I'm trying to do some rotational vibration measurements using a slotted disk with 120 gaps and an optical sensor to get the pulses.
There are 3 discs with 3 optical sensors in total and all of them are exhibiting the same behaviour.
We used the RV4 aquisition card to acquire the speed. Looking at the pulse train in tracking setup, the trigger is set up in the middle and the signal looks right, but when looking at the rpm trace the signal is very noisy.
Is not the first time it happen, I've used also magnetic pickup sensors in the past and had similar issue but I was able to partially fix it (up to a certain speed) by changing the trigger flank (from rising to falling). Please find attached a print screen with the rpm trace and please send me your opinions or solutions about how can I fix this.
Solved! Go to Solution.
without seeing the pulse data it's hard to say, but it looks like you have some imperfections on the disks, or a missing pulse or possibly the trigger level is not in the middle of the pulse.
I suggest T'ing the voltage into another channel and measuring the raw voltage. You can then look at the spikes in the rpm channel and see what is happening on the voltage channel at the same time. If you send the voltage channel and rpm channel to your local support team they can assist. We have options in the tacho pulse to rpm command to handle missing pulses and we have other tools in the Time Signal Calculator that can help.
If you do not need the 120 pulses per rev you could also try using 1 pulse per rev and skipping 119.
I have a slightly different view on what the issue might be.
The picture you attached looks like rpm measured by a lasertach. Some lasertachs produce a stair effect in the rpm like below.
This happens when the lasertach is not a "continouous" lasers. For example, the laser might turn on and off 1800 times a second. To the visible eye, you cannot tell the laser is turning on and off. By doing this, the lasertach consumes less energy.
This turning on and off causes the laser to not detect the edges of the pulses correctly.
The best solution is to get a lasertach with a continously on laser.
In the Testlab software, in commands like TACHO_PULSE_TO_RPM there are tools to try and compensate for this problem. The smoothing parameter can help. In the menu below, the smoothing is set to 10 - the rpm will be averaged over 10 revolutions.
Here is what smoothing can do on the stair step rpm.
The best solution, to eliminate the stair step in the rpm, would be a new lasertach. Smoothing can be used to try and compensate, but only in a post-processing calculation.
This makes sense. I'm using an optical Omron sensor with a response time of 20μs. But I guess it will still induce errors if it doesn't have a continuous light.
Thank you very much for your help.