Aporte: Estacionamiento asistido (parking aid) NO ultrasonido

[rascueso]

Hola seaarg te cuento que tengo la placa lista de tu proyecto solo me falta programar el pic y montarlo. Pero hoy me encuentro con una duda referida a otro tema y opte por preguntarte a vos.
El tema es el siguiente arme un circuito que de 8 lugares diferentes pueden accionar un pulsador y lu ego en una central marca mediante un led cual fue el pulsador que se disparo la persona ve que pulsador se disparo y puede setear el sistema. Lo que no logro lograr es encontrar un circuito que haga una melodía agradable para acoplársela a este circuito cuando alguien acciona uno o varios de los pulsadores
Usando el pic 16f628a y el código de este proyecto puedo lograrlo? Mi idea seria que haga un bip bip bip pausa bip bip bip.
Saludos ras

 

[seaarg]

Si te referis al de estacionamiento el mismo solo hace bips de onda cuadrada a una F determinada.

Si queres algo mejor que eso, hay en este foro (no se bien donde) un programa para grabar un archivo wav en la rom del pic y reproducirlo. Eso si, come mucha memoria.

 

[rascueso]

grax seaarg voy a seguir buscando. esto de no saber programar me limita mucho
Abrazo!

 

[seaarg]

Hola seaarg te cuento que tengo la placa lista de tu proyecto solo me falta programar el pic y montarlo.

Tengo hecho un prototipo de la version 3, mucho pero por mucho mas facil y estable. El problema es que ando falto de tiempo para ajustarla y publicarla aqui. Contame que resultados obtenes con la que estas haciendo.

 

[rascueso]

ok rey. estos dias las termino. tenia problemas para programar el 16f628 pero el amigo juanka ya me explico como conectarlo al programador asi que pronto lo termino. espero la fase 3 de tu proyecto

 

[FFT]

Hello @seaarg,

I've seen your topic accidentally on internet and I found it so interesting.

I just want to make a park sensor for my old car as hobby. I already understand a bit from electronic. I know C and PIC.

I don't understand Spanish, I use translator to read the site.

I've read that u r making the 3th version of the circuit. I also can search for you about XTall oscillators.

I wanted to know about the making of the sensor practically and some tricks about the circuit.

1- Could you explain how to make the capacitive sensor?
2- When you think to publish the 3th version of the circuit?

I hope I'll get an answer.
Sincerely

 

[seaarg]

Hi FFT.

Sadly on this forum we're not allowed to write in english. Thanks for your comments. I hope moderators will allow this message. We cannot provide email addresses in the forum. I'll reply in english and spanish and perhaps we can continue the talk as this.

Lamentablemente en este foro no se nos permite comunicarnos en ingles. Gracias por tus comentarios. Respondere en ingles y en español.
------------------------------------------------------------------------------------------
1)- The sensor is made from two aluminium stripes with an aislator as dielectric in between them. This is a simple capacitor with no polarity. When something aproaches the sensor, it slightly changes it's capacity. The stripes are about 3-4 cm height and goes trough the full lenght of your car's bumper. The stripes are separated with the aislator by about 4-5 mm making a capacitor with a capacitance of about 40pF (not very important, depending on the crystal used). The bumper has to be non-metal, i think (mine is glass fiber).
--------------------------------------------------------------------------------------------
1)- El sensor esta hecho de 2 tiras de aluminio con un aislante entre medio que actua como dielectrico. Esto es un capacitor simple sin polaridad. Cuando algo se aproxima al sensor cambia levemente su capacidad. Las tiras son de aproximadamente 3-4 cm de altura y van a traves de todo el parachoques. Estan separadas por un aislador de aproximadamente 4-5 mm y la capacitancia es de aproximadamente 40pF (no muy importante, depende el cristal que se use). El parachoques no puede ser metalico, creo.
------------------------------------------------------------------------------------------
If you assemble almost any crystal based oscillator, you can replace one of its capacitors (the one at the output of the NOT gate) with this "sensor" and the oscillation frecuency will change (or sweep) a little bit being affected by the proximity of an object. This way you can build a circuit with a microcontroller that samples that frecuency and analyze changes on it to estimate the distance to object. Valid range is from about 70cm to 1cm from the capacitive sensor.
------------------------------------------------------------------------------------------
Si armas casi cualquier oscilador basado en cristal, puedes reemplazar uno de sus capacitores (el que esta a la salida del inversor) con este sensor y la frecuencia de oscilacion va a cambiar de acuerdo a la proximidad de un objeto. Podes armar un circuito con un microcontrolador que samplee la frecuencia y analice cambios en la misma para estimar la distancia a un objeto. El rango valido es de aprox 70cm a 1cm.
------------------------------------------------------------------------------------------
This method is not exact as ultrasonic, in the means that you cannot measure an exact distance. Google for EP parking sensor, or electromagnetic parking sensor. There's a commercial development on England wich i tried to immitate.

