ScanSpeak 3-way project Running

[SGregory]

They are playing!!!!!!! I am sure response isn't flat yet but it can't be too far off.
Tweeter SS Rev 2904
Mid SS Rev 12W
Bass SS Rev 15W
All 4 ohm. Will start to measure soon. Waiting on USB sound card.
Cabinets are Cherry and Hard Maple. Top third of speaker is sand filled to dampen low freq. Mid Enclosure is presently 1 liter. I have one channel mid stuffed with bamboo fiber, and the other with poly stuff. (for future tests). Sound Baffle is laminated strips of hard maple and accented with cherry (breadboard style). 1.5 inches thick. The speaker walls are 7/8" thick.
The bass is rear ported. bass chamber is lined with rock board and 1.5inch open cell foam. There are several braces internal also.

As it stands now the crossover is First Order (by design) 350/3500+. The mid has a zobel and the tweeter has a series notch (Thanks to those that helped me on this board). There is no audible resonance or break-up now. I will start testing/measuring and tweeking shortly. I'll post the graphs as I develop them.

The sound stage is very stable with fantastic depth. Vocals are smooth and open. At this time, they are aggressive but not hard or harsh. I think I need to pad the tweeter a little more. At this point I am applying the 24 hour rule. (or more). I am just going to play them for a while to break in the tweeters. Steely Dan will make your eyes blink.

I am driving them with a Simple SE and a 12b4 pre. No issues with loudness except for extreme movies passages.

[ahegyes]

Interesting cabinets. Great job...but you still have some work to do on that crossover.

[SGregory]

lots of work on the crossover. I made these mostly with parts I had laying around. Can't wait to get my measuring equipment up and running again.

[SGregory]

I have started to measue and could use advise from experienced builders.
My measurement system is:
Laptop with Sound Easy v14. I am using a Emu0404 USB Sound Card Driving an Aleph5. The standard voltage divider probe (22k/47k) described in the manual with an 8.1ohm resistor in series with the speaker seemed to calibrate fine. I have a behringer mic but have not set it up yet(must make a holder/stand for it.) Tomorrows project.

The tweeter is a SS D2904/710003. The first graph is showing the raw speaker, 1mH/40uF/5.5R, and then 1mh/90uF/5.5R. The second graph is the final configuration which was installed 1mH/90uF/5.5R. I did other runs with different values of L and R but this was the best I was able to obtain. Does this seem reasonable?

The third graph shows both the tweeter and the mid hooked up with full compensation. For the mid I added a Zobel and a Series notch filter. The Zobel is 10uF/4R and the Series notch is 4.5mH/270uF/7R. I question the series notch as I don't have too much of a feel for it. To me the values seem big.

The final graph show both the mid/tweeter and the bass/mid/tweeter. I have no compensation in the bass nor have I tuned the bass cabinet yet, as I am waiting to get the mic running.

Based on your experience, does the above seem reasonable at least for the electrical side of things?

The HP is 8uF or 10uF (used 8 for these graphs)
The BP is 100uf/.22mH
The LP is 1.25mH

[SGregory]

Bump

Being my first design from the ground up, I am curious if the above plots are reasonable?
I won't have spl until my mic comes back from calibration. I'll post that info when I get it back

[SGregory]

Learning on this project all of the time. Actually have some SPL data to share. The data collected is using HOLMIpulse and simulations were using EDGE. Thanks to both of these guys for contributing this valuable software.

The first plot is the frequency response using just electrical impedance correction and theoretical xover points. No too flat, but playable. This is the configuration from the earlier impedance plots.

After considerable tweeking and response shaping I arrived at the second response. Response is flat with the exception of two dips. approx2.1kHz and 11kHz. I think I know the answer to the 2.1kHz and is shown in the third plot using EDGE. By compressing the drivers and moving them off center in the baffle I come up with the fourth graph, again using EDGE.

I had initially hoped that the edge rounding I applied to the baffle would have minimized the diffraction problem, but I was wrong. Before I route out the baffle on my speaker and put in an insert, does my thinking on this make sense?

Time to clean up the prototype crossover. It is one big goober of parts. Who said first order is easy?

[SGregory]

Here is some raw driver data mounted on the baffle. No crossover, just the driver. Particularly the midrange shows a similar response to the Edge model above.
First graph is of D2904/710003 tweeter, the second is of the mid 12M4631.

I have been searching the threads and see several people with the issue. It seems that many offset the driver and then use wedges of felt. Would 1/2" work if it is at least partially inlayed into the baffle?