PICKINI V4 - an easy to build, self adjusting PI detector

[Goaty]

Some success today - found the 470R chip resistor R2 had gone open circuit and replaced it. Now I can get it to work by adjusting the sensitivity pot after the auto-tune has completed its procedure but the range is very limited and still unstable. The smd BC848's for Q3 arrived today but fitting one made no difference. Tomorrow I will check some waveforms with the scope and then most likely swap out the LM318. Still waiting for the pack of chip resistors to arrive. For now I plan to check out the New Years Eve fireworks. HAPPY NEW YEAR TO ALL!

[Goaty]

Hi Bernard
Thanks once again for your assistance.


I have followed your suggestions and made some voltage measurements with the following results-


The base voltage of Q3 starts off around 0V-0.5v depending on residual charge in the caps and quickly climbs to 1.5V – 1.65v with a couple of minor fluctuations when power is applied. It settles at either of these two voltages at the end of the auto-tune period.


Pin 3 of the LM318 settles at either 8.4v or 8.9v


The auto-tune period takes around 10secs.


@dbanner


“

Assuming you have the genuine LM318 (fake chips are always in the mix nowadays), Maybe replace C1 and C12 with other capacitors, those caps might not be best quality.

“



I heartily agree with your comments and plan to pull these components as a last resort. I have been putting this off as, being SMD devices, they stand a good chance of being cooked in the removal process. Using components from China adds an extra degree of difficulty to building projects but being 70km from the nearest component retailer or adding $7au postage to the cost of even small local purchases often has to be taken into account

[dbanner]

Assuming you have the genuine LM318 (fake chips are always in the mix nowadays), Maybe replace C1 and C12 with other capacitors, those caps might not be best quality.

[BC547]

The DC offset of the opamp is adjusted by the circuit around Q3 and controlled by the DAC output = pin 13 of the controller.
The output of the opamp is divided by 2 and fed to pin 8 and 9 of the controller. Make sure that these 2 pins are shorted.
One is the ADC input that is used for this auto DC adjustment after startup: the "steady state" DC level is measured after the flyback pulse and the DAC output is adjusted accordingly.
The other input is the comparator input that is used to measure the pulse width.
If seems that the automatic DC offset is not working in your case. This probably results in a relatively long adjustment phase after startup and ends with some extreme value that results in a too high DC offset. If you have a DC voltmeter, you can measure the DAC output that performs successive approximation after power on - pin 13 of the controller or R22 = 10 Ohm to the base of Q3.
After setup, the collector voltage of Q3, going to the noninverting input of the opamp (pin 3) should be around 8V.

[Goaty]

Still waiting for components but in the meantime I’ve been playing around with the SMD Pickini board over the last couple of weeks (when time permits) and have still been unable to locate the cause of the instability with the auto tuning.
No amount of adjusting the sensitivity pot seems to be able to stabilise it. I have tried several coils which work OK with the original Pickini board, tested it outside away from most interference and checked all over the board with the multi-meter but to no avail.

Back in post #812 I listed my voltage readings against what they should be but when I uploaded the table it lost the column formatting and is not very clear to read – mine are the RH figures of each pair.
It appears that the output of the LM318 is too high giving 4.6v on pins 8 & 9 of the pic instead of the original 3.7v. I feel sure this is closely related to the problem but can’t see why or how.
While I think I am OK with the basic principles of this circuit, I am still struggling to come to grips with the finer details (Old dogs and new tricks!) so any suggestions would be much appreciated.

[Goaty]

Hi Bernard


Thanks for your reply.
I have checked the polarity of the electros and they all seem to be correct but it is a possibility that one or more could be leaky especially since I could not find a stable spot for the sensitivity pot when I powered it up today. I was intending to put the CRO on it but gave up when I could not stop it bleeping and carrying on!
BTW, I did not change any of the settings so it is not likely to be that.



In regard to using Mint, it’s good to hear you are a convert too. Virtual box seems to be a good option for the few occasions that Win is needed and I find that I am slowly getting more accustomed to using command lines in Terminal in Linux and am starting to enjoy that aspect of it as well as using the GUI. As a matter of fact, the reason for my tardiness in replying to you is that I got caught up trying to install an UPS monitoring program in Linux for my newly acquired UPS. The program which came with the new UPS is quite good except that its instructions were rather poor, having lost a lot of detail in the translation from Chinese. It took most of the day and several calls to the supplier to get it up and running but I got there in the end.


