TL12000

[Asanka Sanjeewa]

Hi Detectorist#1,

probably Asanka means the 555 timer circuit that feeds the LM7809 voltage regulator circuit part. There is a nonsense connection from the TX oscillator to pin 4 of this 555 timer. This connection is not present in the original Sandbanks PI or PI Polones. See the red arrow below. Pin 4 should be connected to pin 8 of the 555 timer IC, just as it is shown in the original Sandbanks PI and PI Polones schematic.

Dear Detectorist#
I agree, you are right, connecting the 1st 555 ic output frequency to the 2nd 555 reset terminal is wrong. I am not saying that everything is right in this circuit. Its ability to reject iron is much more success than other pulse induction devices.Sorry, I changed it to sound oscillator. Instead of LM 331, I used the 555 oscillator used in POLONESE device. This increased the ability to soundly express the changes made by the sample delay unit.

[Asanka Sanjeewa]

Hi GeoMax,
It is not easy to understand what Asanka means in his posts but we try to help him in his efforts.

Dear Detectorist#
What I mean is, from the devices I've tested in practice, the Delta pulse really has the ability to see deep into large metal parts. But I am trying to improve tl 12000 because of metal separation capability.The barracuda and felezjoo devices have the same separation ability, but from my field tests, it seems that they are not suitable for the mountain rock environment with high mineral content. But tl 12000 was success in those places.
My target is to have a device capable of rejecting iron with high penetration strength in a highly mineralized rock environment. But that doesn't mean that TL 12000 is the best device. But it is the simplest option for me now.

[ArchibaldSTM]

I am not saying that everything is right in this circuit.

Seriously? When you LIED, then or now?

[GeoMax]

Seriously? When you LIED, then or now?

Hello Archibald,

don't say he's lying... Asanka noticed that something was wrong with the power supply circuit, but maybe he just didn't notice the stupid connection on the 555 timer before. In addition to this error, there are other strange connections in the circuit. Just look for example at the connection of the damping resistor. It's actually very strange too. But remember, the TL12000 somehow still works even with these errors...

[GeoMax]

Hi all,

Last year, just for fun, I drew a schematic of the TL12000 from the TL12000 PCB posted by Asanka. Below is the schematic diagram as per Asanka's modified PCB. Remember, this is a one-to-one copy that still includes all the weird connections made on this particular PCB! I copied the original layout into the schematic for recognition so everyone knows we are discussing this PCB and not some other similar looking PCB. If anyone notices a mismatch between the PCB and the schematic, let me know. I can update and re-upload the schematic anytime. As for the coil shield connection that was not shown on the PCB, I am not sure if it is used in Asanka's design. I have not seen a reference to it so far. But the detector coil should have a coil shield/connection to the circuit. That is why I included this.

The scheme is now “open” for discussion about what can be improved.

[Detectorist#1]

Hi GeoMax,

Yes, TL12000 still works even with errors. The names of the voltages in the schematic are a little confusing. If they are measured against "GND", the voltage "-15V" have to be marked as "+3V". The voltage "+9V" have to be marked as "+27V". Voltage "Vbat (+18V)" is the real "analog ground".

[GeoMax]

Hi GeoMax,

Yes, TL12000 still works even with errors. The names of the voltages in the schematic are a little confusing. If they are measured against "GND", the voltage "-15V" have to be marked as "+3V". The voltage "+9V" have to be marked as "+27V". Voltage "Vbat (+18V)" is the real "analog ground".

Hello Detectorist#1,
yes, the voltage names are a little confusing. But I label them according to the circuit board. Always remember that I drew almost everything exactly as it is on the circuit board and that also includes all the errors! And there are still a lot of circuit parts that are not OK. That was also one of the reasons why I drew the circuit from the PCB. Only with the help of a circuit diagram can you see all the design errors.

[Detectorist#1]

Hi GeoMax,

I agree, your schematic and the names of the voltages corresponds to these of the PCB. This is maybe more convenient for the analysis.

[ArchibaldSTM]

don't say he's lying....

You and I have different concepts of *tested and 100% working circuit and board*... From my point of view, this implies a faultlessly executed circuit and board. And not that *well, it works!*.

