Jump to content

Subjective testing of AR-11 speakers


r_laski

Recommended Posts

Subjective testing of AR-11 speakers before and after capacitor replacement – an experiment.

I’ve been following the recent threads on capacitor replacement on this forum. One of the questions expressed by forum members as they contemplate capacitor replacement in a classic AR speaker is “will I screw up my speakers?”

Some have suggested the desire/need for objective (lab) testing of repaired / restored / modified speakers vis-a-vis an all original “brand new” classic speaker. I believe it is generally agreed that finding a classic AR speaker in “Brand New” condition as a benchmark speaker is practically impossible, given what we know about foam surround deterioration, capacitor aging even while sitting idle, and the as yet unexplored changes to a particular driver’s performance over time.

What I have attempted over the past few weeks is a purely subjective (no cost, listening) test of three AR-11 speakers. The primary goal was to see if I could discern audible differences between three AR-11 speakers. Ideally this type of testing would be conducted as double blind listening tests with one person controlling the stereo system and the other person doing the listening. Unfortunately, I didn’t have a volunteer to be one of the “blind persons” playing or listening to repeated playing of short (5-30 second) clips of music a couple hours a day for three weeks.

I have two pair of AR-11s with me right now. My Dad’s AR-11s are very early AR-11 models (Brass logos). Mine are AR-11Bs (Aluminum logos). Except for having the woofers re-foamed (by a speaker repair shop in Omaha), my Dad’s AR-11 drivers are the original drivers as installed by AR. The tweeters on these speakers never had diffraction foam. The foam grills on these speakers disintegrated a long time ago. During testing, I installed my genuine AR-11 foam grills on the speakers being compared.

To prepare for this test I measured the capacitance of all 120uF, 40uF, and 10uF capacitors I’d removed from AR-11s and AR-90s. I was very fortunate in that I found a 120uF Callins cap that measured 119.7uF, a 40uF that measured 39.5uF, and a 10uF that measured 9.91uF. For this test, I consider these to be “original spec” AR capacitors.

Speaker 1: One of my Dad’s AR-11s. I first measured and then removed the three capacitors and replaced them with the “original spec” capacitors. The capacitors I removed were either very degraded/failed, (120uF Callins cap started measuring over 144uF and leaked badly – the reading continuously climbed. I stopped measuring after 155uF) or were not as close to stated value as the “original spec” caps (40uF cap measured 44.1uF and 10uF cap measured 10.45uF). When I installed the three “original spec” capacitors, I duplicated the original installation as precisely as I could – lead attachments to terminals, soldering, etc. Given available resources, this was my “best effort possible” attempt to create as close a possible, an original spec, all original AR-11.

Granted, the caps I installed have aged and the woofer is re-foamed. All I ask is that even the staunchest AR purist allow me a little latitude in establishing a benchmark speaker. For this test, I designated this speaker AR-11A.

Speaker 2: One of my AR-11Bs. This speaker has all North Creek Zen polyprop capacitors, North Creek resistors, and new wires. For this test, I designated this speaker AR-11B.

Speaker 3: My Dad’s other AR-11. This speaker has the White/yellow tweeter dome and the yellow fiberglass, black wire mesh over the midrange (AR-3A carryover). This is a very early AR-11 – SN 03219. The woofer has been re-foamed (by a speaker repair shop in Omaha). Other than that, this speaker is in the same condition as when it was purchased. I didn’t even measure the condition of the crossover components prior to testing. For this test, I designated this speaker AR-11X.

For all tests, attenuation switches on all speakers were set to flat (0 dB) position.

Objectives:

1) Determine if there is an audible difference between the benchmark speaker and an “aged” speaker.

2) Determine if there is an audible difference between the benchmark speaker and a repaired speaker with polypropylene capacitors.

Listening Tests, Part I:

For the first portion of this test, I placed two speakers on 7 inch stands side by side, about one inch between cabinets, directly in front of my listening position. I conducted many hours of listening tests over several days, comparing two speakers at a time.

These tests were done by setting the preamp to mono. I sent the same source material to both speakers using the balance control to “switch” between speakers – no extraneous speaker switching components in the signal path. I played short clips of music from several CDs, choosing content that contained a limited number of instruments so I could concentrate on comparing the sound of specific instruments played alternately on both speakers. Listening position was varied from as close as one foot to ten feet in front of the speaker being played*.

*Something I “discovered” - due to the horizontal alignment of the midrange and tweeter, you could not listen while positioned centered between the two speakers, when they are this close together. As I moved closer to the speakers, the sound of each speaker varied audibly as I varied my position (positioned closer to either the midrange of the left or tweeter of the right speaker). Therefore, for each sound clip, I positioned myself centered directly in front of the speaker being played.