Anyway, in regard to the SMD Pickini project, I will get back with more info once my missing components arrive.


Regards
Goaty

[BC547]

Hi Goaty,

The DC values you measured look relatively normal.
It would be very interesting to see the flyback pulse on a scope at the input of the comparator (pin 8 and of the controller).

I have posted my voltages below and, as you can see, some of them are way off. Little wonder the auto tune is struggling. I think I programmed the pic with vers 4.2 but at the time of doing it I was more concerned with getting the program into it as I had not used the programmer for a long time and this was the first time using win10 on a virtual machine in a Linux computer and having to concentrate on getting the USB to work! (After many hours of work my head was full!)

I also use virtualbox with ubuntu, running Win7 to control the Pickit3 programmer. You can select the USB devices at the bottom menu: machine / devices /USB ( or something like that, don't have it in front of me right now...). There you select the programmer. You need to do this each time you restart the virtual machine. I haven't found a way yet to save this setting in virtualbox.

With the build I have v4.2 running here, the sensitivity range is quite large, not from 0 to 5V, but something like 1V to 4V at least. When you go too high and then return until you get the normal signal again, that is where you have the highest sensitivity. Doing this indoors may lead to a lot of chatter. When you turn down the sensitivity this noise will decrease and possibly disappear, but your sensitivity will be a lot less.

I suppose you did not change anything to the project settings - like enabling the MCLR = reset functionality or anything ?
Another thing to double check is the polarity of the electrolytic capacitors. Especially the signal coupling C9 - but also C1 and C12 are critical.

If you say the coils worked on other Pickinis, they should also work on an SMD version.

Do you have a scope available ?

- Bernard

[Goaty]

If you programmed the 4.2 hex file, the comparator slicing level is influenced by the sensitivity setting. It is normal that you can adjust it out of the range of the analog flyback pulse, resulting in the error tone. Make sure the potmeter voltage goes from 0 to 5v and not the other way round. Turn it up until the normal tone turns into the error tone and then down again until you get detection tones again. This would be the best sensitivity setting.
So, everything appears to be normal.

Hi Bernard
The sensitivity pot range seems OK but the point at which the detector behaves normally is very narrow (if at all some times!). I am using a variety of coils which have all worked successfully on other pickinis
I have posted my voltages below and, as you can see, some of them are way off. Little wonder the auto tune is struggling. I think I programmed the pic with vers 4.2 but at the time of doing it I was more concerned with getting the program into it as I had not used the programmer for a long time and this was the first time using win10 on a virtual machine in a Linux computer and having to concentrate on getting the USB to work! (After many hours of work my head was full!)






U1 - PIC16F1824 SMD Vers
PIN 1 : 5.0V 5.0V
PIN 2 : 2.7V 3.6V
PIN 3 : 1.2V - sensitivity potmeter 2.1V
PIN 4 : 5.0V 4.99V
PIN 5 : 4.7V 4.75V
PIN 6 : 0.1V 0.01V
PIN 7 : 0.0V 0.00V
PIN 8 : 3.7V 4.6V
PIN 9 : 3.7V 4.6V
PIN 10 : 2.7V 3.7V
PIN 11 : 0 ..5.0V – audio -
PIN 12 : 2.8V 2.87V
PIN 13 : 2.0V 1.52V
PIN 14 : 0.0V 0.00V

IC4 - LM318 opamp
PIN 1 : 9.5V 0.01V
PIN 2 : 6.8V 9.00V
PIN 3 : 6.8V 8.98V
PIN 4 : 0.0V 0.00V
PIN 5 : 4.8V 0.01V
PIN 6 : 7.4V 8.99V
PIN 7 : 11.9V 12.35V
PIN 8 : 1.4V 0.00V


PS Sorry the table did not turn out too well - I cut and pasted from another doc.



@Marty – Re board making, I may have mentioned before that I use a dry negative resist film which works quite well. I make my transparencies by printing onto thin wrapping type tissue paper with a laser printer then oiling it with cooking oil to make it translucent. I have a primitive light box (cardboard carton) with strips of UV leds for exposure and use a washing soda (Sodium carbonate) solution as a developer so no nasty chemicals. So the gear is pretty basic but it does a reasonable job.