I say crystal oscillator since its more predictable than other oscillators. You can build this around a schmidtt trigger NOT gate such as CD40106 and a 4mhz crystal. Then use a frequencimeter to make some tests.
------------------------------------------------------------------------------------------
2)- My project is on pause for a while now. The 3rd version is based on a crystal oscillator just as i described. I made it work but not finished software nor make any schematics yet. Dont have the time to finish it and publish it now.
------------------------------------------------------------------------------------------
2)- Mi proyecto esta en pausa hace un rato. La 3ra version esta basada en un oscilador a cristal tal cual como describi. Lo hice funcionar pero no termine el software ni hice esquematicos aun. No he tenido tiempo de terminarlo y publicarlo.
------------------------------------------------------------------------------------------

If you just want a park sensor for your car and doesnt matter how it's done, you can google for several good ultrasonic circuits. If you're interested in research other methods, then this is the post for you

PS: You can google about capaciflector patent from NASA. It uses similar, but obviously more advanced techniques.

Regards

 

[FFT]

Hi again,

Thanks for your replay. I can imagine now the idea.
I want to research and make a park sensor on my own, otherwise I can buy one, they are cheap.

1- You said than "cannot measure exact distance" with this method, so how you calibrate it?

As far as I understood is you made an oscillator (whatever type u want) which contains a capacitor and you make the capacitor using the aluminium folios. Frequency of the oscillator changes by changing the value of the hand made capacitor ("sensor"). Your microcontroller side is just a frequency meter which does some works by changing of the frequency value. Is that correct?

If so, how do you define the base frequency value which oscillator generates in normal conditions?
How do you define the other limit frequencies for estimated distances? - In other words, how do you understand when the car is very near to a wall/car etc.. ?

2- Does it detect everything? For example children, people, vehicles, kerbs, pillars, walls, fences, barriers, skips etc..

We are waiting your 3th version someday
Sincerely.

 

[seaarg]

1- You said than "cannot measure exact distance" with this method, so how you calibrate it?

This method has a base frequency when nothing is on it's field range. (Let's say 4mhz) Then calibration is via software. The pic stores the nearest object frequency value.

As far as I understood is you made an oscillator (whatever type u want) which contains a capacitor and you make the capacitor using the aluminium folios. Frequency of the oscillator changes by changing the value of the hand made capacitor ("sensor"). Your microcontroller side is just a frequency meter which does some works by changing of the frequency value. Is that correct?

That's exactly right. I found that the best oscillator is a crystal one. RC oscillators are unstable.

If so, how do you define the base frequency value which oscillator generates in normal conditions?
How do you define the other limit frequencies for estimated distances? - In other words, how do you understand when the car is very near to a wall/car etc.. ?

When you turn on the circuit, it assumes that nothing it's on the way, so that's the base frequency for greatest distance. Then you have a frequency value stored on eeprom that corresponds to the nearest distance and you make "divisions" to change the beeper tone. (This value has to be saved on eeprom on first run) Other software solution is to take the base frecuency and change this "divisions" according to frequency changes.

I was surprised on my experiments. If i power up with an object in the "near" field it reaches the nearest distance frequency in about always the same place.

2- Does it detect everything? For example children, people, vehicles, kerbs, pillars, walls, fences, barriers, skips etc..

We are waiting your 3th version someday
Sincerely.

