My A7-500s. A journey

[Panomaniac]

Thanks for the info and the encouragement, guys! It very nice.

OK, going to start talking crossovers. This is just the beginning. Crossovers can make or break a speaker. They are a full 1/3 of the speaker - drivers, box, crossover. All equally important. Many speakers do not live up to their full potential because of poor crossover design. There has been a lot of progress in the past decade on computer crossover design, but it only helps, it does not do the whole job - at least not well. It takes a lot of experience, time and effort to design a great crossover. I'm a novice, but learning.

So here we go!

My A7-500s came with the stock N-501-8 crossover. This is the 500Hz crossover for 8 ohm drivers. I have attached a schematic that I drew from the actual part. I measured all values. The inductors measure 2.7mH outside the crossover, but probably rise to ~3mH when mounted in the steel frame. I have called them 3mH. Also included is a photo of the insides. The caps are marked 28uF 125V and measure spot on. L-Pad is 24R & 5.6R Just a basic 2nd order crossover with L-Pad. No Zobel, no traps, no EQ.

The values for a nominal 500Hz, 2nd order Butterworth filter are 28uF + 3.6mH. So the N-501 has slightly smaller inductors than "textbook." The problem is, although the drivers are nominally 8 ohms, they are not 8 ohms at the crossover frequencies. (See impedance graphs in the next post.) So the frequencies are shifted a bit. Simulations show the 416-8A electrically crosses a bit higher, the 511/802 combo pretty close to 500Hz. Electrically speaking. I will try to measure both the electrical filter function and the acoustic points for further reference.

Here is the schematic I drew from the actual crossover and pix of the guts. Nice big L-Pad!

[Panomaniac]

When doing box tuning and designing crossover, driver impedance curves are a great help!
One of the best audio gadgets I ever bought is the WT3 woofer tester from Parts Express.
Just a little USB gadget, it makes measuring drivers very fast. Only a couple of seconds to measure everything but Vas. And that only takes a few seconds more.

Plots show here are from the WT3.

Below we see 2 graphs, 3 plots. One is the woofer in the A7 box. This is a 416-8A. You can see the single peak at resonance (~23Hz) and the rising impedance typical of a woofer. Also obvious are some small bumps that indicate box resonances. I'll work to get rid of those.

Do notice the single peak. Does not look like the classic twin peak bass reflex curve, does it? More like infinite baffle. Maybe that's because of the big port. Currently 206 sq. inches or about 160% of cone area. We can see why Marvin brought his port size down!
I'll do the same and run a sweep.

The other graph is the impedance of the 802 driver on the 511 horn. One is with loading cap, the other without. As the peaks come near the crossover point, this makes things a little tricky. An LCR trap could be used to flatten this peak, might be worth trying.

Just for fun, I tried running the drivers with an active crossover. 3rd order at 610Hz. It works well and certainly sounds better than the stock crossover. Worth investigating further.
FWIW, the way my 511s are mounted, there is only 4 cm of depth offset between the acoustic centers of the woofer and horn. Very close to time aligned. Just a small fraction of a wavelength. Nice!

[Old Guy]

Altering the xover point will alter the sound a lot.

500, what you have, results in a more mid heavy, very "up front' aggressive sound.

This can be very desirable with a straight bass reflex, but the 828's short mid bass horn tends to boost that area too.

Going up to 800 will mellow them a lot. To 1200 even more...but how much is good is personal preference.

An electronic xover can help , play around with crossover points till you get the sound you like, then design your passives for that point...or just stay electronic.

Personally, I will never go back to passives. Never.

And this from someone who insists on never saying never...

Really glad the horn worked out for you, in the old days would have had a spare fram to send, but those days are gone...

[GM]

Do notice the single peak. Does not look like the classic twin peak bass reflex curve, does it? More like infinite baffle.

The little bump is the reflex's upper impedance peak which is highly damped due to the cab being tuned so much higher than the driver's Fs, ergo if tuned a similar amount below Fs it would become the ~single peak with the lower one just a small bump and of course equal in amplitude if tuned to the driver's Fs.

Anyway, if you recall my remarks about driving vintage speakers with the type of high output or matching impedance sources they were voiced with, this high tuning relative to Fs is desirable since it yields much better woofer control/higher low distortion SPL while still having extension to slightly below Fs same as if tuned to Fs with the trade-off being it sounding a bit 'loose' due to the amp 'stalling out' against the high impedance.

This same high output impedance source would also flatten the HF horn's 'hot' mids somewhat while adding scharron's 'gentle climb' EQ to flatten it out with increasing frequency due to increasing driver inductance.

GM

[Panomaniac]

Hi GM, I was hoping you would "tune in."

Yep, I will be working on the port tuning, for sure. Will aim for equal peaks to test how that sounds. And for sure, a high output impedance or resistive network on the horns sounds better. Not quite sure about it for the bass, yet.

[Panomaniac]

Sorry to report no real progress over the weekend, too busy with yard work and other chores. Mostly just listening. =)

But I did try to figure out the internal volume of the box. I have read and heard various numbers, but nothing definitive. So I started measuring. I figure this:

Gross internal volume = 15 cubic feet
Horn section + woofer = 3 cubic feet.
Net volume = 12 cubic feet = 340 liters.
(for box with horn flares open in back)

Actually did all my calcs in cubic centimeters then converted. Years ago I had figured 11.5 feet going from published plans. So pretty close to my measured results. I sure looks like the horn and woofer should take up more than 1/5th the volume, but I measured and calculated 3 times. Anyone have an opinion?

Given an Fb of 340L, and the port size I have, my cabs must be "tuned" to about 65Hz. Or about an octave too high. A starting point!

[aditya]

Hi Panomaniac,

This is really a great detailed kind of a thing from your side....and many many thanks for this. I was really wanting to see something like this as I am in the process of hand building a pair of A7s for myself. Pls keep it up.

The values for a nominal 500Hz, 2nd order Butterworth filter are 28uF + 3.6mH. So the N-501 has slightly smaller inductors than "textbook.".............

Pls correct me if I am wrong.

I think iron-core inductors are like that. Unless a sizable amount of current passes thru them, they will be less than the prescribed value. They are not linear in the graph. But for the VOTs it was not a setback as they were meant to pump away at very high & constant volumes. Also the reduced resistance of the iron-core inductors were an added bonus.

Another interesting point to note here is that, in a standard 12db crossover, if the value of the inductor is a little reduced (as is the case here) while the capacitance value is alright, it will infact put a slight bulbous rise (IIRC...nearly 2 db), at around the crossover point. Which is probabely why the A7 sounds much better at high levels compared to lower ones. Also as they are very efficient units, the mismatch becomes much more apparent.

For home use, IMHO one should either go for air-cored inductors or active X-overs.

Thanks again for the great thread.

[Panomaniac]

Hey guys, thanks for the kind words and good ideas.

I did a little work and measuring and found that the 291/803 will actually play flat down to ~250Hz. Roll off under 250 is 6th order Butterworth. I.E., it drops like a stone. To get this, the midrange has to be pulled down about 7dB. That flattens it out.

Trouble is, phase has gone wacky at that point, so good luck matching phase at the crossover point. Distortion is higher at the bottom of the range, too. So minimum 400Hz, or 500Hz to be safe. Would be interesting to see if the "Giant Voice" drivers could do a better job down low.

So happy Thanksgiving. See you soon!

[-ec1288]

One of my DIY audio projects - Altec Lansing A5
Configurations are as follows:-
Cabinet - 825
Woofer - 515B
Horn - 1005B
Driver - 288C
Crossover - DIY

For details, go to link showed in my signature.