Listening Tests, Part II:

I moved the speakers to where I normally place them in the room for stereo. I played source material in both mono and stereo. To ensure room acoustics were not causing or contributing to the difference in what I was hearing, I played each speaker in both the right and left position.

Results:

1) AR-11A vs. AR-11X –

Test 1: Very slight and subtle difference between the two. There is a slight difference in pitch, with the AR-11X sounding sharper (higher pitch). As an example, cymbals sound like brass on the AR-11A, they sound “tinny” with the AR-11X.

Test 2: The difference in pitch was more apparent in these tests. Also, the sound is slightly “unbalanced” between the two speakers. There is more left speaker / right speaker than a “soundstage” across the room between them.

I did some listening to vocals – mostly Norah Jones. In stereo, I did not get a focused vocal. There were several times when I got the impression there were two voices (pitch difference) instead of one.

What was also audible in both tests when comparing in mono is a difference in separation between instruments. With the AR-11A there is an audible “space” between the instruments and each instrument is clearly heard. With the AR-11X the instruments sound like they are on top of each other. For example, with guitar and snare drums it is very difficult to hear both instruments individually.

2) AR-11B vs. AR-11X –

Test 1and Test 2: Same as 1) above. No surprise here – I was hearing very much the same differences as when I conducted the same tests after I built a prototype replacement crossover for my AR-11s.

One other thing about the comparisons above. Although the difference between the “voices” of these speakers is very slight and subtle, when played together in stereo for an extended period, the difference became “annoying” to the ears and listener fatigue sets in. I experienced the same sensation with my AR-11s and a pair of AR-90s before I repaired them. It is the main reason I decided to repair/restore them.

3) AR-11A vs. AR-11B.

Test 1: Probably the most surprising results of all were in the comparison of these two speakers. I had a very difficult time discerning any audible difference between these two speakers. Whatever differences I “heard” are so slight that I believe if I had someone come in and set the system up so I didn’t know which speaker was which and was blindfolded so I couldn’t see the speakers, only listen to them, I could not tell them apart.

Test 2: WOW! These two speakers together sound identical to each other and are fantastic together. There doesn’t appear to be any difference in relative loudness or pitch between these two speakers. The sound stage across the room and accuracy (clarity) of individual instruments across the soundstage has returned. Norah Jones’ voice is tightly focused and sparkles like crystal.

Final Measurement:

AR-11X capacitors: This was the last thing I did as part of this test. More surprises. This speaker has 72uF and 50uF capacitors paralleled for the woofer (as shown on the original AR-11 schematic on the Classic Speaker Pages). Also, there are two (2) Sprague 20uF caps paralleled for the midrange capacitor. These caps have AR part numbers on them and are identical to another pair of Sprague 20uF caps I pulled out of an AR-LST.

Woofer cap – 122uF measured 134uF (9.8% over)

Midrange cap – 40uF measured 44.9uF (12.25% over)

Tweeter cap – 10uF measured 13.72uF (37.2% over)

Conclusions:

Objective 1): Benchmark speaker vs. “aged” speaker: Capacitors, particularly electrolytics, do degrade over time and this degradation is audible.

Objective 2): Benchmark speaker vs. repaired speaker: I am very surprised how virtually identical the “voices” of these two speakers are. Especially since one speaker has all “original” AR components and the other has many new parts and polypropylene capacitors.

While it would be nice to have a lab test these speakers, it would take someone with deeper pockets than mine to contract with a lab to conduct the extensive testing needed to quantify changes to a speaker’s performance before and after repair / restoration to compare to any documented measurements, either from AR or a reputable equipment testing lab, made when the AR-11 speaker was first introduced. Besides, I spent too much time and effort restoring the cabinets to “showroom new” condition to have them banged up in transit or by a lab tech.

When all the lab tests are done, it still comes down to how the speakers sound to the individual anyway. I’m convinced, based on what I hear, that aging capacitors whose capacitance value have changed, do alter the voice of a speaker over time.

Replacing aged and degraded crossover capacitors with either new electrolytic or polypropylene capacitors of the same capacitance value is not an unpardonable "acoustical" sin. It does not significantly alter the voice of the speaker. Rather, it returns it to a near perfect restoration of its original voice. While it may not be exactly the “voice” of a brand new AR-11, to me, the “voice” of my repaired/ restored AR-11s is exactly “right” when compared to the closest thing to an original AR-11 that I can construct.

Rich

Link to comment
Share on other sites

Guest Bret

Rich, that was a lot of testing you did there and a lot of work to have gone to.