Hell yeah! that's what i like about this technology. On my tests it detected everything. Of course massive objects affects the capacitance in a greater way but i found that the "near to crash" limit is triggered on almost any object.

Attached is the base oscillator i used. I replaced the 7404 with a CD40106 (worked for me), removed both resistors and used a 4mhz xtal with a 47pF trimmer on the input (replacing the 22pf cap) and the sensor replacing the 22pF cap on the output.

If you have access to a cap-meter to measure your sensor, then you can adjust those values best.

I hope this clarifies things for you.

En español: No traduzco porque la mayoria de lo que digo ya lo he dicho a lo largo del post

 

[FFT]

Short but very important tips. Thanks so much, I hope u help not only to me, but also to the other people who cannot speak Spanish Also I think that every person who is in electronic have to understand English, no way.

This project can be made for cap meter and also you indirectly made this

I absolutely understood all, except the "divisions" part.

When the circuit is worked for first time, it measures the freq and stores the value in eeprom, well. But how does it detect the distance levels ( level1 = so far, level2 = near, level3 = very near, level4 = crash etc..). Maybe I can understand this by reading your C code.

If you use the same mechanism in the second version, I'm gonna start to read that one. if not could you publish your unfinished code? Just for understanding the definitions of distance levels.

Best wishes

 

[seaarg]

Attached is the C code for my latest attempt. Still buggy an undocumented code.

As sample frecuency decreases (approaching to an object) the delay between "beeps" is decreased too, so beeps sounds more often.

calib_value variable is the first sample when you turn on the circuit.

This is just an approach and it's intended to auto-calibrate. As you can see, nothing establish fixed distances here, just audible approximations to an object.

Another aproach is to save the nearest object frequency value on eeprom (with a device first run programming routine) and then do this math:

first_sample value: sampled when you turn on the circuit, then:

div_value = (first_sample - eeprom_min_value) / no_of_distance_divisions_you_want

That way, you can assume about:

first_sample = 1000 (just example) corresponds to 70cm.
eeprom_min_value = 700 corresponds to 15cm.

Then you do the math to estimate distance to target on each obtanied sample.

Notice than in my 3rd version i wasnt using this approach. My idea was to set an audible beep that increases it's rate (or tone) as object approaches. Then when object crosses some "about to crash" frecuency barrier, it changes to a constant beep (not implemented on the C program)

You can set several frequency barriers and that way you can have several tone pattern changes or something that's useful to you.

The only thing i'll ask you is to publish your researches and tests here, this way your progress will be useful to the forum & me

I plan to continue this project until i get something fully functional but i had to finish several other projects too as time permits.

This is what i managed to accomplish so far (of course this video is a commercial product)

I just need to stabilize it, debug it, etc.

And this:

Is my own project. Distance was small on that opportunity (about 50cm) but worked. Later i achieved greater distances.

You can see the "sensor" here attached to the table.

------------------------------------------------------------------------------------------

Adjunto el codigo C de mi ultimo intento. Todavia con bugs y codigo no documentado.

A medida que la frecuencia sampleada decrementa (al acercarse a un objeto) la demora entre "beeps" tambien decrementa, entonces los beep suenan mas seguido.

La variable calib_value es el primer sample cuando se enciende el circuito.

Esto es solo una aproximacion y se intenta que auto-calibre. Como se puede ver, nada establece distancias fijas aqui, solamente aproximaciones audibles a un objeto.

Otro punto de vista seria guardar la frecuencia para objeto mas cercano en eeprom (con una rutina de programacion en la primera ejecucion) y luego hacer lo siguiente:

div_value = (first_sample - eeprom_min_value) / no_of_distance_divisions_you_want

De esta forma, se podria asumir algo como:

first_sample = 1000 (ejemplo) corresponde a 70cm.
eeprom_min_value = 700 corresponde a 15cm.

Luego se hace la cuenta para estimar la distancia al objetivo para cada sample obtenido.

 

[FFT]

Thanks for the detailed explanations.

I made the oscillator on my breadboard, but the 1M ohm Rf resistor is mandatory to use, otherwise the circuit doesn't work.
I used 22pf caps (which I have) and I get 4MHZ output from 4011 IC.

