Although special-purpose frequency-to-voltage chips help you to create an F-to-V circuit block easily, they suffer from a major drawback. For given external components, their gain constant, and hence voltage output, can vary more than ±5% from piece to piece (e.g, the LM331 and the LM2917). Their currentsource value also varies piece to piece by a similar amount, thus causing piece-to-piece variations in the maximum voltage-output beyond which F-to-V goes into saturation.
This creates problems for mass-produced solutions requiring F-to-V conversion if the application can’t tolerate these variations. The result is messy trimming procedures for external passive timing components for each mass-produced piece.
This is where the old workhorse LM555 outshines these special-purpose ICs. The 555 generates a mono-shot time-period equation:
TD (mono-shot delay) = 1.1 × RC
Timing accuracy is typically 1%. (For National Semiconductor’s LM555 and LM55C, it’s 0.5% and 1%, respectively.) Thus, designers can use timing components with precision tolerance to meet an application’s requirements, without unduly worrying about piece-to-piece chip variations.
A mono-shot-based F-to-V works by converting incoming pulses to a fixed-pulse-width, variable-frequency pulse-width-modulation (PWM) signal. The duration of the pulse is fixed at the period of the mono-shot (1.1RC in this case). This duration also forms the on period or duty-cycle of the PWM output.
Lower-frequency signals yield lower-percentage duty cycles, since the period forms a lesser percentage of the total period of the incoming pulse train. As the frequency rises, the output’s duty cycle increases, since the “on-period” is now a greater percentage of the incoming pulse-train period (Fig. 1).
This dc value of the output PWM signal (given by percentage duty cycle) is directly proportional to the incoming pulse-train frequency. However, the situation gets a bit tricky when the incoming pulse train has a period shorter than the on-period. The F-to-V is now supposed to saturate to its maximum value for all of the frequencies with time periods less than the mono-shot on-period.
However, if the mono-shot is not re-triggerable, it will expire “asynchronously” with the incoming pulse triggers, causing random output variations. As a result, the example circuit uses a transistorized discharge-switch to allow the 555 to re-trigger, which it otherwise could not do (Fig. 2).
Transistor Q1 forms a discharge path around capacitor C2. Resistor R4 and capacitor C2 are the timing components of the mono-shot with a time period of about 0.39 ms. Hence for a frequency of about 1 kHz, the output duty cycle is 39%, giving a voltage of about 4.68 V for a VBATT of 12 V. C1, R1, D1, R2, and R3 form the base trigger circuit for Q1. The 555’s internal transistor at pin-7 is not unused in this circuit. Rather than leaving the pin un-terminated, R5 and C3 terminate it as recommended by some CMOS 555 chip vendors. It can also be used as an auxiliary source of a sawtooth wave if required by the application. R6 and C5 form an output low-pass filter to extract dc voltage from the PWM pulse-train appearing at the IC’s pin 3.
The circuit’s output voltage increases linearly at the rate 4.68 V/kHz up to 2.5 kHz (Fig. 3). The output then saturates to its maximum value as desired (as a consequence of the re-triggerable feature of this mono-shot circuit).
Thanks again Bob, I have a project that requires a reasonably low ripple at encoder frequencies under 2 khz and the 555 circuit did not make the grade. We are creating a stationary virtual travel device using analog joystick inputs for wheelchair propulsion, YouTube "trekease". I have gone over your notes again and will build my prototype circuits next week. I am a little confused about the differences between the LM331 and the LM2917
Christopher Stanford -September 12, 2009 (Article Rating: )
I'v just finished projects that used both F-to-V and 555 timer circuits. Please keep in mind that the output of this is not a DC value, it's a pulse-width modulated waveform. Figure 3 is completely misleading. To get a DC voltage out, you's need a whoppingly slow lowpass filter, or op amp/ diode circuitry to charge a holding capacitor (which will have a slow discharge problem to deal with). Don't call this a low-cost F-to-V convertor without including the rest of the components to make the "V."
Randy Butturini - September 08, 2009 -September 08, 2009
The Inventor claims that this circuit avoids the drawback of the 5% gain tolerance of the LM331. If you like that, fine, build this circuit. But the LM331 V-to-F Converter, which I designed 30 years ago, has 0.06% per volt max gain regulation, and it avoids the +/- 10% gain error when the "12-V battery" changes + 10% or -10% as it charges or discharges. The LM331 has excellent guaranteed linearity, which the 555 circuit does not. So I just want to make sure I vote AGAINST this circuit, No. 21612. And Jim Williams does, too. / rap
Bob Pease -September 04, 2009
Your Comments:
Enter the text from the image below
Please refresh the page if you have trouble reading this text.
Search Electronic Design
Email Newsletter
Sponsored By:
The Find Power Products monthly newsletter brings you the most important new developments within the world of power design. The newsletter includes exerpts from industry leader Sam Davis's exclusive blog, as well as overviews of the latest new products.
Enter Email to Subscribe
Web Seminar
Sponsored By:
Title: Exploring How Good GUIs Drive Adoption in the Digital Power Management Space
Reprints
)
