Time to winnow down the speakers. Cornwalls? Valencias? 605A's?

[Zilch]

Zilch, are you saying this crossover doesn't have MF EQ (compensation)?

http://home.earthlink.net/~jmarkwart/sitebuildercontent/sitebuilderfiles/605a_brev1.pdf

Nope, that one does, via the "T" filter there.

The one you linked to for the 605A/B does not.

http://home.earthlink.net/~jmarkwart/id3.html

The N1500A for the 604E does, in a rudimentary form.

We don't know what N1600 has, yet, but I'm assuming it's much like N1500A in this respect....

 

[Billfort]

Ah huh.

I linked to the guy's home page the first time - both versions can be found there.

 

[Zilch]

I have 2 original N1600 here, but I don't know how easy it would be to test the L and DCR of the choke nondestructively.

There's another issue, also. N1600A was used with 604B/C. If N1600C is the "Right" crossover for 605A (B?), it's reasonable to assume that there are differences for use with the reduced-sensitivity 806A HF driver in 605s.

I'm having the same difficulties with this as I did with Vals. Nobody seems to KNOW anything about this stuff, or, if they do, they're keeping it to themselves, or just don't care, and it requires considerable effort to unequivocally nail it down.

Once you DO nail it down, however, you find Altec engineers knew what they were doing in many unanticipated respects. Model the N1500A crossover. It'll knock you out what's there....

I linked to the guy's home page the first time - both versions can be found there.

Oops, sorry.

Yup. I did the "Easy" one.... :

I guess my point right now is that the Altec series crossovers of that vintage had some compensation in them, and users should be aware of that. Using a circuit which does NOT incorporate compensation is a step backward in this respect.

 

[Billfort]

I guess my point right now is that the Altec series crossovers of that vintage had some compensation in them, and users should be aware of that. Using a circuit which does NOT incorporate compensation is a step backward in this respect.

Now there is a point we can solidly agree on.

 

[Bauhausler]

FWIW, I believe these N1600C crossovers are original to these 605A drivers. I opened one up. The caps are marked 16uf and 5uf. The fixed resistors are 4 and 20 Ohms. The pot is 25 Ohms. The choke measures between .97mh and 1mh. My LCR meter isn't really accurate in that low a range.

I followed a different schem and used 4uf, so I guess I'll splice in another 1uf in parallel, bringing the tweeter in 20% lower. That should make an audible difference.
This might be why I've been running them with the pots wide open.

THEN I'll stop fooling with them for a while...

I found a Jethro Tull Best Of record behind the couch and it cleaned up pretty well on the VPI. There must be a turntable around here that will work without heroic efforts at rehabilitation.

 

[Zilch]

FWIW, I believe these N1600C crossovers are original to these 605A drivers. I opened one up. The caps are marked 16uf and 5uf. The fixed resistors are 4 and 20 Ohms. The pot is 25 Ohms. The choke measures between .97mh and 1mh.

Which is connected to the red output, the pot or the 5 uF capacitor?

Oddly, 4 uF doesn't shift the HF highpass frequency much, rather, it provides ~0.5 dB more compensation (2 dB total) and raises the lowpass "Q" about the same amount....

 

[Bauhausler]

Which is connected to the red output, the pot or the 5 uF capacitor?

Oddly, 4 uF doesn't shift the HF highpass frequency much, rather, it provides ~0.5 dB more compensation (2 dB total) and raises the lowpass "Q" about the same amount....

I cracked it open again & had a better look. It's exactly as shown in this pic. Exactly the N1600C circuit posted earlier except 4uF is 5uf and 5 Ohms is 4 Ohms.

 

[Zilch]

Here's the voltage drives as I model them, then, assuming nominal 16-Ohm resistive loads. We don't know the actual impedances or inductances of the drivers. Cursor "a" is at 1.6 kHz. Yellow and green are 4 uF:

 

[Bauhausler]

Here's the voltage drives as I model them, then, assuming nominal 16-Ohm resistive loads. We don't know the actual impedances or inductances of the drivers. Cursor "a" is at 1.6 kHz. Yellow and green are 4 uF:

Wow. Not much difference unless the actual impedances deviate a whole lot from 16 Ohm resistive in the crossover region.

I just took the crossovers apart, changed to 5uf and 4 Ohms like the N1600C's that I have and cleaned up some more of the wiring. I don't hear a colossal difference except now it's a trifle brighter because I had a wrong value R in there for the 20 Ohm.

I still like the overall balance, though I'm struggling to the subwoofer level right. I'll have it right on for one recording, then on the next dishes are falling out of cupboards 50 feet away.

 

[Zilch]

Since you have the iron-core originals, and I presume your clones are made with air cores, it'd be interesting to know if you hear the same difference that Patten did.

