Keeping depth without false signals

After giving the Z a run for the past few days in a few different locations across the GT, mostly i have been able to keep the balance nice and smooth.
Today with a little breeze in comparison to previous days, i stuggled to keep the Z quiet.

After pondering about only shallow targets, bullets and rust etc, i decided to dig a hole about 250mm deep and drop a small splitshot sinker i carry for a depth test.
I was running factory default.
I tried Normal for ground, and set Sensitivity to about 3-6.
I could get a signal from the lead sinker well, whereas factory default had no signal at all. Which made me wonder how much i actually walked over if i could not pick up targets at this test depth.

However, when in the modified settings, after a bit of time passes, i found the Z would start finding signals all over the place.
After a scrape, still there. After a dig, target gone.
I was getting a lot of false signals.
At one point i was getting my pin pointer signaling on iron stone rocks by the tens in a single dig hole.
After 4 days of use and looking back, i must have had 40% of targets being false signals, either in factory or modified settings.
Has anyone else found these false signals on ground?
Is there a setting suitable in Difficult that gains depth in the GT?

I don't get any false signals, either in factory or different settings, sometime I find some hot rocks but other than that usually every signal is a target, a lot of pellets and bullets, but targets.

If you're in factory default settings High Yeild/Difficult,with smoothing in the off position,then expect some ground noise and falsing to occur over hot rocks and noisy sections of ground.Hence the low and high smoothing modes.These settings are designed to reduce alot of the false ground noise signals you're experiencing. Yes you do lose a fraction of depth/sensitivity when in low smoothing,more so in high,but you can make up for it by cranking up the gain,whilst maintaining a smoother,stable threshold with minimum falsing.

Cheers Kon.

I found that I had to do a "Reset all" each time I went to a new location. This was the only way I could get rid of the false signals.
It was unworkable on high yield,difficult and low smoothing (GT) until I did the reset all. As you state, if you are in the "normal " setting then I suspect it would be even more unstable and false targets all over the joint.
I am sure frequent users have there own preferred method to settle the Zed but i find resetting all is better than whipping,yelling or swearing at the bloody thing.
cheers Bill

I also agree with Bill on this one. Every new area/location of ground one intends detecting over,it is wise to dump all previously stored detector info,by doing a full "Reset all". That way the detector now can amass as much information over this new type of ground,for maximum detector output,without the memory of any previously stored ground info,kicking in.

Cheers Kon.

I agree, every session it is wise to reset all, and start again.

Just one thing puzzles me.

It has been emphasized that
"Every time the Quick-Trak button is triggered it dumps the oldest GB data in the loop."

In the tutorial video it shows how to perform the ground balance, and then the ability to change Gold Mode and
re-balance [using the quick trak button], thus saving the GB info for both modes.
Which is great for testing targets before the dig.

So the GB info was not "dumped" during that 2nd process?

The questions is, Just how many "loops" are there?

So hypothetically could I Ground Balance in all combination of modes ie 3 x 2, and will it remember them all in that session.

High Yeild x Normal
Gen x Normal
Deep x Normal
High Yeild x Difficult
Gen x Difficult
Deep x Difficult

Here is an example.
First video is a target after srcaping debris away.


Next video is after a bit more soil is removed.


As can be heard from 2nd video the target is gone.

Sometimes this occurs after an small dig, and sometimes after deeper that 250mm target disapears.
Yes, i did also check the removed soil/dirt for targets.

From memory the settings were:
High Yield, Difficult, threshold 24, sensitivity 8.

Each time i go out, i use the noise cancel including a reset.
If the balance seems a bit off (check with ferrite), the i use the noise cancel, but not always the reset option.