I appreciate it. I suspect that you suspected what I suspected which is; of all the important specifications that capacitors might have, capacitance is chief among them.

I suppose that's why they aren't called resistors.

But you've given me additional courage to do what I suppose needs doing.

Bret

Link to comment
Share on other sites

Bret,

It was a labor of love – of music and AR loudspeakers. I still have the AR-11A and AR-11B hooked up and can’t get over how perfectly matched they are and how incredible they sound together.

>> I appreciate it. I suspect that you suspected what I suspected which is; of all the important specifications that capacitors might have, capacitance is chief among them.<<

I agree with you that when choosing replacement capacitors for speaker crossovers, matching the capacitance value is most important. While a capacitor’s Equivalent Series Resistance (ESR) is an important consideration, when you are working with the resistance of a speaker’s voice coil which varies with frequency and temperature, and resistors with 10% tolerance, the difference in ESR of an electrolytic vs. polypropylene capacitor, or a new vs. old electrolytic capacitor is less significant (mathematically speaking) than a difference in capacitance value in a speaker crossover network.

I believe capacitor ESR would be a much more important consideration if we were discussing a more sophisticated circuit such as a solid state preamp or amplifier.

>>But you've given me additional courage to do what I suppose needs doing.<<

Then my experiment was a worthwhile success for both of us. Good Luck with your projects and let us know how they turn out. I now need to get started repairing / restoring my Dad’s AR-11s.

Rich

Link to comment
Share on other sites

  • 2 weeks later...

Ken,

>>How did you test the caps? (What frequency and level?)<<

To test the caps I either remove them from the crossover or de-solder leads to create an open circuit at the capacitor to disable the high/low pass filter. I use a Mastech 390-735 capacitance meter (Parts Express) to measure capacitance. From the user’s manual: the test frequency for the 20uF range is 80Hz and for 200uF range it is 8Hz. The excitation voltage is 2.8V peak.

>>I personally doubt that a 10% difference is going to prove very audible in truly controlled listening tests, but a 30+% difference would be.<<

I agree. During my listening tests, I could not clearly or consistently identify an audible difference in the bass response between the three speakers, where the difference between the capacitors was slightly more than 10%. Even though they measure high, the paralleled caps in the AR-11X were “in spec” (within 10% of rated capacitance).

However, to me, the differences I “heard” were in both the midrange and the high range. The difference in pitch that I described was more audible in the upper frequencies. The smearing or overhang of individual notes from instruments was present from the midrange up.

I would say that the “threshold of audibility” is somewhere between 10% and 15% difference in measured capacitance. Maybe a hypothesis for a future experiment under more controlled conditions with more than one listener.

Link to comment
Share on other sites

One would hope that a 10% capacitor tolerance would not produce a difference between speakers, as that is a typical mfg. spec. ESR also increases with age in an electrolytic and one effect of increased ESR is to decrease the crossover frequency. Increasing C also decreases the crossover frequency so the two act in concert.

I have access to an excellent LCR meter, and will be most happy to measure the caps you removed-am doing that for another as we speak. Send me an e-mail and I will give you my home address.

Cheers,

Link to comment
Share on other sites

  • 9 months later...

Two AR11 crossovers.

I recently finished rebuilding the crossovers of my Dad’s AR11’s. These are the speakers I used in the subjective listening testing I did last year.

Something I found interesting –

One of these speakers had a different crossover from the other AR11s I’d worked on, like this one:

http://www.classicspeakerpages.net/dc/user_files/335.jpg

Instead of being all on one board, the crossover is split between a board on the back and one on the bottom of the cabinet:

http://www.classicspeakerpages.net/dc/user_files/336.jpg

This crossover reminds me of the AR LST crossover the way it is laid out with all the connections to a terminal strip. This crossover more closely matches the AR11 crossover schematic Tom Tyson provided to the archives. It is the earlier (older) of the two crossovers.

Note the resistors in the equalization network on the back board. Both switches have a 3 Ohm and a 1.5 Ohm resistor. The pair of 20UF Sprague caps are paralleled to make a 40UF midrange cap. The two paralleled caps at the top of the picture are a 50UF and a 72UF cap for the woofer. There is no fuse in the crossover on this speaker as shown in the AR-11 crossover schematic.

Rich

post-100657-1111446808.jpg

post-3-1111446808.jpg

post-3-1111446809.jpg

Link to comment
Share on other sites

Rich,

Great pictures, thank you !

I looked and then answered Roy's question about the crossover schematic (as best I could) and walked away to do something completely different when I was stopped in my tracks by a "Yeahbut."