But my sensor-making experiment has failed.
I used 4cm height, 1.2 meters length aluminium strips with 1mm plastic band as dielectric. And I connected it using short crocodile cables to the oscillator circuit.

When I connect it to the circuit, the output became 3.966MHZ, when I disconnect it the output is 4MHZ
Also tried 2 parallel 22pF (44pF) caps on one side of XT.
I didn't see any frequency changing on the output when I'm near to the sensor or not. But when I touch the sensor, it is 800Hz

I think maybe the problem is in the diameter of the dielectric, or in the type of the dielectric I use.

What do you think about this?

What is the (base) frequency when your sensor is connected?

EDIT:
I also now tried to remove the crystal from the working circuit and saw that my multimeter measures 4.2MHZ. I've put there 12MHZ crystal and the output is 4.6MHZ. Put there 3.7..KHZ crystal and the output is 3.5MHZ. Put there a 26MHZ crystal, the output is 4.8MHZ.
Is this normal?
I'm using nand gate of 4011 as NOT gate and metal crystals.

Best regards.

 

[seaarg]

I made the oscillator on my breadboard, but the 1M ohm Rf resistor is mandatory to use, otherwise the circuit doesn't work.
I used 22pf caps (which I have) and I get 4MHZ output from 4011 IC.

First of all, im glad you're trying.

you said "caps" not "cap". You have to replace the inverter output cap to ground entirely with the sensor, otherwise the cap will stabilize the frequency. (our goal is to unstabilize it)

But my sensor-making experiment has failed.
I used 4cm height, 1.2 meters length aluminium strips with 1mm plastic band as dielectric. And I connected it using short crocodile cables to the oscillator circuit.

Try using a 4mm dielectric. Mine is assembled with this: http://www.papeleradelmar.com.ar/ap.../product/1d77fbc1b0ce1bcbbfa5db205840e7da.jpg

I dont know the english word for that material (perhaps Styrofoam??)

When I connect it to the circuit, the output became 3.966MHZ, when I disconnect it the output is 4MHZ
Also tried 2 parallel 22pF (44pF) caps on one side of XT.
I didn't see any frequency changing on the output when I'm near to the sensor or not. But when I touch the sensor, it is 800Hz

Do you have a cap-meter? the input side cap should be with capacitance as close as possible with your sensor. Also the crystal to use is related with those values.

For instance, with a 20mhz crystal should be about 10pF, with a 4Mhz one, 47pF (dont know the formula... those are values used on the pic crystal)

It's normal the fact that when you touch the sensor the frequency changes a lot.

I think maybe the problem is in the diameter of the dielectric, or in the type of the dielectric I use.

What do you think about this?

What is the (base) frequency when your sensor is connected?

I'd look that way, try different dielectric sizes. My base frequency is about 3.9mhz (not exactly but it's been a while since i tried last time) with a 4mhz crystal. Also tried 20mhz crystal to have more pulses to count on a 100ms sample but it was more stable at 4mhz.

Also, check datasheet for the gate you're using, perhaps doesnt work on those frequencies. (40106 isnt supposed to work at 4mhz but it works!... weird)

EDIT:
I also now tried to remove the crystal from the working circuit and saw that my multimeter measures 4.2MHZ. I've put there 12MHZ crystal and the output is 4.6MHZ. Put there 3.7..KHZ crystal and the output is 3.5MHZ. Put there a 26MHZ crystal, the output is 4.8MHZ.
Is this normal?
I'm using nand gate of 4011 as NOT gate and metal crystals.

Best regards.

This is normal based on my experiments. In order to get the frequency of the crystal, the caps should be of certain values. I wish i knew the formula for that.

What happens when you remove the crystal is that the gate still oscillates, this is normal. The crystal is there to only stabilize the frequency. (as far as i know)

My practical formula is: As crystal is bigger, caps should be smaller. Try to measure your "sensor" cap. If it's around 10pF use 20mhz, 47pF, use 4mhz and so on.

Try power it with 9-12v instead of 5v (you'll have to adapt the output signal to a pic later)

Also, when you post messages, please paste a spanish google translation under it. Otherwise, moderators will sent our messages to the trash.

