Look at my new babies...Altec 846Bs w/EV Tweeters!

[Zilch]

What about for 846a- 16 ohms. Is there a board of parts and schematic yet?

O.K., here's the deal on the 16-Ohm 846A:

Jack has made a PC board, and the circuit is simple, and tested at this point. However, it does not address the 19 kHz spike well, so I've deleted that part. Since doing that one, I have done three additional iterations for other systems and learned better how to do this.

Jack is getting set up to do measurements, and wants to measure the performance of the present circuit in his system. He's also likely to use the BMS in his third system, and if so, I have asked for an opportunity to work with those new drivers and finalize the design for 846A.

If you're going to do BMS, PM me, and we'll work a way to do it with your 16-Ohm drivers as well.

The compensation filter he's presently using appears below, which I tested with your N-800-F:

http://audioheritage.org/vbulletin/s...329#post160329

Jack has built new crossovers with a 1.2 kHz frequency, which appear to work with this compensation filter as well. That's one reason he wants to do measurements first.

This filter will work with BMS drivers in 846A. I'm saying there may be a better version available soon....

[SMKSoundPro]

10-4.

message rec'd.

standing by for advice.

Thank you!

scott.

[johnaec]

I was able to listen to a couple examples of the BMS drivers with the 811 and 511 horns over at Zilch's yesterday. A definite improvement from the last time, when he was still tweaking the Altec drivers. I also thought the 511 sounded better than the 811 with the same driver, even though both were crossed over the same, somewhere above 800hz. The 511 just seemed to sound a little more open.

That said, Zilch also had the horn from the 4348 hooked up to a JBL driver, (2435 or 2452? I can't remember...). As expected, this combination really shines! Zilch has some interesting plans ahead for these!

A side comment - I was over there to get some CLIO curves run on the 2447/2381 HF combos from the SP215-9's I recently purchased. Since that cabinet was last produced in 2002, the drivers had to be at least 5 years old, still with the original diaphragms. We compared them with a brand new 2447/2381 combo, and both older drivers measured within 1 to 2 dB of the new driver, (the used ones actually had a little smoother curve), and the really amazing thing - both older drivers measured on average within .5 dB of each other! If that's not a testament to JBL quality, I don't know what is!

John

[jackgiff]

Zilch, these pix are of a Trusonic 80CX coax which is convenient for testing/learning purposes. Do the responses look like something to be expected from a speaker of any kind?

Thanks, Jack

[Steve Schell]

Jack, if those divisions are 15dB then the response may be a bit like a ride on the Cyclone Racer, but many of the strange old speakers I've got measure worse than that in their raw response. At least the overall response is gently downsloping instead of having some killer upper mid /treble peaks.

A friend recently gave me a printout of an article by Nelson Pass, "Current Source Amps and Sensitive / Full-Range Drivers." You are dealing with a coaxial, but there may still be something useful here. On the passdiy.com site there is also his newer article, "Current Source Crossover Filters."

I have experienced dramatic bass improvements with amps having high output impedance... the phenomenon is real, like taking one's hand off the cone. Perhaps one of his filter suggestions could help level out the midband as well.

http://www.passdiy.com/articles.htm

[Zilch]

Hi, Jack.

Well, it's certainly rolled off above 12 kHz. I've sent you a PM with detailed comments. I'm ignoring everything below 630 Hz as contaminated with room effects for now, but if the rest of it above there is accurate, it looks like a '60s speaker with "issues."

I look at your curves and my fingers are itching to twiddle knobs and push buttons on your new RTA to figure out what's up....

[jackgiff]

I have been playing with my RTA most of the afternoon, and moved on to a couple of Valencias which were out in the garage. These are not the BMS drivered versions yet, because I wanted to get some experience using the RTA. By the way, it is a BLAST!!!

The first picture is the FR of an 846A which has an Eminence APT-80 Supertweeter mounted in the horn. The connection for the woofer/806A phasing is as Altec shows it on the wiring diagram on the back of the Valencia. This tweet is coupled through a 4 Microfarad cap, and the 806A is allowed to go as high as it can. The tweeter then just comes into the picture slowly and adds to the decaying response of the 806A. As can be seen by the dip at about 1200 Hz, the 806A should be hooked backwards, so that it doesn't subtract from the woofer during the crossover overlap.

The second picture is the FR after the 806A connections are reversed. Much nicer. The HF was also reduced before this pic was taken.

The third pic was after the HF was adjusted to normal listening levels by ear, not by watching the RTA. I tried to better the FR by using the RTA, but it seemed to be a tossup.

I haven't listened to this Val much yet, but will spend some time with it tomorrow. The listening will surely be better after using the RTA to get the phase corrected between the woofer and 806A.

Jack Gifford

[Zilch]

Tweeter's not running in the second pic, right?

I'm certainly interested to see how your stock Vals measure without HF augmentation, of course.

At 1.2 kHz, the phase looks "corrected" there, yes.

[jackgiff]

These pics show the FR of a Val with an APT-80 supertweeter mounted in the horn, and crossed over from the 806A at 3500 Hz. Again, I haven't spent a lot of time listening to this Val, but will tomorrow.

The first pic is the FR of the unit as wired according to Altec's wiring diagram, and as it was wired for the comparo. Again the big dip in response at about 1200 Hz is due to the 806A subtracting from the woofer output during the crossover period. It is hooked backwards.