The oldest crossover did receive bad marks from you. Now, I can hardly wait for someone to come along and call me an idiot and tell me to shut-up, but if ESR in capacitors works just like resistance in resistors, then shouldn't two caps parallelled have half the ESR in-circuit as one cap of equal ESR and capacitance? Two "low ESR" 20uF caps parallelled ought to have lower combined ESR than one "low ESR" 40uF cap, right?

So, wouldn't the ESR of the parallelled caps have to be twice the original ESR to sound like ONE cap if ESR made any difference????

IF that's true, as the parallelled caps aged and their ESRs rose wouldn't the net rise be fractional? Shouldn't those two capacitors have lower ESR together than either of them individually? Shouldn't it have always been thus, even when they were new?

And if changing ESR changes ANYthing in the crossover's design, if it were taken into account AT ALL wouldn't there be a difference in inductors or resistors or SOMEthing to compensate for the parallelled capacitors???? I mean, if ESR is important to the design, surely changing a 50uF and 72uF parallelled cap would dictate changes to the crossover in order to accomodate the use of a new, single 120uF cap. . . right or wrong?

If, and it's a big IF, anything I just said is true, then the differences you were hearing in the untouched speaker vs the in-spec old capacitored speaker almost can't have been ESR because it is almost a given that the "old but in-spec" caps' ESR has risen, too. I know you didn't measure for it, but it's implied by the paralleling of everything but the 10uF isn't it?

Or are you using John's old caps and he HAS a reading for ESR on them? If so it would be fascinating if he could / would check the ESR of the parallelled caps. If we find that they have lower ESR than the "in-spec" ones

Absorption, dissapation factor. . . something else is audible but either ESR is a throw-away spec or AR (being a major manufacturer) used the parts it had even though substituting changed the crossover and changed the specs on its speakers.

OR the crossover did change. Are there different inductors in them? Different anything?

Suddenly I'm not as worried about matching ESR as I was an hour ago. I suspect it was ignored and the choice of "Compulytic low ESR" capacitors was to get some other special spec. Like "ripple current" measures (whatever that does), or speed, or longevity or . . . something. I'm beginning to smell a flawed assumption in this whole "low ESR" thing: Low ESR as a by-product, not a design element.

Bret

Link to comment
Share on other sites

Bret,

>>Now, I can hardly wait for someone to come along and call me an idiot and tell me to shut-up,…<<

Okay, if you insist, “you’re and idiot and shut-up!” :7 (LoL)

I’m not even going to try to answer all your questions on ESR.

>>The oldest crossover did receive bad marks from you.<<

Yes. As far as the test goes it did not fair well. I noted that the paralleled woofer caps were marginally “in spec” and that I could not discern an audible difference in bass response between the speakers during the test.

However, NOT part of the test was a comparison of “AR-11A” to the other speakers before I replaced the caps to make it “AR-11A”. Briefly, AR-11A and AR-11X were a close “voice match” before the test. They did not “voice match” AR-11B. During the test AR-11A and AR-11B “became voice matched,” AR-11X did not “voice match” either speaker. So, my point was, over time, the caps age and the speaker’s “voice” changes. New (good quality) caps can restore the voice to original or as near original as I could make an AR11. I did the test for my peace of mind that spending $200 + to rebuild a pair of AR11 crossovers were restoring not just changing the speakers.

>>And if changing ESR changes ANYthing in the crossover's design, if it were taken into account AT ALL wouldn't there be a difference in inductors or resistors or SOMEthing to compensate for the paralleled capacitors???? I mean, if ESR is important to the design, surely changing a 50uF and 72uF paralleled cap would dictate changes to the crossover in order to accommodate the use of a new, single 120uF cap. . . right or wrong?<<

There is NO difference in the woofer inductors. There are NO resistors in the woofer network. The crossover frequency remained at 525 Hz.

>> OR the crossover did change. Are there different inductors in them? Different anything?<<

The resistors in the equalization network are different, BUT, the entire switching network is different. However, with the switches at 0dB, my test setting, the resistors are switched out of the circuit. The only other resistor in the crossover is a 15 Ohm resistor in the midrange. It is the same in both crossovers.

>> Or are you using John's old caps and he HAS a reading for ESR on them? If so it would be fascinating if he could / would check the ESR of the paralleled caps. If we find that they have lower ESR than the "in-spec" ones<<

I used capacitors I had on hand from AR11 and AR90 crossovers. Right after I finished my testing I had to put my audio projects on hold for several months while I completed some home renovation / remodeling projects. I just recently (last moth) rebuilt these AR11s.

I did not ignore John’s offer to measure the capacitors. I just had to set the entire project aside for several months. I would like to have him measure these caps if the offer still holds.