------------------------------------------------------------------------------

dijiste "caps" y no "cap". tienes que reemplazar el cap de salida del inversor a tierra en su totalidad por el sensor, de lo contrario el capacitor estabiliza la frecuencia. (Nuestra meta es desestabilizar)

Trate de usar un dieléctrico de 4 mm. El mío es montado con esto: http://www.papeleradelmar.com.ar/ap.../product/1d77fbc1b0ce1bcbbfa5db205840e7da.jpg

No sé la palabra Inglés para ese material (tal vez de espuma de poliestireno?)

¿Tiene un cap-meter? el capacitor del lado de la entrada debe ser con capacidad lo más cerca posible con su sensor. También el cristal que use está relacionado con esos valores.

Por ejemplo, con un cristal de 20 MHz debe ser de 10pF, con un 4 Mhz, 47pF (no sé la fórmula ... esos son los valores utilizados en el cristal de los pic)

Es normal que el hecho de que al tocar el sensor de la frecuencia cambia mucho.

Lo veria de esa manera, tratar de diferentes tamaños dieléctrico. Mi frecuencia base es de 3.9mhz (no exactamente, pero ha sido un tiempo desde que intentó por última vez) con un cristal de 4MHz. También trató de cristal de 20 MHz a tener más impulsos de contar con una muestra de 100 ms, pero fue más estable a 4MHz.

Además, verifique el datasheet de la compuerta que está utilizando, tal vez no funciona en las frecuencias. (40106 no se supone que funciona a 4MHz pero funciona! ... Raro)
Esto es normal, sobre la base de mis experimentos. Con el fin de obtener la frecuencia del cristal, los capacitores deben ser de ciertos valores. Me gustaría saber la fórmula para eso.

lo que sucede cuando se quita el cristal es que la nand sigue oscilando, esto es normal. El cristal es sólo para estabilizar la frecuencia. (Hasta donde yo sé)

Mi práctica es: como el cristal es más grande, los caps deben ser más pequeños. Trate de medir su "sensor". Si es en torno a 10pF, 20MHz , 47pF, use 4 MHz y así sucesivamente.

pruebe con 9-12V en vez de 5V (que tendrá que adaptar la señal de salida para un pic más adelante)

 

[FFT]

Hi,

I've increased the distance between plates and I think the cap value is decreasing in such way.

I use this scheme below exactly but I'm changing all values by trying...


The most successfully try was with
2.2MΩ resistor,
5.5pF output capacitor (4x22pF in series),
12Mhz crystal at 12V DC
(assembled on a breadboard)

My base frequency was ~24Mhz. When my hand was 1 cm near to the capacitive "sensor" I see 37Mhz on output. The freq is changing on 0-15 cm distance of my hand. This is the best result for me.

I think this cannot be a trusted circuit for my car.
Also I don't have a capmeter

I need to find the best configuration of the component values with the sensor or remake the sensor again...

1- Could you measure and give me the exact dimensions of your cap "sensor" which works well with 4Mhz crystal ?

2- How is changing the freq value with values of capacitors in input and output of the NOT gate? Shall they be equal for the best stable freq? Or how would be occur when I change the input cap? output cap?

I saw that sometimes freq is increasing by coming close to sensor, but sometimes (with different configuration of xt cap values) is decreasing...

Sometimes the freq is (lets say) 20Mhz when I'm the center of the room, it's 25Mhz when I'm so far from the sensor and it's 27Mhz when I'm with 1 cm distance with the sensor. How could you explain this?

Best wishes

----------------------------------------------------------------------

Hola,

He aumentado la distancia entre las placas y creo que el valor límite está disminuyendo de esa forma.

Yo uso este esquema de abajo exactamente, pero voy a cambiar todos los valores al tratar ...
[url]http://img220.imageshack.us/img220/4739/xtallosc.png[/url] [/ IMG]

La prueba con mayor éxito fue con
2.2MΩ resistencia,
Condensador de salida 5.5pF (4x22pF en serie),
12Mhz de cristal a 12 V DC
(Montado en una placa)

Mi frecuencia base se ~ 24MHz. Cuando mi mano fue de 1 cm cerca de la capacitiva "sensor" Veo 37Mhz en la salida. La frecuencia es el cambio en la distancia 0-15 cm de mi mano. Este es el mejor resultado para mí.

