Voodoo Main Discussion

[surfdetector]

Anybody can give me some advice?
George, can you post the image that's on pin 6 and 7 of U8?
thanks in advance

P.S.: powered by a 3-30V 10A desk power unit.

I did a series of build videos for the Voodoo detector. You might find them helpful
https://youtube.com/playlist?list=PL...Ny2u3m-SRFjPFy

[Qiaozhi]

I finally got almost all the components and I built the circuit (lacks mosfet tx and ic for filters),
first of all I wanted to test the voltages and the signals generated by the CPU:
these signals are the exact same as those described in the manual, voltage -5V ok but +5V is not present!!!
On tp9 the signal is perfect, on U8 the tensions on pin 8 = 0V (+VB), on pins 3 and 5 = -12V (-VB)
on pin 6 and 7 there is a signal very similar to SYNC but not perfectly square and amplitude around 8-9V starting from -12V
on pin 2 there is no oscillation.
I changed the 78l05 and 4 pieces of Lt1054 it does not change anything.
If I disconnect D6 and/or take off U7 nothing changes (no oscillation on pin2), same thing with a icl7660s.
Anybody can give me some advice?
George, can you post the image that's on pin 6 and 7 of U8?
thanks in advance

P.S.: powered by a 3-30V 10A desk power unit.

U8 pin6 provides a reference voltage of +1.8V relative to -VB.
Therefore there should be a pulsed waveform (SYNC) at U8 pin7 with an amplitude of 1.8V relative to -VB.
At U8 pin2 there should be a pulsed waveform that swings between 0V and -12V (measure relative to 0V).
At cathode of D6 there should be around +8V DC relative to 0V.

Now ... this may be a stupid question, but you do have U8 inserted the correct way round? This IC faces downwards on the PCB, whereas all the other 8-pin devices face upwards.

[andon80]

How is the correct fuse settings for the PIC18F4520?
I am getting an error with TL866A minipro programmer

[surfdetector]

How is the correct fuse settings for the PIC18F4520?
I am getting an error with TL866A minipro programmer[ATTACH]55139[/ATTACH]

I think most of us used a PICkit to load the hex file

[Qiaozhi]

The hex code contains the fuse settings, and a PICkit uses those when programming the PIC.
I've never used a TL866A minipro programmer. What does it require for fuse settings? Is it a file, or do you simply type in some numbers?

[nixie]

U8 pin6 provides a reference voltage of +1.8V relative to -VB.
Therefore there should be a pulsed waveform (SYNC) at U8 pin7 with an amplitude of 1.8V relative to -VB.
At U8 pin2 there should be a pulsed waveform that swings between 0V and -12V (measure relative to 0V).
At cathode of D6 there should be around +8V DC relative to 0V.

Now ... this may be a stupid question, but you do have U8 inserted the correct way round? This IC faces downwards on the PCB, whereas all the other 8-pin devices face upwards.

Finally! I have the solution!
Obviously the circuit was correctly assembled!
I noticed that the signal on pin OSC wasn't right and not wide enough (hence why I asked you to post the images) then when I looked over the lt1054 datasheet. I though about taking off transistor Q6 and resistance R48 (pull-up 20k between Vref and Osc) in order to make the circuit independent as Positive-VoltageDoubler and, magically, I got on pin 2 a perfect 10khz wave and my +5V!!!
I reconnect transistor Q6 and my +5V are still present and on pin Osc there is the wave present on TP9 with an amplitude of 12V and perfectly squared!!! OK! Maybe I had connected something the wrong way...
But then I reconnect Resistance R48 (pull-up) and the circuit stops working!
This makes me think that on pin Osc there already is an internal pull-up resistance not mentioned in the datasheet (in my opinion this is proved by the fact I got the correct square wave without external pull-up) and when we connect the external pull-up this makes the functions of pin Ref go haywire.
Solution:
in my case you just need to disconnect the 20k R48 resistance and the +5V will magically appear.
My LT1054 from Linear Technology have been built the third week of 2019.
I hope this can be useful to you.

[Qiaozhi]

Finally! I have the solution!
Obviously the circuit was correctly assembled!
I noticed that the signal on pin OSC wasn't right and not wide enough (hence why I asked you to post the images) then when I looked over the lt1054 datasheet. I though about taking off transistor Q6 and resistance R48 (pull-up 20k between Vref and Osc) in order to make the circuit independent as Positive-VoltageDoubler and, magically, I got on pin 2 a perfect 10khz wave and my +5V!!!
I reconnect transistor Q6 and my +5V are still present and on pin Osc there is the wave present on TP9 with an amplitude of 12V and perfectly squared!!! OK! Maybe I had connected something the wrong way...
But then I reconnect Resistance R48 (pull-up) and the circuit stops working!
This makes me think that on pin Osc there already is an internal pull-up resistance not mentioned in the datasheet (in my opinion this is proved by the fact I got the correct square wave without external pull-up) and when we connect the external pull-up this makes the functions of pin Ref go haywire.
Solution:
in my case you just need to disconnect the 20k R48 resistance and the +5V will magically appear.
My LT1054 from Linear Technology have been built the third week of 2019.
I hope this can be useful to you.

Look at Fig.12 in the LT1054 datasheet. It shows an external resistor between pins 7 and 6. Also, it states in the text:
"Synchronization can be accomplished by adding an external resistive pull-up from Pin 7 to the reference pin (Pin 6). A 20k pull-up is recommended. An open collector gate or an NPN transistor can then be used to drive the oscillator pin at the external clock frequency as shown in Figure 12. Pulling up Pin 7 to an external voltage is not recommended."