Like you, that has not been my experience, so long as the DCRs are close.... :yes:

 

[Bauhausler]

Since you have the iron-core originals, and I presume your clones are made with air cores, it'd be interesting to know if you hear the same difference that Patten did.

Like you, that has not been my experience, so long as the DCRs are close.... :yes:

The clones are made with air cores. I think they're 14ga or better. They're from the AR3 cabs & weigh a ton. DCR is the only significant difference I can think of. The iron cores have much higher gauge (smaller) wire, but of course less of it because of the core. I couldn't measure DCR of the iron core choke because I didn't have a meter handy that measures with DC. For some reason, my LCR meter uses AC to measure R, making it useless for DCR of chokes.
The only other differences could be nonlinearity of the core material, but that should be insignificant below the tens of Watts level. I just can't believe that an iron core would sound *that* different at the fractional Wattage levels I'm using.

 

[Steve O]

The clones are made with air cores. I think they're 14ga or better. They're from the AR3 cabs & weigh a ton. DCR is the only significant difference I can think of. The iron cores have much higher gauge (smaller) wire, but of course less of it because of the core. I couldn't measure DCR of the iron core choke because I didn't have a meter handy that measures with DC. For some reason, my LCR meter uses AC to measure R, making it useless for DCR of chokes.
The only other differences could be nonlinearity of the core material, but that should be insignificant below the tens of Watts level. I just can't believe that an iron core would sound *that* different at the fractional Wattage levels I'm using.

Actually, there may indeed be "non-linearities" in the Altec iron core inductors at low levels. When I was rebuilding N800-F xovers and prior to the schematic(s) showing up, I measured the inductance @ ~3mH using the typical LCR meter. My meter uses a low level 1000Hz signal at this range. When 3mH was plugged into the model, the simulated LF response didn't match the measured LF response when the actual xover was terminated with resistors and driven with a couple of volts (to match constraints of model). The schematic revealed the inductance to be 3.5mH and this value matches measured response more closely. However, best match was when 3.75mH was plugged into the model. This is the value I used when I cloned the N800-F but with air core inductors. Curiously, the inductance of the air core coils as measured with my LCR meter nearly exactly matches the stated values and the measured response overlays the simulated response within measurement error. Therefore, I conclude that the Altec inductors behave differently at really low levels as might be used with an LCR meter or maybe even a bridge compared to typical operating levels.

At the end of the day I have to wonder:
1. Was Altec was aware of this and did they compensate for it in their designs?
2. Was Altec aware of this and chose to ignore it?
3. Was Altec unaware of this?
4. Has the core material somehow aged and no longer performs as designed?

Regardless, it seems that with Altec xovers we're left without much insight as to original intentions, esp subtleties such as the built-in compensation hidden in so many of them. OTOH this may be a good thing because it requires us to reach beyond what Altec could do back in the day and maybe improve things a bit

 

[Bauhausler]

STEVEO;
Thanks for that contribution. I'm glad to know someone else has run into this and given it some thought. I'm aware of inductors having different characteristics under DC current and different levels of excitation. That's well documented in tube amp output chokes and transformers. I just didn't think any of those mechanisms applied here.
Perhaps the thing to do is develop a more subtle test procedure that mimics the inductor's use in circuit. I have basic bench equipment. Maybe I should take a batch of different physical sizes and electrical values of inductors and test them under different conditions.
A really easy test would be resonance frequency with a capacitor (Fc = 1/(6.28sqrt(LC)) I think) versus excitation level. If the functional inductance shifts with excitation level, the resonance frequency will move as well. It's easy to test for and quantify.
Magnetics is not my strong suit. Actually it's as strong as any of my other suits. I'm just weak all over.
You know who would know about this? My brother. He's a materials scientist and once expounded on GOSS's inductive properties at length with no prompting. If he's too busy to give it any thought he'll know a good reference for the information. I'll pester him about this tomorrow.

STEVEO:
You're in SE MI? I'm in Huntington Woods, over by the Zoo...

 

[dlp4341]

Gentlemen:
When I measure crossovers etc , I use the proper load ... a speaker. After all that is the way I listen to Them .
Try it ; run test curves with 8 or 16 Ohm resistors, then do it again with the speaker. This may explain what you are hearing or not hearing
Don

 

[Bauhausler]

Well, today I yanked both crossovers apart and changed the chokes to 1.4mh. I also put a 12 Ohm resistor across. I'm beginning to regret the burst of enthusiasm that caused me to solder these up the first time, as they've been apart 3 times now for changes. Sometimes a rats nest of clipleads is your best temporary option.
I suppose I could try the N1500 crossover for grins. I think I have all of teh parts and another 6 pairs of dual binding posts for in/out connections.