>> Absorption, dissipation factor. . . something else is audible but either ESR is a throw-away spec or AR (being a major manufacturer) used the parts it had even though substituting changed the crossover and changed the specs on its speakers.<<

It is obvious that Teledyne / AR changed vendors for its capacitors. This is probably about the time Chicago Industrial and Sprague went out of business or were bought out. AR never changed the published specs of the AR11. AR did not manufacture and sell AR11s as matched pairs / consecutive serial numbers only.

My advice, don’t give yourself a wedgee over ESR. While you need to be aware of ESR, it’s more important to match capacitance. If you were building exotic preamps, amps, SACD / DVD-A transports or highly accurate test equipment then you need to fret about ESR, etc.

Consider this: The 15 Ohm, 10% resistor in the midrange can be anywhere between 13.5 Ohms and 16.5 Ohms and still be in spec. The 40uF, 10% capacitor can be anywhere between 36uF and 44uF and still be in spec. That takes care of mille- a micro- Ohm differences in capacitor ESR, length of capacitor, inductor, and resistor leads, quality of solder connections, connectors, contacts, etc, etc, etc…

Rich

Link to comment
Share on other sites

Rich,

What is considered to be the "optimum" switch setting for the AR-11's? Is it the setting that bypasses all of the series resistors in the tweeter and mid circuits?

What is your preferred setting?

thanks,

Roy

Link to comment
Share on other sites

>My advice, don’t give yourself a wedgee over ESR.<

DARN ! And here I thought I had a wedgee-worthy subject . . .

Well, what I was really up to was reiterating support for my much earlier, very provocative, terribly controversial, highly stylistic, and completely groundless wild guess that ESR was not a crossover parameter that was designed-around, taken into account, or otherwise originally given a flying rat's-wedgee about.

Being willing to inflict my own agony of defeat in the name of full-disclosure: I suppose it could have been designed-for then been totally ignored in order to use-up a supply of parts.

Personally, henceforth I vow to ignore it entirely.

Have you noticed that the 303 has a mylar cap on the tweeter? Wonder if we could get Ken to explain how that was arrived-at? Would be educational.

Now, call me a taxi. . .

Bret

Link to comment
Share on other sites

The AR-3a Limited crossover of the early 90's has a mylar cap on the tweeter as well, the rest being electrolytics. (That crossover is very much like the AR-11 with 8 ohm l-pads instead of "resistor" switches.)

Back in the days when I took a stab at comparing caps, I always thought I liked the sound of the mylars better....or maybe it was their color..I really liked the yellow ones :-).

Don't be too hard on yourself Bret, we're dealing with one of the great mysteries (myths?) of the universe here.

Roy

Link to comment
Share on other sites

>I did not ignore John’s offer to measure the capacitors. I

>just had to set the entire project aside for several months. I

>would like to have him measure these caps if the offer still

>holds.

It holds: please e-mail for postal address

ohanlon at ece dot arizona dot edu

john

Link to comment
Share on other sites

Roy,

>> What is considered to be the "optimum" switch setting for the AR-11's? Is it the setting that bypasses all of the series resistors in the tweeter and mid circuits?<<

In the AR ADD/Truth in Listening Owner’s Instruction Manual there are 2 pages on placement in a room and room acoustics and another 2 2/3 pages (not counting large pictures of switches / binding posts) on adjusting the midrange and highrange level controls.

Here is what AR said about the level controls:

“In most rooms and situations in which there is normal acoustical absorption and treble diffusion effected by drapes, shelves, book cases, furniture, etc., and where associated equipment and source material of very high quality is being used, the switch or switches should be placed in the “0” position.” 0dB is also the position AR specifies you start at when adjusting the level controls.

>>What is your preferred setting?<<

I have found that in almost every room I’ve had AR-11s in (carpeted floor, drapes, book cases, furniture, etc) the 0dB positions provide the best overall balance to my ears. In a couple of rooms that were overly bright, I had to tone down the highrange to -3dB.

A couple of other noteworthy points –

-- AR also stated, “These level controls are designed to permit matching of the midrange and treble of the speaker to your individual environment and are meant as important adjuncts to (to be used in conjunction with) the tone controls in your receiver or amplifier.”

One could draw the logical conclusion that AR supported the use of tone controls, INCLUDING EQUALIZERS, to fine tune their speakers to each individual room’s acoustics. The level controls were provided as a “coarse” adjustment to get you close and your tone controls / equalizer make the “fine” adjustments to get just the right balance for each placement.

-- The adjustment procedure require adjusting each speaker separately by sending a mono signal to one speaker (using mono or A+B and balance control) and conducting lengthy listening comparisons with the switches in the 0dB position and each of the other two positions. For the AR11, with separate midrange and highrange controls, they recommend setting the midrange level first and then the highrange level.