The second pic shows the FR after reversing the leads of the 806A.

It seemed like the notch at roughly 2 KHz was also gone, but I tried reversing the leads of the tweet just to check it out. Sure enough the big dip came back, so the tweeter leads were put back to the same phase as before. The ultimate response of this Val is shown in the second pic.

The RTA is going to keep me busy for week or six learning what it can do. But it sure looks like FUN!

Jack Gifford

[jackgiff]

Tweeter's not running in the second pic, right?

At 1.2 kHz, the phase looks "corrected" there, yes.

Yes, the LPad for the tweeter was turned way down, essentially taking it out of the circuit. The drop off of the 806A is pretty apparent in that pic isn't it?

Jack Gifford

I will get the stock Vals in the next day or so, as well as the BMS guys. This is going to improve my photography as well. Come over and help me tweak knobs. That will shorten my learning curve.

[Zilch]

Yes, the LPad for the tweeter was turned way down, essentially taking it out of the circuit. The drop off of the 806A is pretty apparent in that pic isn't it?

I'd say you're confirming the ZilchLab findings with respect to Val response in actual use, yes. I didn't have real Vals to work with here, just the horns and drivers.

*****

It certainly appears that crossing in the tweeter at 3.5 kHz is too low, creating a 5 dB "hump" at 4 kHz.

Thus far, just the 4uF would appear to be a better solution, at least insofar as frequency response is concerned. It's probably making a bunch of phase-interference combing as the two HF drivers interact in the overlap region, tho. Finding the optimum frequency to do a "real" crossover would be a better approach, as you've previously stated.

Having measurement capabilities is pretty much essential for doing this stuff, in my experience.

More data ... less wank....

[jackgiff]

I'd say you're confirming the ZilchLab findings with respect to Val response in actual use, yes. I didn't have real Vals to work with here, just the horns and drivers.

*****

It certainly appears that crossing in the tweeter at 3.5 kHz is too low, creating a 5 dB "hump" at 4 kHz.

Thus far, just the 4uF would appear to be a better solution, at least insofar as frequency response is concerned. It's probably making a bunch of phase-interference combing as the two HF drivers interact in the overlap region, tho. Finding the optimum frequency to do a "real" crossover would be a better approach, as you've previously stated.

Having measurement capabilities is pretty much essential for doing this stuff, in my experience.

More data ... less wank....

Zilch, sometimes a good wank is OK. Better than no wank at all.

The parts for a 5 KHz crossover will be here on Friday, and it may do a much nicer job. I feel that any higher, and the 806A is getting in pretty treacherous territory. But then, what do I know?

And messing with the value of the 4 Microfarad cap can move the point where the tweet comes in. That will get played with a little bit in the next few days.

Having the capability to actually measure the response is BITCHIN'! Wish I had done that much, much sooner. And when I finally get to the time delay capabilities of the DEQX, what will be in store at that point? WOW

One other thing to remember is that my Vals have the fretwork grilles, which can interfere with the HF a bit. I have a large piece I could send you to cover the front of your horns for trial purposes. But then again, we are almost done with the Vals, yes?

Thanks, Jack Gifford

[Zilch]

One other thing to remember is that my Vals have the fretwork grilles, which can interfere with the HF a bit. I have a large piece I could send you to cover the front of your horns for trial purposes. But then again, we are almost done with the Vals, yes?

I think some high-resolution measurements with the grilles would answer the questions about their influence, so send the piece up here, and I'll study it a bit.

With respect to being done, there's just a few measurements more I have yet to complete. My goal is to pack everyone's drivers up and return them yet this week. More work to do with tweeters, especially, and I still have to check out the "light" diaphragm Gary sent.

I'd say we're not gonna be done with the Vals 'til the Comparo is repeated with the several approaches optimized. However, the fundamental result seems conclusive at this point -- they need and warrant upgrade.

[Gary L]

With respect to being done, there's just a few measurements more I have yet to complete. My goal is to pack everyone's drivers up and return them yet this week. More work to do with tweeters, especially, and I still have to check out the "light" diaphragm Gary sent.

Play away Zilch! I have at least a solid month before I am ready to mount those 802s on any thing.

I am just sitting back at this point watching you guys tweek the daylights out of these things and hoping you come up with some directions for us die hard Altec fans who simply must maintain the traditional drivers.

I have not yet ordered a pair of BMS's but will admit I am learning to appreciate the Eminence PSD 2002s in current use. So far I can't find a thing not to like about them. After extensive yet unscientific A+B testing along side of the 802-8Gs I can honestly hear more content from the 2002s and they appear to be much cleaner when the HF spikes to high levels.

Gary

[Zilch]

Having the capability to actually measure the response is BITCHIN'! Wish I had done that much, much sooner. And when I finally get to the time delay capabilities of the DEQX, what will be in store at that point? WOW

Everybody's gonna scoff, but there's an inexpensive analog active crossover available that has adjustable delay on the LF. You can tweak the "delay" knob and watch the time-alignment notch disappear on the RTA. I may dig in there and figure out how it's done for DIY. Perhaps Ian or someone knows already.

Swipe mouse here for the product ID -----> Behringer Super-X Pro CX3400, <$130 at Guitar Center and similar.