurred in the input and the op amp saturate and the detector will become a non-motion type ? It is right ?

Best regards,
Max

[GoldscanIV-addicted]

ops

I wrote non-motion (I know that original goldscan IV is already non-motion) instead of motion, so the question will be : using autotune, the detector will become a motion type ?

Best regards,
Max

[Reg]

Hi Max,

Yes, the Goldscan will become a motion type detector. Now, if you want a non motion, just add a switch in series with the resistors I show on the schematic so the resistor circuit can be opened. This way, you could have either a motion or a non motion detector.

If the resistor circuit is opened, the detector will drift so the pushbutton labeled switch auto hold will be needed to retune back to the original threshold. Now, this pushbutton (pb) switch is labeled incorrectly. It switches from non motion to a very fast retune and is needed to adjust the various pots both before and after the pb. As an example, the two pots, one for each half of IC 2 are to be adjusted for zero out of the amp. By holding down the fast retune button, external signals are eliminated and the adjustment is much easier.

Reg

[GoldscanIV-addicted]

next step will be sheet number 3

Hi Reg,
this morning I've finished to mount the "sheet 2 of 3" part of the circuit on a protoboard
with soldered joints and ic sockets just to see what happens. Has one can expect from the
schematics considerations I've got two opposite signals at out pins 7 and 14 of IC2 due to the
symmetry in the configuration of the two channels. That's fine. I pushed down the "reset"
labeled pb and see on my scope the two signals convergence, then I just set the two 50K
(IC2 diff integrators offsets, right ?) trimmers to find the overlapping position that
cancels out any residual asymmetry (well not at all but most of I think).

About the "sw auto hold" labeled part (1/4 IC2, switch, 470nF capacitor etc) (please correct if
what I suppose is wrong):
1. the labeled "sw auto hold" seems in the posted schematic a normal switch (I mean 2 stable positions)
but in real (commercial) goldscan IV units it is a push button (?) and serves only for tuning back
to the original threshold in a quick and reliable manner after performing a reset operation

2. it seems to be working like this (I'm not sure anyway :
- when the "sw auto hold" labeled is open then the amplifier provides a strong negative feedback
to stabilize the two sections of diff integrators mantaining outputs levels near gnd and so avoiding
saturation on these stages, doing so with a tank capacitor (470NF) that holds a control
voltage (vgs) limited in range from gnd to (about) -0.3 volt (forward voltage on germanium type diode)
(J113 has a Vgs cutoff of -0.5 volt minimum --> so, after transients, it works in linear zone acting as
a voltage controlled hi value resistor, maybe of some megaohms thus resulting in a non-motion behaviour
,but I think here imbalance can be introduced by temperature related drifts and leaks in some time)

- when the "sw auto hold" labeled is pushed, thus connecting pin 9 of IC2 to gnd, then we have an
offset voltage (negative - schematic say --> 15mv in absolute value but do not indicate the test point -
is it on the middle of the trimmer or at output of the amplifier ? I think is on pin 10) on the output, depending on
the trimmer regulation, that gives a stable vgs polarization of the 2 fets (J113) stored in 470nF capacitor
resulting in more conduction (Rdson typ= 100ohm for a Vgs=0 volt in J113 and some hundred Kohms near pinch-off) and
a discharge path for input capacitors (470NF) --> so we have motion type detection and easy regulation;
when the pb is released negative feedback works again Rds increase in the megaohm(s) level and non-motion behaviour
is then restored.

So I think that the true difference between "reset" and "auto hold" labeled is only in their usefulness :

Maybe reset is used in lab regulation time (gnd balance nulling) and to fast reset huge input charges due to
target overloads instead of waiting that negative feedback do the work (could take many seconds I think).
The "auto hold" seems to me, like you wrote, best choice for accurate regulation of threshold in worse cases.

If it is true (I never used a commercial goldscan before and don't Know) it'll possible to do regulation using
only reset push button, but with less accurancy.