Creo que esto no puede ser un circuito de confianza para mi coche.
Además no tengo una capmeter: (

Tengo que encontrar la mejor configuración de los valores de los componentes con el sensor o rehacer el sensor de nuevo ...

1 - ¿Podría medida y me dan las dimensiones exactas de la tapa de "sensor" que funciona bien con cristal de 4 Mhz?

2 - ¿Cómo está cambiando el valor de frecuencia con los valores de los condensadores de entrada y salida de la puerta no? Deberán ser iguales a la mejor frecuencia estable? O cómo se producen cuando cambio la tapa de entrada? salida de la PAC?

Vi que a veces frecuencia va en aumento al llegar cerca del sensor, pero a veces (con diferente configuración de los valores de x, la tapa) está disminuyendo ...

A veces la frecuencia es (digamos) 20 Mhz cuando soy el centro de la sala, es 25Mhz cuando estoy tan lejos de los sensores y es 27Mhz cuando estoy con la distancia de 1 cm con el sensor. ¿Cómo podrías explicar esto?

Los mejores deseos

 

[seaarg]

You're having the same problems i had on my first experiments.

Im guessing your frecuency meter is capable to measure such high frecuencies, have you check that?

1)- Sadly no, the sensor is already installed on my car (from my last tests) and i dont have the time to dissasembly. The sizes are about the same i explained earlier. I'd suggest you to rebuild the sensor. 15 cm max distance is not good enough. With thin sensors this happened to me, that's why im suggesting to use a 4mm dielectric. (and yes, it should decrease capacitante but cannot be sure with home made caps.)
2)- The caps should be as close in values as possible. Of course (this is important) C2 in your circuit is NOT a capacitor, it's the sensor. I'd suggest use a 47pF trimmer for C1 (i think it's the blue, or green one) C1 & C2 (the sensor) should have a value as close as possible to make this circuit stable. If you dont have a cap-meter, then use a trimmer and adjust (with plastic screwdriver) until freq seems stable. If not, you gonna have "jumps" in each sampled frequency.

"I saw that sometimes freq is increasing by coming close to sensor, but sometimes (with different configuration of xt cap values) is decreasing..."

Right! when you approach to the sensor, it's capacitance should increase, so frequency should decrease.
As you can see on this program line: if (value < calib_value)
The best behavior is obtained this way, so adjust values to match this behavior.

You can try increase R1 (in some tests i used 10mohms, 2.2mohms)

"Sometimes the freq is (lets say) 20Mhz when I'm the center of the room, it's 25Mhz when I'm so far from the sensor and it's 27Mhz when I'm with 1 cm distance with the sensor. How could you explain this?"

Unestability is my best shot. On my last test if i went out the detection range the frequency is somewhat stable (changing a few hertz on each sample) if i standing still.

Remeber, this is an experimental project, perhaps someone with more advanced electronics knowledge could help us to stabilize it.

If you can, try replace the nand gate with a 40106 inverter. It worked for me. (i tried another, faster ttl inverter and was unstable. Dont know why)

Remember, try different sensors, i think that's the key.

I hope this helps a little and you can achieve your expected results.
-------------------------------------------------------------------------------------------
Estas teniendo los mismos problemas que yo tenía en mis primeros experimentos.

1) - Lamentablemente no, el sensor está instalado en mi coche (desde mi última prueba) y no tengo el tiempo para desarmarlo. Los tamaños son del mismo he explicado antes. Yo le sugiero a la reconstrucción del sensor. 15 cm máximo de distancia no es suficiente. Con sensores delgados me pasó esto, es por eso que im sugiere utilizar un dieléctrico de 4 mm. (Y sí, se debe disminuir capacitante, pero no podemos estar seguros con tapas caseras.)
2) - Las tapas deben estar tan cerca de los valores como sea posible. Por supuesto (esto es importante) C2 en el circuito no es un condensador, que es el sensor. Te sugiero usar una podadora 47pF para C1 (creo que es el azul o verde) C1 y C2 (el sensor) debe tener un valor lo más cercano posible para que este circuito estable. Si usted no tiene una tapa de metros, a continuación, utilizar un condensador de ajuste y ajuste (con un destornillador de plástico) hasta frecuencia parece estable. Si no, vas a tener "saltos" en cada frecuencia de muestreo.