Loops,what loops are you referring to Redfin?  The moment a "Reset All" has been performed,all and everything stored in terms of previous ground memory is gone,or should be,or the "Reset All" would play no role nor have any meaning. Once the 7000 has gone through it's first initial Reset All,tuned & ground balanced (to the ground you intend swinging over) procedure,it don't matter what other gold modes you switch over & "quick track" to.All you'll be doing is reground balancing to that particular section of ground with the new gold mode your in.
Lets not confuse Quick Start,with Quick Track. Quick Track only performs a ground balance in all/any of the gold modes one decides to switch over to.It dumps none of the initial "Reset All ground balance/tune procedure". Only by being in "Reset Settings" & pressing reset audio and all current detection settings,is when all previously stored information gets dumped.So in theory,it don't matter what other gold/soil modes you ground balance to,using the quick track button only,does not change,but adds on to the initial reset all ground balance procedure memory. That's what I thought it does anyway's,unless I'v completely misunderstood your question or the 7000s stored memory procedure in general?

Cheers Kon.

I am a tad confused with this as well Kon.

The "Loop" explained by JP.
https://www.minelab.com/aus/treasure-talk/my-method-for-ground-balancing-the-gpz-7000

The GPZ 7000 ground balance algorithm behaves similarly to a loop recording on dashboard video cameras. After a period of time dash cameras record back over the top of the oldest recorded information allowing the operator to pause or stop the video and go back in time to view their recordings. From switch on, the GPZ has to generate material from scratch to store in the loop record (as it only has the last piece of GB data saved from the last detecting session to work with). Hitting the Quick-Trak button too early just dumps the oldest information in the loop recording, which is still not complete. This is counterproductive and reduces performance. The detector’s algorithm refines its GB based on the information stored in the loop record. After 15 to 20 minutes of sweeping the coil it reaches optimum levels for your chosen area.

In a video released to compliment the paper,
https://www.youtube.com/watch?v=5ezuOB1yREQ
he GB's in High Yield / Diff. Then goes on  and GB"s in Gen / Diff.

My question is why on the second GB, didn't the data get dumped?

Is there a 2nd "Loop, or more, or one for each gold mode or combination of settings?

Confusing indeed Redfin. During the first initial ground reading/tuning and balancing procedure (cosidering the detector is in auto ground balance to start off with) the quick track button should not be deployed for the detector is still amassing ground info within that first 15/20 minute period.Deploying the quick track before the time period stated,dumps the ground information the detector has been processing & storing. It is after that initial,so called 20 minute  ground reading period,that the majority of ground info is permanently stored and the detector should be performing at its peak. As stated,pressing the ground track button to early,tends to dump ground info,before the detector has had the chance to reach and store,its full ground reading potential.After that,you can change into any mode and re-ground balance away,without losing any previously stored ground info,until a new "Reset All" procedure is put into effect.

Cheers Kon.

I am finding after about 20mins the Z starts to get a little unbalanced.  When I usually re-noise cancel without reset.

The video is asking to sign in.does anyone else have this problem

In what way coreytroy? Unbalanced in the sense of,the coil needing to be ground balanced,so as to compensate for a rough running noisy coil? or re-noise cancel,meaning a retune for EMI?  There is no need to reset settings once the detector has gone through its initial Reset All and 20 minute ground info gathering.The quick track and retune is more than sufficient for this,unless the detector is constantly running noisy and cannot be ground balanced for sufficient operation. Then a full switch off/on restart will be required.

Kon.

deutran wrote:The video is asking to sign in.does anyone else have this problem

Yep

Coreytroy? None of us can get to see your video without signing in.

Cheers Kon.

Redfin wrote:

Just one thing puzzles me.

The questions is, Just how many "loops" are there?

So hypothetically could I Ground Balance in all combination of modes ie 3 x 2, and will it remember them all in that session.

High Yeild x Normal
Gen x Normal
Deep x Normal
High Yeild x Difficult
Gen x Difficult
Deep x Difficult

I think that is right Redfin, plus High Yield x Severe

At least that's what I recall JP saying somewhere.
After a full reset they are all blank, hence JP switching modes one after an other and doing the GB routine, in each combination he intended using.

Do you have straight links for the vids? My phone is blanking on them

Cheers
Kev.

Hopefully video's fixed to allow playing

alchemist wrote:

At least that's what I recall JP saying somewhere.
After a full reset they are all blank, hence JP switching modes one after an other and doing the GB routine, in each combination he intended using.

Do you have straight links for the vids? My phone is blanking on them

Cheers
Kev.

Kev, Video working fine here.
I've never had one do that to me. Clay perhaps?