[MartyJ1963]

Hiya Goaty,,,,Nice to hear you made a workng unit,,,,,I know what you mean about the ''Impatience'' game,,been there, done that,,Lol. Hopefully when your new SMD parts arrive you'll have had the chance to work out what's what with the present board and then etch a new pcb and be able to apply your new found results to that one knowing everything will be perfect at first 'switch-on',,,I love these little pickinis,,I've made quite a few now and there all still going great (Thanks again Bernard!),,,I'm still yet to etch my first SMD Pickini though but I'm busy at present with tuning some DD Coils and swapping parts out of my Mirage's so I can experiment with them. Good luck with your results and keep us posted,,,Regards,,,Marty

[BC547]

If you programmed the 4.2 hex file, the comparator slicing level is influenced by the sensitivity setting. It is normal that you can adjust it out of the range of the analog flyback pulse, resulting in the error tone. Make sure the potmeter voltage goes from 0 to 5v and not the other way round. Turn it up until the normal tone turns into the error tone and then down again until you get detection tones again. This would be the best sensitivity setting.
So, everything appears to be normal.

[Goaty]

@Goaty,,,Any progress on the Pickini4 SMD yet? Regards,,,Marty

Hi Marty
I’m afraid “ I have some good news and some bad news” as the saying goes.


The good news is that it works and the bad news is that I have a problem with mine.


The story is that my impatience got the better of me when it came to waiting for parts to arrive from overseas so I cobbled up some shortcuts to get it going. These involved stacking smd resistors to get the required values in a couple of places and tacking in place a th BC548 instead of the sm version.


When I applied power I was relieved to get the startup tones and the auto-tune ticking as this meant that the programming of the soic pic using my home-brew socket had been successful. But there the joy ended as I was greeted by the fault tone.


I eventually found that I could get past this by tweeking the sensitivity pot and it came to life and responded to metal waved near the coil but if I powered it down then up again it would not work until I again played with the sensitivity pot.


So, at this stage I plan to go over the board with the multi-meter and CRO and collect some info and also wait for the parts to come.


In the meantime I have been reading through all of the posts in this topic to glean as much info as possible on how the whole thing works. I have an idea that it may be to do with the comparator slicing level mentioned by Bernard in posts #739 and #552 so I will work on this for a start.


I will post more info when I get it

[MartyJ1963]

@Goaty,,,Any progress on the Pickini4 SMD yet? Regards,,,Marty

[6666]

Thanks F117, much appreciated, you are very clever .

[BC547]

The PIC 16F1824 has 2 analog comparators onboard. They work exactly as any other comparator but have an input range of 0 - 5V.
And of course a lot of settings in SW: hysteresis, output inversion, output internal/external, clock sync, ...
1 of them is used to measure the pulse width of the flyback pulse.
The attenuated ( /2 - to match the 0 - 5V input range of the comparator) pulse is connected to the negative input of the comparator = pin 9
The positive input = pin 10 is connected to a DC slicing level coming from the PWM output pin 2 + RC filter.
The DC level of the flyback pulse is measured with an ADC in steady state, and the slicing level is set 0.5V lower.
This results in a positive output pulse = flyback pulse width at slicing level.
This pulse width is measured by gating a timer running at 32 MHz with it ( resolution = 32 ns).

You can hack the schematic by connecting an external voltage to the positive input pin = pin 10 with a potmeter, eg.
To do this, you need to disconnect R19 = series resistor between PWM output and comparator input.

This manual adjustment of the slicing level can be an advantage in very noisy environments.
This is also what v4.2 of the software does: adjusting the sensitivity setting will increase/decrease the comparator slicing level - this is done internally by modifying the PWM output.

http://users.telenet.be/willaert/MD/pickini/pickini_en/V4/software.html

[6666]

Hello F117
I have been experimenting with some analogue comparator circuits and how I can use them to detect a signal in a PI detector,
so I am just wondering how you configure the comparator inputs in your PIC.

Do you use the comparator inputs as inverting mode or non inverting mode, or do you do something different as I am not sure how a PIC comparator works compared to an analogue circuit
thanks.