-- Make your tone control / equalizer adjustments after you return to stereo mode.

-- AR noted it is NORMAL for the midrange and highrange switches to NOT be in the same position on both speakers, “… due to the differences in the immediate acoustical environment to the speaker(s).”

Rich

Link to comment
Share on other sites

Very nice work Rich, I enjoyed reading about your work and seeing a rational approach to this question. I agree capacitance value is most important and ESR is sometimes important it depends we cannot make generalizations. I'll try to comment further when I get a chance.

Pete B.

>Subjective testing of AR-11 speakers before and after

>capacitor replacement – an experiment.

Link to comment
Share on other sites

  • 1 month later...

I sent all the capacitors I removed from AR speakers to John so he could measure them for Capacitance © and Dissipation Factor (DF) at different frequencies (f). He also measured Dielectric Absorption (DA). I must give John all the credit for his extensive measuring, research, and analysis on these capacitors. What he found may surprise some, but should give all cause for concern to have their capacitors inspected and measured, and replace all aged, degraded, damaged Non-Polar Electrolytic (NPE) capacitors. Below are spreadsheets of raw data capacitor measurements for my capacitors and John’s capacitors removed from pairs of AR-18bs and AR-11s.

http://www.classicspeakerpages.net/dc/user_files/428.jpg

http://www.classicspeakerpages.net/dc/user_files/431.jpg

http://www.classicspeakerpages.net/dc/user_files/430.jpg

John’s Analysis:

Capacitors Enclosed in Aluminum Cans:

(Sprague, Callins and Industrial Condenser Corporation).

· The cans were clearly marked “np.”

· C and DF versus f were made on an HP-4246A LCR meter at three frequencies. DF was high as expected for large NPE capacitors. All capacitors eventually go inductive, but electrolytics do so at lower frequencies than other capacitor types due to winding and position of lead wire attachment.

· An HP-4284A LCR meter was used to measure C at various frequencies for a 350- and a 120-uF cap (limited access to meter). See attached graph below. In those cases, the cap was subjected to a 5-V dc pulse for 1 second before each measurement to simulate a large, single woofer excursion. C was measured within a few cycles after the pulse ended. Notice the device goes inductive at a lower f than without a pulse, because the electrolyte is flipped +/- quickly. During that 1 sec it cannot completely reform its oxide; both electrodes can conduct. E.g. it becomes a resistor at lower frequency for a very short time after large sudden voltage excursions.

http://www.classicspeakerpages.net/dc/user_files/432.jpg

· The capacitors used a four-layer composite winding of (Al-paper-Al-paper) in a tight roll. See picture below. The 65-um-thick paper layers were saturated with an (oxalic acid) electrolyte. The lead wires were attached at one end of each foil. Thus, both sides of each foil are anodized and the two foils in series form the back-to-back, or NPE structure. The large roll was potted (same material as used in transformers) within the aluminum can.

http://www.classicspeakerpages.net/dc/user_files/433.jpg

Summary (Al can):

· The poor high frequency response (very high inductance,) is primarily due to the leads being connected at the end of the wound foil and because the oxide cannot reform quickly after a fast voltage pulse. NPEs are not capacitors at all frequencies.

Recommendation (Al can):

· Replace woofer crossover caps with something like a Solen polypropylene.

Tubular, PVC-Encased Capacitors (Black sleeve, Red ends)

(Callins)

· All the black/red caps (4—40 uF) were aluminum foil NPE. The imprinted “50 VAC” should have told me right away; they had the same inner construction as aluminum can types (See picture below). Ten uF and smaller used smooth Al foil electrodes, whereas 24-uF and larger used etched Al foil for increased surface area (up to 4x), which yielded increased capacitance/volume.

http://www.classicspeakerpages.net/dc/user_files/434.jpg

· All had one wire contact per foil with the contact located near the center of each 1.5-m-long strip. The reduced inductance helped to keep the capacitance constant to over 10 kHz.

· Measured C remained rather constant; DF increased rapidly above 1,000 Hz.