The cap and resistor changes didn't make a huge difference in the sound, but I hadn't listened for more than 10 hours in any configuration. Difficult to tell. Maybe a little more lower mids/upper bass. Maybe not. I also moved them back against the wall at the same time & I'm constantly fooling with the subs to get the balance right. It's no way to do critical evaluation.

I'll leave them up in some state or another for at least a week. Gradually, something might become audible that needs changing. Maybe not. I like the balance a lot as it is.

I also made some nice grilles for the subwoofers to match the 605a cabinets.

 

[Steve O]

Gentlemen:
When I measure crossovers etc , I use the proper load ... a speaker. After all that is the way I listen to Them .
Try it ; run test curves with 8 or 16 Ohm resistors, then do it again with the speaker. This may explain what you are hearing or not hearing
Don

I readily admit to being a speaker designing grasshopper so I probably don’t appreciate all of the finer points of the practice. However, I do find the use of standard loading resistors quite useful for verifying proper electrical operation of a crossover. The primary reason is that a resistor is a standard, predictable and repeatable load compared to a speaker. Therefore, it’s easy to compare the measurements of a physical crossover implementation with the circuit model. I’ve found that models closely predict the electrical performance of a crossover and will clearly reveal basic characteristics. What-if scenarios also facilitated. I recognize there are simulated speaker load designs out there, but it would be pure luck if any of them were to match a particular driver I happen to be working with.

Also, since there is usually little meaningful (to me) info regarding how an electrical input to a speaker translates into real-world acoustic output, I don’t see a lot of value in measuring the electrical output of a crossover with a speaker as the load. Instead, I’ll measure the acoustic output of the speaker and combine this with listening and adjust crossover components accordingly. An example is the acoustic midrange prominence of certain vintage Altec driver/horn combinations when fed a relatively constant voltage input. Measurement of the voltage at the crossover output doesn’t really predict what’s heard or measured acoustically. But, as I note above… I’m no pro at speaker design so if there’s a fundamental flaw in my logic, set me straight.

BTW, I found your website article regarding the time-delayed crossover for the 605A interesting. I would never have guessed that a high level, 8th order, passive LP section could be mated successfully with what appears to be a 1st order HP section (do I have that correct???). Unfortunately, it appears that you’ve encoded the article such that images can’t be printed or saved so a personal copy is a real PITA. Any suggestions?

I’m interested in trying this topology with other Altec duplexes but the physical offset, nominal impedance and xover frequencies are different. I did find that a model of the LP section scales nicely WRT to frequency and impedance but the delay magnitude also moves. Can you point me to any references where I can learn about applying this concept to the general case?

 

[Steve O]

STEVEO;
(snip)

STEVEO:
You're in SE MI? I'm in Huntington Woods, over by the Zoo...

Yes, I'm in SE MI...Waterford to be more specific. To those who don't know the area, I usually say I'm near Clarkston or West Bloomfield.

 

[Bauhausler]

Yes, I'm in SE MI...Waterford to be more specific. To those who don't know the area, I usually say I'm near Clarkston or West Bloomfield.

Small world. Clarkston and West Bloomfield don't have a tight, tree-lined roadracing track though. I like Waterford. I spend a lot of time stomping around Bald Mountain park in the summer, but that and concerts at The Palace are the only times I go North these days.

I don't know what I'll do 'permanently' with the 605A crossovers. For now I'm listening to them with the slightly modded N1600C version & it's fine. So far these are beating out the Valencias for a few reasons: footprint, bass extension with easily implemented subs, looks, height, placement versatility, imaging. The Cornwalls got stomped in round one of the competition, but they'll make someone else happy. I'm listening to the great Joe Strummer & The Mescaleros CD Global a Go Go right now.

I had an evil idea the other night. I have a single JBL C40 Harkness cab. It has a D130 in it. I wonder how a 605 would work in that....

 

[dlp4341]

Steve:
My tech ref is : STANDARD HANDBOOK FOR ELECTRICAL ENGINEERS 12th (>10) EDITION (1987) Edited by DONALD FINK. Chapter 2, FILTERS. It also quotes the work of ZOBEL in 1923. I use a filter source Z= 0 and load Z= 13.
To find the delay required, use the pulse and pickup the acoustical signal with an omni mic 12" on axis of the Duplex. Measure the time on a oscope or Mac/PC and software.
Re the web : all I did was use the Mac Web setup. Sorry I can't help how it came out
Don.

 

[Zilch]

Re the web : all I did was use the Mac Web setup. Sorry I can't help how it came out.

I requested above that you upload here the stock 605A response and voltage drives, the last two images on that page.

I cannot read the scales or legends, and it's apparent that the stock compensation is greater than shown in my sims.

Would you do that for the benefit of this thread, in a larger size, please, Don?