I do but on looking at the part again, I see I screwed up - it has a 400Hz low pass filter built in. Teaches me for not looking closely.
The accelerometer that I'm using now is home-made from a piezo buzzer out of a broken clock radio. I wanted something to compare it to.
It's a conspiracy - with so much to do this evening upon getting home I'm greeted by a black out - no power for 3 hours. Oh well, I needed time to relax anyway.
To satisfy my curiosity last night I did test the box but the results are inconclusive. I will test again (soon, I promise) and post the results. What I didn't do in the "before" tests was record the peak output so there is no point of reference for the new measurements. Arta shows it worse than it was before, at first glance. I'd like to find a CSD plot software that doesn't use the peak as 0db but plots it relative to the output.
Yes, hard to explain and I'm tired and frustrated (despite cooling my heels in the dark here for 3 hours).
I do but on looking at the part again, I see I screwed up - it has a 400Hz low pass filter built in. Teaches me for not looking closely.
The accelerometer that I'm using now is home-made from a piezo buzzer out of a broken clock radio. I wanted something to compare it to.
It's a conspiracy - with so much to do this evening upon getting home I'm greeted by a black out - no power for 3 hours. Oh well, I needed time to relax anyway.
To satisfy my curiosity last night I did test the box but the results are inconclusive. I will test again (soon, I promise) and post the results. What I didn't do in the "before" tests was record the peak output so there is no point of reference for the new measurements. Arta shows it worse than it was before, at first glance. I'd like to find a CSD plot software that doesn't use the peak as 0db but plots it relative to the output.
Yes, hard to explain and I'm tired and frustrated (despite cooling my heels in the dark here for 3 hours).
Sorry about your luck!
I did a quick internet search and came up with this link:
Using LAUD with Accelerometers
Digikey part number: MSP1001-ND. Cost US $45.00.
I think I will order one and give it a try.
Last edited:
If that is the ACH-01 it is cheaper at mouser. Mine is still unopened on the kitchen table.
dave
True, but I have a business order pending with DigiKey, so it is easier to simply append it to that order and it becomes a business asset tool instead of coming out of my personal pocket.
I will check out LAUD.
Send it to me, I'll open it
I had an ACH-01 added to a Digikey order nearly 2 years ago (when it was "just" $26) but deleted it in favour of some other parts I needed more. My loss.
Still haven't measured the "sili-surround" speaker but I did notice the my quick measurement the other day (the one I said looked bad) was taken with only 2 screws holding the driver in. It may make a difference if I put all of the screws in before testing...
Too much to do, not enough sleep takes its toll.
yes
dave
Oddly enough it didn't make much of a difference.
Measurement of the box in its original state:
Measurement after I added the magnet brace:
And measurement of the "sili-surround":
Without a point of reference between these, the results are pretty much useless. I didn't record the output peak on any of them and Arta putting the peak at zero db regardless of the output level makes this utility only useful for eye-candy.
To give an example of what I'm saying, here is a measurement with no output - the speaker not making any sound:
Hi John, i see your problem
I'm assuming you always stick the same signal at the same level into the speaker? If so couldn't you simply measure the voltage output after the preamp to get an idea of the peak?I'm sure different software would be an advantage, but as a quick fix it might well work.
Mark.
Hi Mark,
No need to actually measure it, Arta does and displays the peak in mV. Now if it could be made to alter the plots with this number as a reference, then it would be useful.
Right now you cannot visualize any improvement based on these measurements.
In general, from my own observations, the plots with the most "noise" are the quietest panels, output wise. A panel with a high Q resonance will look very clean because the peak is what the plot is showing, move the db scale downward to equalize it and you see all of the "noise" that shows up in the other measurements.
Frustrating and I'm glad I didn't spend $50 on my accelerometer.
No need to actually measure it, Arta does and displays the peak in mV. Now if it could be made to alter the plots with this number as a reference, then it would be useful.
Right now you cannot visualize any improvement based on these measurements.
In general, from my own observations, the plots with the most "noise" are the quietest panels, output wise. A panel with a high Q resonance will look very clean because the peak is what the plot is showing, move the db scale downward to equalize it and you see all of the "noise" that shows up in the other measurements.
Frustrating and I'm glad I didn't spend $50 on my accelerometer.
Sadly true. That last post gives an idea of the background noise of your "house"
dave
I don't think the furnace was running at the time but clearly seen is the fan noise (?) from the computer at 400Hz. It is the closest source of noise in any case.
And if you study the 3 plots above you can see that 400Hz peak - it could give some perspective on the actual output.
Last edited:
Ah, i see
In actual fact you can get very useful information from it then, it just doesn't display too well as 0db is always the peak of the chart. However, you'll know how much better things are by the reduction in peak voltage & that can easily be converted in Db.It just doesn't display to well so things end up looking similar even if they aren't. How about simply adding the peak voltage to each graph, after all if we can't convert things we shouldn't be here
You know what i mean chap, by the way, where did this software come from as i'd be interested in it myself...
Keep up the good work
Mark.
I have done this but it makes no difference in the CSD plots - they still look exactly the same. If the peak output is 1mV or 50mV, no difference in the plot and that is the difficulty.
All of this is difficult to describe but I'll do a demo next that will give a better picture.
gotcha - there is a lot of noise somewhere in the circuit. Anytime I have that much noise i assume there is a ground problem and I also find it is correctable. I have confidence that you will be able to get this to work.
Hmm, maybe it would be that the "sound" of the box is irrellevant as some claims. For me it would be the best explanation, no need for excessive box work. In my line of work, this is crucial. Do not make the product better than necessary.
Anyway John, I admire your persistance...
Cheers,
Peter
Anyway John, I admire your persistance...
Cheers,
Peter
The circuit that pre-amplifies the accelerometer? It contributes no noise whatsoever, totally silent operation.
The accelerometer is picking up background noise in the same way as a microphone will pick up ambient noise. The point of the display was to show how Arta is plotting - setting the max peak at zero db regardless of the output level.
See, I said it was hard to explain.
oops - i should have read what you wrote more closely instead of just looking at the graphs. It is a bad habit i have developed from having to read too many papers. sorry.
I'd like to see some raw data. In fact, if you can post a file I'd be happy to run some analysis.
gotta get my son up now from his non-nap - will make for a long afternoon.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.