· The errant caps from Rich’s AR-11s seemed to have clustered dates of manufacture (1068 yyww). This is because the four AR-11 speakers from which the caps were removed were assembled near the end of 1977. All the bad capacitors were not isolated to one speaker or one pair of speakers. The common element appears to be the capacitor mounting technique -- the use of hot glue to secure capacitors to crossover boards. John O concludes that the bad caps were a result of an overzealous operator’s *liberal* use of hot glue near the red end potting; some were encased in **huge** globs of hot glue. It is guessed that the rapid heating near the end of the capacitor caused the seal between the red end plug and sleeve to delaminate in some areas. The bad caps all had one region where the sleeve popped free from the red plug. This leak would have allowed slow out-diffusion of water, which would have slowly increased electrolyte strength. Oxide growth and dissolution compete. At increased electrolyte concentration, dissolution increases, resulting in thin oxide and increased capacitance. This is consistent with the reduced thickness of the Al foil in the “bad” caps. The metal thickness decreased from 48 um (good) to 42 um (bad) in a 6uF cap. For the 10 uF: 42 um (good) to 36 um (bad), as the simultaneous oxidation and dissolution slowly consumed the metal foil.

Summary (Black/Red Tubular):

· Some of Rich’s concerns resulted from listening to a speaker with a bad crossover cap. Ex: an AR-11 tweeter using a 10-uF cap measuring 20-uF would x-o at 2,500- not 5,000-Hz. Likely, neither the driver nor the listener were happy.

Recommendation (Black/Red Tubular):

· Replace all mid- and hi-range crossover caps with high-quality polypropylene film caps (e.g. Zen). Perhaps bypassed with a film and foil capacitor.

· Don’t use hot glue to secure capacitors to a crossover board.

Rich’s Comments on Quantitative Measurement vs. Subjective Testing:

AR-11A vs. AR-11X:

· 10uF capacitors - Stable C vs. f throughout 80 Hz – 10KHz range. AR-11A DF is substantially lower than AR-11X capacitor. Although DA for both caps is very high, Callins (AR-11A) has lower DA.

· 40uF capacitors - Both paralleled Sprague and single Callins measured similar across the freq range. DF readings are not significantly different. DA is better for the paralleled Spragues, but both are within an expected range. Only significant difference between these capacitors is a consistent 10% higher C vs. f for the Sprague pair over the Callins. This may account for some audible difference.

· 120uF capacitors - No trends or correlations between paralleled caps in AR-11X and single Callins in AR-11A. Largest differences in C occur at the lowest and highest freqs. Higher DA of Callins cap may be significant. These widely varying readings may indicate why it was hard to “hear” audible difference in bass region.

AR-11A vs. AR-11B:

· I didn’t send the Zen capacitors to John. Instead he measured one of his Zen caps as a representative sample. Readings indicate exceptional quality of new polypropylene vs. “old, but in spec” NPE capacitor.

· Measurements indicate one would expect to hear more of an audible difference between AR-11A and AR-11B. I’m at a loss to explain why I didn’t “hear” it. I suspect however, that if I had conducted the listening test over a longer period of time more differences would have become audible. I do know that over time the AR-11s with Zen caps are just a joy to listen to.

Conclusions:

· Difficult to quantify subjective results with just capacitor measurements.

· Measurements do indicate capacitor aging / degrading at varying rates over time.

· Measurement and forensic analysis (taking capacitor apart) revealed probable cause of why some capacitors measured so far off: Damage to end cap seal caused by hot glue during assembly of crossover.

John and Rich -- a collaborative effort

post-100657-1116763437.jpg

post-3-1116763437.jpg

post-3-1116763438.jpg

post-3-1116763439.jpg

Link to comment
Share on other sites

Guest Bret

>overzealous operator’s *liberal* use of hot glue near the red end potting; some were encased in **huge** globs of hot glue. It is guessed that the rapid heating near the end of the capacitor caused the seal between the red end plug and sleeve to delaminate in some areas.<

Great Scott! Good job gentlemen.

Bret

http://www.classicspeakerpages.net/dc/user_files/435.jpg

post-100690-1116785107.jpg

Link to comment
Share on other sites

Guest Bret

>It is guessed that the rapid heating near the end of the capacitor caused the seal between the red end plug and sleeve to delaminate in some areas.<

Uh oh.

Maybe there's more to the story.

I replaced the 8uF caps today. One of them had an end just buried in hot glue. It measures 8.7uF.

Here's something else to chew-on a bit.

This cap was 8.7uF, the one in the other speaker measured 8.6uF.

Changing the 8.6uF capacitor did not produce much difference in the sound of the cabinet. Replacing the 8.7uF capacitor made a largish difference in the sound of that cabinet (thank goodness, I was afraid I had a bad driver).

Weird, huh?

Bret

Link to comment
Share on other sites

Guest Americain

Wow. An absolutely fascinating bit of experimentation here Rich. What a great read. I do have one question though...when you use the term "pitch" would that be what most of recognize as "timbre" or do you literally mean pitch? When I think of pitch I think of specific notes but timbre would be something altogether different. I hope I haven't added confusion here.

Thom

Link to comment
Share on other sites

Thom,

Thank you.

