High frequency to voltage converter

[eam]

Hello everyone<p>I am working on a circuit to convert a high frequency signal (800khz, 2V, AC) to a low voltage signal (10V, DC). Does anybody konw how to build it or where I can buy that kind of things?<p>
Thanks a lot

[Engineer1138]

When you say "convert" what do you mean:<p>a) It's a power source, and you need to level-shift it?
b) It's a signal and you need to output 10VDC when you see it?
c) None of the above?<p>a: use a high frequency transformer and a rectifier & filter
b: use a PLL or other frequency detector circuit
c: well, you get the idea.<p>In any case, you'll need to offer more information before we can really give a good solution.

[dyarker]

Or even 1 to 10V depending on frequency?????????

[Robert Reed]

One diode , one resitor, one capacitor. Arrange as an RF detector used in test equipment probes. Feed a Transistor Ampl, with this and pick off the DC you desire. Cant give exact values without knowing more parameters, but for starters a 1N914 seriesed into a1000 PF to ground then feed transistor base thru a 100K resistor from diodes cold end.A 10K cllector load to a 10VDC supply and you should have what you're looking for. Dont know what polaity you need . May have to invert?

[haklesup]

Sounds like you want an integrator followed by an amplifier.<p>The integrator is an op amp with a cap in the FB loop. The ouptut is the area under the input waveform. Choose a non zero reference voltage or your periodic waveform will integrate to 0<p>THe second stage is an amplifier with gain appropriate to scale the integrator output to the voltage and current you desire.
Select op amps with appropriate bandwith and supply voltage and output current to meet your minimum needs.<p>Google on op amps, integrator circuit, differentiator circuit and amplifier or read some OP-Amp app notes and data sheets for specimin circuits to emulate.

[eam]

Sorry about my dimness, I neeed 1 to 10V depending on frequency

[Enzo]

SO you genuinely need a freq to voltage converter. Your FM radio has one. But that doesn't help much.

[dyarker]

You didn't mention range, but try this for a starter idea:<p>The period of 800KHz is 1.25uS. Half of that is 0.625uS.<p>For every cycle of input generate a 0.625uS fixed width pulse of 10V. At exactly 800KHz in, the average voltage of the pulse stream will be 5V.<p>As input frequency goes below 800KHz, the pulses will be farther apart, the average voltage of the pulse stream goes down proportional to the difference from 800KHz. Reverse happens above 800KHz.<p>Advantage is wide input range, high end almost 1.6MHz. Disadvantages are component values/tolerances for a one-shot to create the pulse width you want, thermal drift of pulse width, getting exactly 0V to 10V pulse amplitude, and messy integration and filtering.<p>-------------------
Or:<p>Divide a crystal oscillator to 800KHz and feed it and the input frequency to a phased locked loop. The control voltage for the voltage controlled oscillator in the PLL is also the output level.<p>Advantage is it is mostly digital (except VCO), accuracy and thermal drift depend (mostly) on the crystal oscillator (only parts per million error). Disadvanages are narrow range above and below 800KHz, the VCO control voltage will only vary by one or two volts for the frequency range. Probably about 3V at 800KHz with 5V Vcc logic. To get 1V to 10V you'll need at least one op-amp for gain and to add/subtract an offset. (Or a microprocessor with ADC and do the scaling in firmware.)<p>If 800KHz is too fast for the PLL, divide the input frequency by 8 (for example), and crystal to 100KHz before feeding the PLL. The output will not respond to changes in input frequency as quickly, but that can be a good thing.<p>-------------------
Added or:<p>Amplify and schmit-trigger the input frequency, and feed it to crystal clocked microprocessor. Do everything in firmware with the built in timer registers. Output of DAC 0V to 5V, externally amplified by 2; and/or serial or parallel binary.<p>Advantages are, minimal analog circuitry for accuracy and stability, range between the previous two methods, fewer components so a smaller circuit board. Disadvantage is learning to program a microprocessor, and downloading the program to the microprocessor. But that is the way things are going to, so might as well start somewhere.<p>Cheers,<p>[ May 13, 2005: Message edited by: Dale Y ]</p>

[ezpcb]

a F-V convertor from analog or TI or Maxim will work.

[Ron H]

Dale, what kind of PLL has two inputs?

[dyarker]

Ooops, none. The phase comparator section has two inputs. The VCO section isn't used.<p>Now that RonH caused me to rethink, the PLL would be too hard to do too. The 800KHz input wouldn't neccessarilly be in phase with the crystal oscillator even if the frequencies are exactly the same. The phase comparator depends on phase difference, not frequency difference.<p>While constant frequency difference will cause comparator output related to the frequency, it would be unstable with very small differences.<p>Sorry 'bout that.<p>That leaves fixed pulse width integration, firmware in a microcontroller, or a specialized F-V IC as ezpcb.com suggested.<p>Later,