Hello All,

I wanted to start a new post on this topic even though I have seen a number of threads about this. I'd like to see if I could get some specifics on what I would need to do for a little project I have.

The research topic that I'm currently working on in mechanical engineering (that's why I don't know much about electronics) is requiring me to have a piece of circuitry that will give me a DC voltage output that is proportional to the primary oscillation frequency of an incoming signal. Here are the specifics I mentioned earlier. I am working with a high frequency actuator that pulses between 4 and 6 kHz. I would also like to have a frequency resolution of about 5 Hz or so. I have a sensor inside of the actuator that is monitoring the unsteady pressure fluctuations (the actuator is fluidic-based). I need to send this signal in to the circuitry and have a DC voltage come out. I'm using a narrow band pass filter the clean up the signal from my transducer, and I can calculate the frequency easily using my oscilloscope or data acquisition hardware. My issue is that I'm trying to record my actuator's frequency on-the-fly, if you will (or at least as close to instantaneously as I can get).

I know that there are a number of circuit diagrams that use the 555 timer to output a pulse width modulated signal and that you have to use an integrator circuit to have that converted to a DC voltage. I am not really sure what all components I need for this as well as the specifics - does the frequency range I can measure depend on the resistors, capacitors, etc? I have wired up a couple of basic circuits in my day, and I do know how to solder. Being an engineer, I ought to be able to figure out how to wire these up from some circuit diagrams. My problem is that I just don't know all of the components I need and where I can get them (if my local Radio Shack doesn't have any).

Can someone help me out here?

Thanks,
Phil

 

There are a number of devices that directly convert frequency to voltage, such as the LM2907 or LM2917. I suggest you use one of those. Google "frequency to voltage converter" for a number of references.

 

What about LM331? Any experience with it in comparison with LM2907? When used for converting < 100Hz to voltage. Thanks!

 

I am working with a high frequency actuator that pulses between 4 and 6 kHz. I would also like to have a frequency resolution of about 5 Hz or so.

If you are looking to resolve to ±5Hz in a 2kHz frequency spread that wont be a trivial task. I doubt you could devise a relatively simple 555 timer pwm + integrator based solution that would achieve your stated measurement goal.

 

I used the XR-4151 in my lab frequency generator to create an analog voltage for the digital display. It is very accurate if you use best quality resitors and caps.

http://www.datasheetcatalog.com/datasheets_pdf/X/R/-/4/XR-4151.shtml

 

I used the XR-4151 in my lab frequency generator to create an analog voltage for the digital display. It is very accurate if you use best quality resitors and caps.

http://www.datasheetcatalog.com/datasheets_pdf/X/R/-/4/XR-4151.shtml

Why is that every time I see an XR part I immediately think it is an obsolete one?

They had incredibly good and imaginative ICs in their catalog.

 

I am working with a high frequency actuator that pulses between 4 and 6 kHz. I would also like to have a frequency resolution of about 5 Hz or so.

The problem I see here is what are you gonna do with the output signal once you have it. If you want to use it in a feedback loop, the delay between change of frequency and change of output might be a problem. Also the resolution you want will be hard to achieve.

Maybe a better solution would be something like a zero-crossing detector and a digital timer which will measure the period and output the voltage via an D/A converter. This way you should get faster response than with analog circuits.

 

Hi All,

Thanks for the responses. I have read a significant amount of information and I'm learning how these devices are working. I am indeed trying to control the frequency of an actuator inside of a feedback loop. A slow response (1 < t < 3 sec) would not be ideal as I am trying to use this as quick as possible (~0.5 sec or less) in a feedback loop. The resolution of ~5 Hz is needed because I can only change my frequency over a range of about 500-600 Hz for some cases, and only 50-100 Hz for others. Maybe increasing that to about 10 Hz would be reasonable.

I think I may have found a software-based solution to my problem however. I am using MATLAB's Simulink and dSpace for the control of this actuator. I have devised a counter-based method to do this with a compromise between the resolution I need and the response time of the measurement.

If anyone still has any input on this, I would greatly appreciate it. The less that I have in my control loop, the better - an analog piece of hardware would be better than using my software-based solution. Again, here are the characteristics that I need to have with this sort of device:

Measurable frequency range: 4 - 6 kHz (would be nice if tunable to other ranges like 4.8 - 5.4 kHz or 2 - 4 kHz)
Frequency resolution: ~5-10 Hz
Voltage output: 0-10 VDC
Voltage input: +/- 5 VAC


Thanks again for the help,
Phil

 

The counter method should have almost no delay, something in the range of 1 period should be no problem. If you sampled the signal at say 44khz, then you should be able to get even better precision than 10hz. The only problem will be detecting the zero crossing, can you post a picture of how the raw signal looks like?

 

Here's a picture of the raw signal. It comes from an unsteady pressure transducer. I sent it through a bandpass filter with a bandwidth of 1 kHz.

 

That seems pretty clean to me, but can you post the unfiltered signal to see the difference?

 

I actually did not acquire the unfiltered signal when I ran this. I can take a dataset that has it when I set this back up again - hopefully within the next 2 days. The unfiltered signal is going to be pretty rough though. It'll have a DC offset as well as quite a bit of noise. I don't think I'll be using the unfiltered signal for anything, but I'll post it when I can.

Thanks.