The good news is that the peaks are all in the right places, which suggests that the box's parameters match the sim.
The bad news is:
(1) the measurement suggests that the driver's Le (and perhaps Re) don't match PE's published specifications. Pull the driver and measure its parameters to confirm. If the specs are off, then the HornResp model needs to be adjusted accordingly.
(2) the upper peak is a bit low, suggesting that the box can be improved with a bit more bracing. Sit on the box and measure the impedance curve again - if the upper peak rises significantly in value, that will confirm the bracing question 🙂.
That should be easy to check - bring up the sim in HornResp, then double-click on the "Sd" input box. The t/s parameters should be displayed in a new dialog box. Check and confirm that the displayed parameters match what's shown on PE's website for the driver.
Oh sorry Brian, i was thinking outloud here. I should have asked KTOKTO if he sim his driver in HR and use that generated data in DATS.
I'll have more time tonight to continue.. Thanks again Brian.
I did sit on the box after reading your page. Both standinging uprigt and on its side , measurements were similar to what I posted here.
I did sit on the box after reading your page. Both standinging uprigt and on its side , measurements were similar to what I posted here.
Last edited:
I ran some test tones through the box at higher power after I did the impedence sweep. I didn't find any air leaks or hear any weird rattling or noises from the box. All of the pictures on the basement and upstairs walls in the house were rattling. This should be a fun box once everything is fixed.
Love this thread! The impedance stuff is beyond my scope at this point... Can someone elaborate what and why this needs to be fixed? I am taking a wild guess that it appears you won't get as much SPL as predicted? (Le/Re off)
Me too, I am just pushing buttons, making saw dust and learning at this point.
I will get the actual measured parameters tonight and plug them back in to this boxes hornresp record and see what happens.
This is so thru. You could have the best impedance graph and sim all you want if the hardware sound like crap, all the fancy measurement can't make your box better.
Most of the time I build prototypes out of MDF, do some measurements and see if there are any issues appearing EQ wise. If this is ok then run at war volume with a couple of mid/high tops, and put on some favorite tracks I always use for testing because I know them inside out and how it sounds on different systems. If it handles it for a few hours and others who listen agree with how it sounds then success!
Before I forget, is this cabinet easy to move or is it of the odd/weird I need a dolly even though it's not big variety?
Depends if you put handles in/on it or not, without a driver it is already heavy for me to carry around.
Mine is similar in size and weight to a 1x18" BR cabinet, it is a few inches wider than my JBL JRX118s boxes but similar weight (I would guess 85 pounds but will get an exact weight this weekend).
I am adding corner handles like the example below and putting casters on the back so we can roll them in. This worked well with the JBL boxes. I wouldn't want to have to carry 4 in by myself without casters.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
1. Measured impedance at the minimum points is higher than expected. This could be because Re does not match published value, box losses are higher than expected (could be caused by stuffing), or measuring device (DATS here) is not calibrated properly. I'm beginning to suspect that it's the last option. Ktokto, did you perform the measuring leads calibration process? To check, just connect the leads together and do and run the impedance measurement. The result should be a graph that shows an impedance of zero ohms at all frequencies. If not, you need to perform the calibration process.
2. The impedance rise at higher frequencies is lower than expected. This usually means that actual Le is lower than published. This would not be the first time I've seen this from a Dayton driver - the PA310 I used in my POC3 build has a measured Le value that's less than half of the published value. Large changes in Le will affect the response in the passband of the TH.
Finally, the measured peaks in the impedance response curve will always be less than the peaks predicted by HornResp, but they should occur at or very close to the same frequencies, and it looks like this is the case for this particular build, which suggests that the build matches the HornResp sim wrt path length, size and expansion. I have found though that the lowest peak always tends to be a little higher than predicted, I'm not sure why. The height of the measured peaks is dependent on how "lossy" the box is, and I've found the uppermost major peak (there are three major ones for TH designs) is the one that's most impacted by bad bracing and panel flex - the higher the measured peak, the better the bracing and the lower the panel flex.
I believe I had everything calibrated but since this was my first time using DATS I may have had something fouled up. I will double check everything tonight and measure again with the driver in and out of the box. I had the box sitting on some really thick carpet and padding. I don't know if that would effect the measurement or not.
Last edited:
Here is the sweep with the leads connected to each other.
And here is the measurement parameters with the speaker out of the box (elevated on a couple blocks of wood so the rear vent on the magnet was off the ground)
The parameters are off of what PE lists.
I plugged in the measured parameters (except vas / mmd) in to the original sim. The curve is definately more saddle shaped.
I was unable to measure vas as I did not know how to accurately measure the piston diameter on the paper surround driver. I also do not have a digital scale on me so I was going to use a roll of quarters if that would work.
I also measured the drive in the box multiple times on its side, back top etc while sitting on all the panels, the measurement was the same every time.
And here is the measurement parameters with the speaker out of the box (elevated on a couple blocks of wood so the rear vent on the magnet was off the ground)
The parameters are off of what PE lists.
I plugged in the measured parameters (except vas / mmd) in to the original sim. The curve is definately more saddle shaped.
I was unable to measure vas as I did not know how to accurately measure the piston diameter on the paper surround driver. I also do not have a digital scale on me so I was going to use a roll of quarters if that would work.
I also measured the drive in the box multiple times on its side, back top etc while sitting on all the panels, the measurement was the same every time.
The best way to measure the driver's parameters is to hook up the DATS leads to its terminals, then stand up and hold it out by one of the basket spokes as far as you can away from your body when the sweep is done. Do it a few times until you obtain consistent results.
Even then, the results you've obtained are quite interesting. Fs is almost spot on (published=37.7 Hz, you measured 37.68 Hz - that's basically the closest I've ever seen). Measured Le is also lower than published Le as I expected, based on the impedance curve you measured.
What I wasn't expecting though was the lower Re and the significant difference between measured Qes (0.52) and published Qes (0.37). That difference seems pretty large.
Concerning Sd, I think it's safe to go with the published spec. When measuring Vas, I prefer to use the delta compliance method - basically connect the driver to a small sealed box and determine Vas from the change in resonance frequency. DATS also provides the means to do this. The good news however is that Vas can change quite a bit before it makes a significant difference in the
Ok, if it measures the same, then bracing isn't an issue. Have you imported the predicted impedance curve for the adjusted HornResp sim into DATS yet to see how it compares to your measured curve?
As for the ripple in the adjusted sim, it doesn't seem to be that significant in level. Box losses are likely to flatten it a bit anyway.
Here is a new one holding the speaker in the air.
An externally hosted image should be here but it was not working when we last tested it.
Ok, so very little change.
Time to measure the Vas, to see how badly off-base we were with the original sim 🙂.
Got any unused small sealed speaker boxes anywhere?
Not available right now. Would the added mass route be accurate enough?
Last edited:
- Status
- Not open for further replies.