[nixie]

Look at Fig.12 in the LT1054 datasheet. It shows an external resistor between pins 7 and 6. Also, it states in the text:
"Synchronization can be accomplished by adding an external resistive pull-up from Pin 7 to the reference pin (Pin 6). A 20k pull-up is recommended. An open collector gate or an NPN transistor can then be used to drive the oscillator pin at the external clock frequency as shown in Figure 12. Pulling up Pin 7 to an external voltage is not recommended."

Yeah I saw that too and that's exactly what you did right!
There are two things to consider:
1) surely someone built the metaldetector without any problems, which proves the schematic is correct.
2) for some others, me included, it doesn't work: in my case how can you explain the perfect wave on pin Osc without the need of an external pull-up resistor? The only explanation is that a pull-up is already present!
Moreover the output and the pin Osc initially have a 10khz wave 50% then when the micro activates they syncronize and produce a wave form exactly like TP9 but at 12V, this proves the correct operation took place.
With the external resistance the wave on pin Osc is extremely tilted on the ascent and the width is around 9V, without touching neither 0V nor -12V and the output doesn't oscillate.

There may be a need to find an errata corrige of the datasheet or wait for someone whose metaldetector doesn't work and have them take out R48 and then see what happens.

[Qiaozhi]

Yeah I saw that too and that's exactly what you did right!
There are two things to consider:
1) surely someone built the metaldetector without any problems, which proves the schematic is correct.
2) for some others, me included, it doesn't work: in my case how can you explain the perfect wave on pin Osc without the need of an external pull-up resistor? The only explanation is that a pull-up is already present!
Moreover the output and the pin Osc initially have a 10khz wave 50% then when the micro activates they syncronize and produce a wave form exactly like TP9 but at 12V, this proves the correct operation took place.
With the external resistance the wave on pin Osc is extremely tilted on the ascent and the width is around 9V, without touching neither 0V nor -12V and the output doesn't oscillate.

There may be a need to find an errata corrige of the datasheet or wait for someone whose metaldetector doesn't work and have them take out R48 and then see what happens.

As far as I'm aware, no-one else has encountered this problem. Also, I've used this configuration in several designs without any issues.

The 7660 does not require a pull-up resistor when driven by an external clock, which makes me a little suspicious of your LT1054's authenticity. Maybe it's a 7660 disguised as an LT1054.
I'm going to set up a test rig in a prototyping board, and test both an LT1054 and a 7660 to confirm that the former does not have an internal pull-up resistor.

[nixie]

As far as I'm aware, no-one else has encountered this problem. Also, I've used this configuration in several designs without any issues.

The 7660 does not require a pull-up resistor when driven by an external clock, which makes me a little suspicious of your LT1054's authenticity. Maybe it's a 7660 disguised as an LT1054.
I'm going to set up a test rig in a prototyping board, and test both an LT1054 and a 7660 to confirm that the former does not have an internal pull-up resistor.

Obviously I bought it from China!
But even if I had bought it in Europe the seller would have probably bought it from China...
and I belive you are right! It may be a remarked 7660, in case you think it could be useful I can gift one to you!!!

[Qiaozhi]

Obviously I bought it from China!
But even if I had bought it in Europe the seller would have probably bought it from China...
and I belive you are right! It may be a remarked 7660, in case you think it could be useful I can gift one to you!!!

I wired an LT1054 on a prototype board and connected it to an Arduino Uno to supply the sync pulse via a 2N3904 with a pull-up resistor (20k) to Vref.
It worked perfectly as per the datasheet.

Then I went back and read your first post concerning this problem. You stated that removing the transistor and pull-up resistor caused the device to run at its internal clock frequency of 10kHz. This is interesting, because the internal clock frequency of the LT1054 is actually 25kHz, whereas for the 7660 it's 10kHz. Very suspicious...

[nixie]

I wired an LT1054 on a prototype board and connected it to an Arduino Uno to supply the sync pulse via a 2N3904 with a pull-up resistor (20k) to Vref.
It worked perfectly as per the datasheet.

Then I went back and read your first post concerning this problem. You stated that removing the transistor and pull-up resistor caused the device to run at its internal clock frequency of 10kHz. This is interesting, because the internal clock frequency of the LT1054 is actually 25kHz, whereas for the 7660 it's 10kHz. Very suspicious...

Well then it's a 7660 for sure.
Tomorrow I'll check with a microscope and see if I can find traces of writing. The most important thing is for anybody that got my same problem to be able to solve it.

[MartinB]

Well then it's a 7660 for sure.
Tomorrow I'll check with a microscope and see if I can find traces of writing. The most important thing is for anybody that got my same problem to be able to solve it.

Well found nixie.
Whilst we all try and save money where we can, I think the answer is to buy from a reputable supplier.

[MartinB]

I have just ordered 5 nano conversion boards from JLCPCB and expect delivery in about 2 weeks. I will have 4 spare boards. Cost each is 3.66 GBP. If any one is interested in buying one, please send me a PM with your location. I can then calculate postage costs and get back to you to see if you are still interested. Payment will be by PayPal. I am not making making any profit on this - just don't want surplus PCB's hanging around the work shop.

This is what I paid:
Merchandise Total: ￡1.46
Shipping Charge: ￡16.83
Order Total: ￡18.29

[nixie]

Hello George,
could you kindly tell me the electrical characteristics (amplitude and frequency) of the signal generated by the eddy current entering the first stage of the preamp?