I was working on a project recently where I needed to take a single frequency output from a liquid coriolis meter and split the signal into two identical signals. One of the signals needed to go to a flow computer to calculate the flow of oil going through the coriolis meter and the other signal needed to go to a prover connection so the meter can be proved periodically.
Apparently, the pulse or frequency signals used for the meter and the prover connection have to come from the same signal generator in the meter and the meter only had one pulse/frequency output. The frequency from the coriolis meter needed to be 10,000 pulses per barrel, or 10kHz.
We ended up using a PE-100 Liquid Sample System Interface made by Angus Measurement Services, LP. It was a simple, off the shelf solution that worked. The PE-100 can do several things and is designed to not only split a pulse or frequency signal, but it also has a relay output to send pulses to a sampler solenoid. I think it might have been overkill for this project, but like I said, it was a simple, off the shelf solution.
This got me thinking though. I was wondering how hard it would be to make a simple pulse/frequency splitter that is accurate. One that could essentially take a single signal in and split that signal into two identical signals without losing any of the original signal’s integrity, thereby maintaining the accuracy of the original signal.
Turns out, not only was this a pretty simple task, I already had the parts to do it in my shop. I wanted the splitter to be powered from a 12vdc or 24vdc power supply since both voltage levels are readily available in most liquid measurement control cabinets. The splitter also needed to be small in size, and as I said before, accurate. I wanted the splitter to be able to handle anything from 1Hz up to 20kHz – one pulse per barrel up to 20,000 pulses per barrel.
What I ended up with was a splitter that can operate from about 8vdc up to 30vdc, splits almost any shape of waveform into two identical signals, and works accurately from 1Hz up to about 30kHz. My splitter works well between 20% and 90% duty cycle. The sweet spot seems to be between 25% and 75%.
It draws between 30mA and 36mA depending on the supply voltage, amplitude of the output signals, and ambient temperature. I used several methods to test the splitter, but as a final test, I connected both output signals to an ABB Totalflow XRC motherboard on two pulse inputs. Normally, a Totalflow XRC pulse input needs at least 6vdc of amplitude to register a pulse signal. I found that my splitter needed the input signal to have an amplitude of at least 7vdc to output two signals correctly. The splitter dissipates about 1vdc of the input signal amplitude. This should not be a problem though.
Overall, this was an interesting side project. I learned a few things about signals and had some fun with it. As you can see from the pictures below, I developed the splitter circuit on a breadboard, then built the circuit on perf-board. Finally, I drew the circuit in CAD. Maybe someday I’ll send the files to a fab shop and have some of them built.
Breadboard
Proto-board
1Hz Signal
10Hz Signal
1kHz Signal
10kHz Signal
30kHz Signal
Pulse Inputs at 1kHz
Pulse Inputs at 10kHz
Pulse Inputs at 17.295kHz
Pulse Inputs at 30kHz
PCB Layout
3D View of PCB
I hope you found this interesting. If so, please share it.