[GeoMax]

Hi ArchibaldSTM,

I don't know how long you've been testing detector electronics... I've only seen a handful of nearly flawless circuits myself in the last 25++ years. In almost every circuit I've analyzed, people claim their circuits are 100% flawless. So I never believe when someone says "my circuit is 100% flawless" until I've checked and proven it myself. You should do the same, because it helps you not to be disappointed when a circuit doesn't work as expected. Take the Detector MiniPuls or Baraccuda for example. Both circuits worked well. But it wasn't until Geotech Members took a closer look and came up with a redesign with no design flaws that the circuit was even better than its originals. This can happen with the TL12000 too if everyone who is interested gets involved calls out all their design problems and then fixes them.

[GeoMax]

Hi ArchibaldSTM,

the question is where to start improving the TL12000 circuit.

I would start with the following circuit parts first.

1) Redesign the power supply circuit and relabel the power pins.
2) Redesign the front-end-receiver using a widely available NE5534 op amp.
3) Throw away the first half of the LM358 (buffer circuit) because the LF398 already has an integrated buffer circuit.
4) Threshold circuit. Use a more proven design for that.

Don't be surprised that after all these changes you have almost turned a TL12000 back into a Sandbanks PI aliases PI Polones detector. Because the TL12000 is based on the Sandbanks PI from 1979. The main difference is still the use of the LF398 sample & hold IC. But that's it (only)! The original designer of the TL12000 uses the Sandbanks PI as the basic circuit and just replaces the switching FET (in front of the integrator) with the LF398 and the CD4011 with the CD4013. I don't know why he changed the clock circuit too. Maybe he did that to hide the basic circuit. He could have also replaced the CD4011 with a CD4538 or CD40106. But that makes absolutely no difference in the function of the circuit. Then he made some other fun changes to the circuit and used a real VCO IC (LM311) to improve the sound. In the end, the TL12000 was finished.

[Asanka Sanjeewa]

Seriously? When you LIED, then or now?

Please understand my friend what i tell . The 4th terminal [ reset pin ] of the second 555 ic is wrongly connected. That's right. But I looked at the output of 555 IC. It shows exactly 6.5kHz.This is correct value [ frequency ] This output is what supplies the voltage regulator.I finally checked the 7809 regulator ic output. It also shows the correct voltage, which means that even if the terminal is wrong, it is working correctly. I did not try to increase the work because it was working properly.I wanted to solve the two obvious grant problems. The hard to find LM 331 voltage to frequency converter ic was replaced by the cheap 555 sound oscillator. BC327 was put instead of the BC337transistor used in other youtube videos.No matter how much sample delay I adjusted until this part was made, I could not reject the iron. He was able to make changes better than others. Please send someone's email address in this forum and I will upload the video. If you look at it, you will be able to understand.I admit that there are mistakes here, I said that it draws too much current [it eats battery]. Because the BU 508 transistor gets a little hot, we need to put HEAT ZINK. Even if the battery voltage is a little lower, the sensitivity will decrease a lot. That's why I said that. Don't misunderstand.But the circuit works very well. I put here a working circuit that I tried and changed. I don't want to waste my time and scam more people by sending mails. I wanted to share my experience with others. Please understand that.Please tell me a way to send the video to an email address or a member's YouTube account. Can't upload mp4 in this forum.

[GeoMax]

Hi Asanka,

if you can't find a LM331 VCO IC, just use a CD4046 IC. The CD4046 is actually a PLL IC, but also has a nice VCO part. Of course, you need to further modify the TL12000. The CD4046 is a DIP16 IC, cheap and widely available. I have already seen some detector circuit modifications using only the VCO part of the CD4046 IC chip. Alternatively, you could use a LM566 VCO IC. But this is also quite hard to find and expensive. The VCO made from a 555 timer is really one of the cheapest options. Even Eric Foster, the metal detector guru from the UK, has used a 555 timer VCO circuit in his excellent detectors. So did Carl in his Hammerhead PI detector. There is actually nothing wrong with using a 555 timer IC as a VCO. It is easy and cost effective.

[GeoMax]

Hi Asanka,

you wrote... ".... I said that it draws too much current [it eats battery]. Because the BU 508 transistor gets a little hot, we need to put HEAT ZINK. Even if the battery voltage is a little lower, the sensitivity will decrease a lot....."

One of the reasons why your BU508A gets hot is the TX oscillator's pulse width settings. This circuit part has no adjustment capabilities for it. That's why the transistor gets hot and the circuit drains the batteries quickly.

[Detectorist#1]

Hi GeoMax,

All these problems in current design are true. For me, the strangest think for this design is that Asanka insists that capability for Fe/non Fe separation is more good than in most simple PI projects. Where is the reason for this or this is only his impression?