Best regards,
Max

[zzy]

Hi Max,
I've finished GS4.I remove the resistors to get non-motion mode. Because the sensitive is very weak at motion mode.
Now my big problem is the detector will drift.i must push the "reset" button often to get quiet.
I think the problem is the "switch auto hold" can'nt work well.May be the schematic incorrect.Because the "switch auto hold" is use to prevent the amp saturate and can't avoid the drift.
Do you know the voltage on IC1 8 pin and IC1 14 pin? positive or negative? I measure both of them are positive. Am i right? and may be the problem is there.If
Best regards,
zzy

[Max]

Hi Max,
I've finished GS4.I remove the resistors to get non-motion mode. Because the sensitive is very weak at motion mode.
Now my big problem is the detector will drift.i must push the "reset" button often to get quiet.
I think the problem is the "switch auto hold" can'nt work well.May be the schematic incorrect.Because the "switch auto hold" is use to prevent the amp saturate and can't avoid the drift.
Do you know the voltage on IC1 8 pin and IC1 14 pin? positive or negative? I measure both of them are positive. Am i right? and may be the problem is there.If
Best regards,
zzy

Hi,
the 4 and 8 of U1 must stay at GND: gnd is the positive rail that goes to the positive lead of battery.

The drift doens't probably depends of U1... U1 is there just to provide the clock for voltage converter / charge pump.

You have to check Mosfet and damping resistor, often drift depends on one of them or both.

Also, some components could have high temperature coeff. and thus giving a big variation in signals cause of temperature changes.

On my GS4 the drift exist but is very few... I have to retune every 7-10 minutes in hot climate conditions... 12-15 when cold.

The original sold unit, as I know, is about the same : user must retune threshold from time to time when no SAT enabled. This is due expecially to the very high gain there and use of old, drifty components.

If you instead hear the tone change within seconds or in a minute... there must be some problem in your PCB / components.

Also...the Mosfet usually doesn't require an heatsink... cause of the low RDSon and duty-factor but sometimes adding one make things stabilize easy and for longer. You can try.

Kind regards,
Max

[zzy]

Hi,
the 4 and 8 of U1 must stay at GND: gnd is the positive rail that goes to the positive lead of battery.

The drift doens't probably depends of U1... U1 is there just to provide the clock for voltage converter / charge pump.

You have to check Mosfet and damping resistor, often drift depends on one of them or both.

Also, some components could have high temperature coeff. and thus giving a big variation in signals cause of temperature changes.

On my GS4 the drift exist but is very few... I have to retune every 7-10 minutes in hot climate conditions... 12-15 when cold.

The original sold unit, as I know, is about the same : user must retune threshold from time to time when no SAT enabled. This is due expecially to the very high gain there and use of old, drifty components.

If you instead hear the tone change within seconds or in a minute... there must be some problem in your PCB / components.

Also...the Mosfet usually doesn't require an heatsink... cause of the low RDSon and duty-factor but sometimes adding one make things stabilize easy and for longer. You can try.

Kind regards,
Max

Hi Max
Sorry my bad expression.I mean 14 and 8 of "IC1" is on the Reg attached image on this thread.In you schematic it's 14 and 8 of "U10".
Yes,if the temperature changes,my GS4 give a sharp beep.I still think my
"SAT" not work, so i need to know the the voltage on 14 and 8 of "U10".

[Max]

Hi Max
Sorry my bad expression.I mean 14 and 8 of "IC1" is on the Reg attached image on this thread.In you schematic it's 14 and 8 of "U10".
Yes,if the temperature changes,my GS4 give a sharp beep.I still think my
"SAT" not work, so i need to know the the voltage on 14 and 8 of "U10".

Hi,
ok, now I understand. It's something near gnd level , the positive lead of battery level. Small deviations there of few hundreds mV are tolerable.

If you read about the positive voltage you're on the right way.

I don't remember much of GS4 tuning... but I think a good value is around 50-100mV there respect to gnd.

Kind regards,
Max

[Alexismex]

Hello Forum,
Just yesterday i have encounter a small problem that i have years ago
in the power supply of the Goldscan and another power supply (based on 7660).
The fact is that sometimes the +5V are absent when you start on the detector , you switch on/off again and the +5V is now ok ...what Happen??? ...i made lot of different PI power supply, but this fail are embarrassing me for years and yesterday i just encounter:
in the gS power it is C8 the value is 100uf it is not good because too HIGH sometimes this value block the 7805 , if you put a 22 o better a 10uf for C8 you did never have this problem...
and this is good for another power supply the electrolytic capacitor before the input of the 7805 must not be very high...
Have all of you a good time,
Alexis.