"He visto que a veces frecuencia va en aumento al llegar cerca del sensor, pero a veces (con diferente configuración de los valores de x, la tapa) está disminuyendo ..."

¡Muy bien! cuando usted se acerca al sensor, que es capacidad debería aumentar, por lo que la frecuencia debe disminuir.
Como se puede ver en esta línea de programa: if (valor <calib_value)
El mejor comportamiento se obtiene de esta manera, por lo que ajustar los valores para que coincida con este comportamiento.

Usted puede tratar de aumentar la R1 (en algunas pruebas he utilizado 10mohms, 2.2mohms)

"A veces la frecuencia es (digamos) 20 Mhz cuando soy el centro de la sala, es 25Mhz cuando estoy tan lejos de los sensores y es 27Mhz cuando estoy con la distancia de 1 cm con el sensor. ¿Cómo podría explicar esto? "

Unestability es mi mejor tiro. En mi última prueba, si salía del rango de detección de la frecuencia es más o menos estable (cambio de algunos hertzios en cada muestra) si sigue en pie.

Recuerde que este es un proyecto experimental, tal vez alguien con más conocimientos de electrónica avanzada nos puede ayudar a estabilizarlo.

Si puede, trate de sustituir a la puerta NAND con un inversor de 40106. Funcionó para mí. (He intentado otra más rápida ttl inversor y era inestable. No sé por qué)

Recuerde, trate de diferentes sensores, creo que esa es la clave.

Espero que esto ayude un poco y puedas lograr los resultados esperados.

 

[FFT]

I've used the C1 as the cap sensor and the C2 was 5.5 - 44pF capacitor on circuit.

I changed their places and now I get the frequency of the crystal which I put there! Tried 4, 12, 16, 20Mhz crystals and all outputs are stable. Changed the normal capacitors but no chance. Crystal makes it so stable. I understood that my circuit generates (primarily) 4MHZ all the time. I'll try with another chip and if not another sensor I'm gonna make.

BTW: My sensor is 5 cm wide, 1.2 meters long and 6 mm high

Are you sure that your sensor is connected on the output of NOT gate (C2 in my scheme) ?

BR.
----------------------------------------------------------

He utilizado el C1 como el sensor de la tapa y el C2 fue de 5,5 - 44pF capacitor en el circuito.
Me cambiaron de lugar y ahora tengo la frecuencia del cristal que me puso en su lugar! Trató de 4, 12, 16, 20 Mhz cristales y todas las salidas se mantienen estables. Cambiar los condensadores normales, pero no posibilidades. Cristal hace que sea tan estable. Comprendí que mi circuito genera (sobre todo) 4MHZ todo el tiempo. Voy a intentar con otro chip y si no es otro sensor que voy a hacer.

¿Estás seguro de que el sensor se conecta en la salida de la puerta NO (C2 en mi sistema)?

 

[Meta]

No estamos en Inglaterra, ejjeejjeejej Y eso que es un foro internacional.

 

[seaarg]

Yes, the sensor is connected where C2 is. There's no much more i can say but to keep trying. Sensor measures should be ok.
----------------------------------------------------------------------------------------------
Si, el sensor va conectado donde tenes C2. No hay mucho mas que te pueda decir mas que seguir probando. Las medidas del sensor que declaras deberian estar bien.

@Meta: Por eso mi pedido de que ponga el post aunque sea traducido por google al castellano y lo esta cumpliendo. Y por eso es que me tomo la molestia de poner mis posts en los 2 lenguajes. La verdad es que se merece toda la ayuda que se pueda, ya que lo esta intentando. Cierto?

 

[FFT]

Hi again,

I've tried with 2 different sensors too but no chance...

First one was 5cm height, 1.3meters length and 4-5mm dielectric distance..
Second one was 2.5cm height, 2.3meters length and 0.5mm dielectric distance.