>>I do have one question though...when you use the term "pitch" would that be what most of recognize as "timbre" or do you literally mean pitch? When I think of pitch I think of specific notes but timbre would be something altogether different. I hope I haven't added confusion here.<<

“There is a slight difference in pitch, with the AR-11X sounding sharper (higher pitch). As an example, cymbals sound like brass on the AR-11A, they sound “tinny” with the AR-11X.”

A very good question. I may have been the one to add confusion. The example I used in the quote above is more of an example of “timbre,” or quality of the notes of the cymbals. Also, when I said, “With the AR-11X the instruments sound like they are on top of each other,” you could take that to mean the speaker exhibited the inability to accurately reproduce the distinctive sound of the two instruments (timbre).

I also heard a difference in pitch with the AR-11X, particularly in vocals. I did give a better example of this later on in my posting: “There were several times when I got the impression there were two voices (pitch difference) instead of one.” With the “two voices,” one was “natural,” the other was “sharp” (pitch).

I hope this answers your question and clears up any confusion. I did not distinguish between the two, but was actually talking about both.

Rich

Link to comment
Share on other sites

Bret,

>>Maybe there's more to the story.<<

During the course of getting the capacitor measurements, our attention turned to finding out why particular capacitors were abnormally far off from labeled values. We then concentrated on these capacitors and explored several theories. It wasn’t until John conducted the forensic analysis that he found the problem with hot glue and capacitor end caps.

Since your capacitors are marginally “in spec” I would NOT expect hot glue damage to be the cause of the poor sound you were experiencing.

It’s hard to tell from your description what was going on with your 8uF capacitors.

If I had to make a SWAG at it, based on the information you provided, my initial thought was a cold solder connection on one or both leads of the 8.7uF capacitor. Something you would not look for, but cure by changing capacitors, assuming you made a good solder connection with the replacement capacitor.

A while back I stumbled on a discussion group for Chevy S-10 pickups. I was trying to find something that would tell me if my problem - erratic windshield wiper control - was in the wiper motor or the switch on the steering column. I found some messages about a severe problem with cold solder connections on the PC board in the wiper motor. I removed the PC board and with a magnifying glass found seven (7) cold solder connections on my board (some were capacitors). I re-soldered the bad connections and have not had any problems with the windshield wipers since. Saved me a bunch of money.

Rich

Link to comment
Share on other sites

>

>Since your capacitors are marginally “in spec” I would NOT

>expect hot glue damage to be the cause of the poor sound you

>were experiencing.

>

>It’s hard to tell from your description what was going on with

>your 8uF capacitors.

>

>If I had to make a SWAG at it, based on the information you

>provided, my initial thought was a cold solder connection on

>one or both leads of the 8.7uF capacitor. Something you would

>not look for, but cure by changing capacitors, assuming you

>made a good solder connection with the replacement capacitor.

>

Rich,

You bring up an excellent point on cold-solder connections! Although speaker manufacturing became increasingly mechanized as time went on, the crossover networks were basically always hand-wired and soldered. A bad day for a production-line worker might result in a batch of poorly soldered connections, many of which would not show up until quite some time later. Quality-control sweep frequency-response testing may or may not reveal this problem. This is probably more common that we would imagine.

I had a pair of Allison: Ones back in the early 1980s with a strange distortion is the midrange, and I went through the motions of swapping parts and drivers and so forth, and was about to give up when I decided to carefully inspect the crossover. Sure enough, there were numerous cold-solder wiring connections that were making intermittant contact, causing the problems. I went through the entire crossover and re-soldered every connection, and the problems went away.

--Tom Tyson

Link to comment
Share on other sites

  • 7 months later...

I was wondering Rich if you don't mind informing us as to what power amplifier and speaker cables you used?

Pete B.

>Subjective testing of AR-11 speakers before and after

>capacitor replacement – an experiment.

>

>I’ve been following the recent threads on capacitor

>replacement on this forum. One of the questions expressed by

>forum members as they contemplate capacitor replacement in a

>classic AR speaker is “will I screw up my speakers?”

Link to comment
Share on other sites

>> I was wondering Rich if you don't mind informing us as to what power amplifier and speaker cables you used? <<

Power Amplifiers: A pair of Leach Super Amps

Speaker Cable: Straight Wire “Waveguide 8”

The cables were made by Straight Wire and are composed of 8 individually insulated conductors - 4 gray and 4 white for positive and negative. The conductors are braided inside a clear jacket.

Looking at the speaker cables on the Straight Wire web site, (straightwire.com) they look like an early version (if I remember correctly, the original version) of their Stage 2 cables. These do not have a "non-conductive core".

Rich

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...