That is what I would like confirmation of "They are all blank"
Is there one, two or multiple banks of memory?
I have asked tech support but they have been sitting on the question/email for about 10 days now.

coreytroy wrote:Hopefully video's fixed to allow playing

Hi coreytroy,
I couldn't see them on my phone, but all OK now on the PC thanks.

I'm not in the GT but what Kon and vasilis says seems on the money.

I'm fortunate being able to run Normal a lot, but on red/orange clay areas I will often switch over to difficult to check targets before deciding to dig. I still get fooled since you need to up the gain a bit to compensate, however since you're concentrating on one little spot you can still hear if the signal is there albeit weak. I've not tried it yet, but using Extra Deep to double check a target might help as it apparently has good ground handling characteristics, well at least the software analysing it has, but of course if it's a deep smaller piece you could loose it entirely.

Because you're ground is "spotty" false, and not all hee-haws, I would persist in Normal and check your targets with difficult. Maybe you're already doing this I don't know but conductive salty clay appears to be the Achilles heel of the 7.

Cheers

Redfin wrote:
Kev, Video working fine here.
I've never had one do that to me. Clay perhaps?

That is what I would like confirmation of "They are all blank"
Is there one, two or multiple banks of memory?
I have asked tech support but they have been sitting on the question/email for about 10 days now.

Hi Redfin,
            From what JP said I believe yes, each combination has its own area of memory. JP has a good grasp of the topography of each model, as ambassador I believe he has been informed.
e.g.    
I imagine that in RAM buffers or pages of RAM are allocated to each search combination.
When you start the machine from a reset all the buffers are effectively, Blank.
I suspect that a starting coefficient (no divide by zeros ever allowed) dependent upon the search combination, is read from FLASH memory (permanent) and written to the first block of the buffer memory and used as the seed for the initial calculations. This may explain why it is recommended to pass over the ferrite within seconds of commencing a GB routine. This would quickly modify the coefficient and successive results.
As you perform the 10 or so second GB routine the buffer begins to fill with data and and build a picture of the surrounding ground.
Slowly new data is fed into the Buffer while old data is pushed out the end. First In First Out.
Manual GB setting would hold the buffer contents as they presently are, and Auto would continually poke new data in to reflect current ground conditions.

Swapping search combinations at this point would start the process off all over again, but in a different Buffer, with a different seed coefficient
If you operated the quick track switch in this mode it would flush this buffer only, not the one above, so you could switch back to the previous one and it would still have all the data preserved from where you swapped over.

When you switch the unit off for smoko, the last data results from all the buffers would be written to a special place in FLASH memory, and the RAM Buffer contents would be lost with the power.
When you power on again, but don't do a reset, the last saved data contents are read from the special area in FLASH and written as the seed coefficients at the start of the buffer.
When we do a full reset the default values are read from the permanent area of FLASH and once again used. The special area in FLASH that save intermediate results is probably erased.

I don't know how Minelab actually do it, I am just speculating upon what I've read and how it "appears" things may be operating very very roughly, just to show that it is a trivial matter to save seven buffers of data.

To anyone who sees the whole picture, the opportunities for this process to be corrupted if operating procedures are not followed to the letter will be obvious and explain much.

Cheers
Kev.

A nice theoretical explanation, if it is in fact so, but why won't they tell me, just a simple yes or no will do.

Redfin wrote:A nice theoretical explanation, if it is in fact so, but why won't they tell me, just a simple yes or no will do.

You must have 'the faith' grasshopper, hold on perch eat grasshoppers don't they?

I've noticed that Minelab very rarely describe any 'nuts and bolts' that might give any advantage to any possible competition. I don't know why they are worried, they've got everything clamped under the thumb of Australian biased commercial law and dizzy patent lawyers. At least it's not Indonesia where you'd be secretly thrown to the sharks under the cover of darkness.
I was surprised just how much Bruce has revealed in his white papers!

You could test it out Redfin if you've got a couple of hours to spare. Find some nasty ground and then balance a couple of combinations to it and see what happens after a restart and a reset, of one combination, and see what the other is doing, i.e. how does it handle the hot ground afterwards. I've not got the time at present myself, but I'm keen to know what would happen.