I've made a simple capacitor meter using 16F628A. Using the principle of RC charge time. I'm charging the capacitor (using series resistor 2.2Mohm or 4.4 or 6.6 or 8.8Mohm) and when the voltage on the capacitor becomes 63.2% of the VCC (5V) stopping the timer.
Discharging the capacitor through a NPN transistor.
Using this method I can measure correctly 10uF, 10nF, 1nF, 22pF etc real capacitors. That means my circuit works properly and if there are any changes on the capacitance of sensor, my circuit can understand this.

But no chance.. Only when I'm 2 cm near to the sensor then I can see some changings...

I hope the oscillator method is the same thing like this, cause I'm measuring directly the capacitance (thru charge time).

I also tried to use a average filter to stabilize the values.

Isn't there anyone who can explain more details for this sensor?
OR
Isn't there anyone other who tried to make this project?

Sincerely

--------------------------------------------------------------------------

Hola de nuevo,

He probado con dos sensores diferentes también, pero ... hay posibilidad

Primero fue de 5 cm de altura, longitud 1.3meters y 4-5mm de distancia dieléctrica ..
Segundo fue de 2,5 cm de altura, longitud y la distancia dieléctrica 2.3meters 0,5 mm.

He hecho un medidor de capacitores sencillos utilizando 16F628A. Usando el principio de tiempo de carga RC. Yo estoy cobrando el condensador (con 2.2Mohm serie resistencia o 4.4 o 6.6 o 8.8Mohm) y cuando la tensión en el condensador se convierte en un 63,2% de la VCC (5V) detener el temporizador.
Descarga del condensador a través de un transistor NPN.
El uso de este método que puede medir correctamente 10uF, 10nF, 1nF, 22pF condensadores, etc. Eso significa que mi circuito funciona correctamente y si hay algún cambio en la capacidad del sensor, mi circuito puede entender esto.

Sin embargo, ninguna posibilidad .. Sólo cuando estoy cerca de 2 cm en el sensor entonces puedo ver algunos changings ...

Espero que el método del oscilador es la misma cosa como esta, porque estoy medir directamente la capacidad (a través del tiempo de carga).

También traté de usar un filtro de media para estabilizar los valores.

¿No hay nadie que pueda explicar más detalles de este sensor?
O
¿No hay nadie más que trató de hacer este proyecto?

sinceramente

 

[seaarg]

Hi,

Bad thing you cannot make it work. It's a hard project (i spend a few months before i can make the first working prototype).

Read the following:
http://www.electro-tech-online.com/...33485-parking-car-electromagnetic-sensor.html

http://bdml.stanford.edu/DML/previous_projects/capaciflector/capaciflector.html

Those were my first reads about it. After some prototypes i end up with a crystal oscillator schematics.

I build the attached LCF meter to measure my sensor. The C meter should be able to measure pF scales in order to measure the sensor. (as you said, only at 2cm away you can get results... probably that's because capacitance increases dramatically that near).

A few days ago i readed about another crystal oscillator with variable frequency. Attached is the sample schematics i got. Perhaps we can set a fixed L and a variable C in series with the crystal. (not tested, just a tought).
---------------------------------------------------------------------------------
Hola,

Que pena que no puedas hacerlo funciona. Es un proyecto complicado (pase unos meses antes de que pueda hacer el primer prototipo funcional)

Lee lo siguiente:
http://www.electro-tech-online.com/...33485-parking-car-electromagnetic-sensor.html

http://bdml.stanford.edu/DML/previous_projects/capaciflector/capaciflector.html

Esas fueron mis primeras lecturas acerca del tema. Despues de algunos prototipos termine haciendo un oscilador a cristal.

Construi el medidor LCF que adjunto para medir mi sensor. El medidor de capacitores debe poder medir escalas de pF para poder medir el sensor. (como tu dices, cuando estas a 2cm del sensor obtenes resultados... probablemente es porque la capacitancia se incrementa dramaticamente estando tan cerca)

Algunos dias atras lei sobre otro oscilador a cristal con frecuencia variable. Adjunto un esquematico de ejemplo que consegui. Tal vez se pueda dejar fija L y poner variable C en serie con el cristal. (no lo probe, es solo una idea)