VibSensor Tutorial #2
As I continue to add to this I will do so in short add-on’s like this to save on time and get the info out. I’ll have Dom link these to the bottom of post #3
. As those that have made these tutorial know, its extremely all consuming.
Here some additional details concerning the data view on VibSensor. Concentrating on the PSD and LogPSD it appears that on the PSD graphs, which displays the Linear PSD, the bins are of equal size and resolution in amplitude (vertical scale of the graph) as they are in the frequency scale, horizontal graph. In the frequency scale/domain we can scroll from left to right and zoom in to view all frequencies between 0.3hz-50hz. We don’t have that luxury in the amplitude domain. Plus we understand the frequency scale and how it allows us to pinpoint specific freqs. I don’t know what the amplitude scale means, i.e, 1e-01, 2e-13.
I thought I had a bead on it as 1e-01 is above 1e-02, but then 3.9e-06, etc, is thrown in to the mix. I know it relates to the G and m/s equation, but until i get an explanation, I’ll won’t use it. The LogPSD, on the other hand, has a very useful and understandable amplitude scale. It matches well with the vibration data.
To this point I’ve used primarily fixed frequency tones, sine waves, when testing. The question came up of can we use frequency sweeps, short and long. With Omnimic you have to use a very specific test track for each different function or the test is meaningless. So I came to the conclusion that sine sweeps won’t work. Well, thats wrong. Once again I fell back to Mic-based thinking as thats the comfort zone. This is a mechanical meter. It samples at 1/100 sec. and unlike a mic its actually resolves with more at lower frequencies as it takes and draws 100 sample points at 1hz., and decreases to 2 samples at 50hz. While all is resolved accurately you get smoother graphing down low. This meter will give you exacting data at 0.3hz. I have a .csv that shows EACH sample point in each of the three axis. I would love to see this in a spectrum display.
Here's a 1-60hz sweep .1 LFE track test run of my front subs only taken the first day I had VibSensor. It matches the Omnimic graph as only the front plane/array of subs are measured. This also shows the importance of making tons of runs. I forgot I had those runs. I didn't understand them at the time but now can pull them up and check out the history of test run and have the data available. Don't be intimidated. This is new to all of us. Test and post. Test and post. There's no SIM for this. We need the data to compile a database. You don't have to understand it to post it. For those decade and bin challenged heres some additional graphics on the PSD graph, in all three axis, displayed logarithmically in both the frequency and amplitude domains.
I started looking at the numerical data more. I think RMS Vibration is the one most here if they use this will go to. I believe this is the SPL reading of the Accelerometer test world. When all the data and axis and vibration and PSD is calc, its spits out this number and thats seems to be the NUMBER. That will be ultimately how we rate a given sub and these three axis rms number will be combined to one final number that tells us a subs tactile power response rating. So it goes like this in the future, Ricci has a Particle Velocity meter and now measure the Tactile Power on each sub. Now we see the DNA of each particular sub. I thought of that last night but Dom used that same word to describe it two years ago. I think now it could very well be true that different sub designs each have a unique tactile fingerprint. Back to DB testing. So now Josh does all his standardize testing and new cutting edge testing that spits out the SPL and TPR (tactile power response) which all gets calculated into one final number, lets call it the Crush Factor. One sub may get the nod SPL-wise, another by way of TPR, but its all about the CF. And I’m sure the DSP will have an sensor that we place in at the MLP with PV tech in it, setting the subs up and making sure we don’t screw things up. We could have a PV meter in our phones in a couple years or sooner if they wanted it. I can see it now, you walk into Bestbuy/Magnolia and they start in about the new Y-axis subs and why their the greatest thing ever. So when your doing your test runs do pay attention to the RMS vibration as it a quick summation of whats going on. And the raw peak can help with peak hold output if needed.
Hard surface measurements. To this point all measurements taken by VibSensor have been by simply laying on a seat/s or surface. When scientific grade measurements are needed, especially hard surface measurements (like different sub cabs), you must use a mount. Its really the best way to get consistent, repeatable numbers. The RAM one is excellent and its available on Amazon.
I want to mention is the VibSensor interface. It a great tool. But it would be great to have an app using VibSensor’s computational prowess in an app specifically designed for us. Heres an example of what I mean. We could upload pics of our system. How about all the data runs being uploaded for analysis. How about a private remote GTG with hundreds in attendance from around the globe. While we’re not their in person to compare in person,we see the ticker tape displaying all the realtime numbers. Possibilities are endless. Anyway to anyone with the expertise to pull that off, I’d gladly pay for that app. Heres some pics of the Dynolicious app I referenced at the beginning. It started out with some basic app functions (original Dynolicious) but now turned into a graphic beast (Dynolicious Fusion) that now costs $20, Sure would be nice and a extremely valuable tool that everyone has in common.
Pics of Dynolicious Fusion for comparison only. I got this app the same time started this thread and thats what started me down this road to a another way to test out subs and rooms.
No cheating. It knows.
You can see the possibilities here. Any SDKer's here.
Lastly, does anyone have Mathworks Matlab capability and experience. VibSensor can email huge .csv and .mat extension files if someone knows how that all works. I dabbled in that enough to know I’m not touching that with a ten foot pole. Someone thats sharp with an engineering background would be a huge help too work with these data and spectrum files. Tutorial #3
I’ll explore more of the different types of tests signal and introduce demo clips to run, multiple measurement placement and comparing multiple measurement platforms used together on shared test runs and some crazy way over limit testing just to see how far you can push this meter.