Cheers
Kev.

Best of luck there Kev.

We are leaving in about 3 hours for a week at Dunolly again.
and have the perfect paddock to use this theory on.
If I chain in 200 yard runs I think I should be able to use at least 4 combinations [after I run over it all in deep first].
Will see how it goes.
Weather will be perfect thats for sure.

Looks great Redfin, all the best and I hope you do well.

Don't rush her, take your time, she likes it nice and slow.
I have to restrain myself from rushing here there and tither, not good for 7000s

You might have more insight for coreytroy when you get back.

Cheers
Kev.

Hi all

What I do, no more then this I turn it on I walk for 15minutes then I do a tune and balance and that's it till I feel it become unbalance again then do it all over again that it .

Detrackozi wrote:Hi all

What I do, no more then this I turn it on I walk for 15minutes then I do a tune and balance and that's it till I feel it become unbalance again then do it all over again that it .

I basically did the same too

Like others on here, I have major trouble with false signals and I doubt the recent repairs to my 7000 will fix that. But as for the initial GB for 15 seconds and then "rebalance as it drifts off". I was doing that every 5 minutes!!

Hopefully, things will improve after these repairs.

I must once again thank the tech repair team at ML, Kevin and Darryl for a very quick turnaround. Great service as usual.

Being not the only one with issues, and GB being an item that must be constantly udjusted, i am thinking whether there is really any major benefit to Auto-GB.

The GPX series i know are a diffferent machine, but it was a simple GB to put it back into rythm.

Is there really any benefit of storing all this GB data for 15-20mins only to have it dumped or give off bad balances anyhow.

Next outing i may just start as norm, and change to manual GB to see if the Z can be kept free of less false signals.
Your thoughts?

It does seem a little odd to me, this idea of collecting GB info over a 15 minute period to obtain a so called "optimum". I reckon your strategy of using "manual" GB could be a winner, Consider,,,,,,,,,,if "manual" is used you may lose a small amount of "optimised" GB but you could be rewarded with a much quieter machine that isn't chasing its own tail and ending up "winding itself" up.

my recent trip out was my first main "get to know you" with the z. so naturally i first headed to where i had found a 109 grammer down about 18 inches with the extreme way back in 2001. the extreme scored well here with more than 6 and a half ounces to its credit. the 4000 didnt improve on that score, (i passed on the 3000 and 3500) and the 4500 and 5000 did find more, mainly a 79gram piece down about 15 inches along with some 30 ish gram pieces down in gravel beds. in all this there were smaller pieces closer to the surface to be had. so yes it had been a very good spot, so good, a commercial miner has now ripped the guts out of the area. it was very interesting to see the hidden gravel beds that he exposed and probably got paid for his efforts.
just outside his lease area, where i had found a 55gram piece down deep, was my first port of call. running the first day in deep, difficult (couldnt get normal to work quiet enough) with a high sensitivy all just produced a few ground noises to check out. nothing. my hopes were dashed, just like probably many of you who have done the same with the new machine.

i gave the area a miss for a couple of days, got to know my new machine a bit better and couldnt resist going back for another lash at my "skite site". this time i was able to run in general normal with smoothing low, dug a lot more ground noises until they became obviously just that, and was disappointed again.

before leaving with my chin draging on the ground, i headed back to a dig hole of "ground noise" which disappeared more than a foot down, was it a target i lost? had it been in the wall, and dropped into the hole? i had to know. initially back at the filled in hole, there was no signal, started excavation of soil, test for signal, nothing, kept this up until hole was now completely redug. nothing. i decided to poke the coil into the hole VERTICALLY and got a wave motion continuous sound, like bad EMI. i then noise cancelled the z ON THE GROUND,and still running in auto GB, did the 12 to15 second jig over the ferrite, the machine was now perfectly tuned. this was the first time i noise cancelled on the ground, like we used to do with the GPX machines at clermont qld, where the emi and electrical interference is miminal, so you tune out the soil and not the air. it worked, no more extra noises, but unfortunately a few more hours did not turn up any nuggets for me.

all a waste of time to make me richer, but i can now let go of that area - DUN AND BUSTED except for my good memories... kev