What type of capacitors should I put in my AR-90s?

[Kuja]

Hello,

I want to recap crossovers in my "new" AR-90s.

In my previous restorations I used polypropylene caps manufactured by ICEL in Italy.

This is a very popular audiophile brand in my part of the world.

BUT...

I run across some threads saying that modern poly caps are not good, since they will change the voicing of old AR speakers.

The theory is that poly caps have lower ESR, and that using them will result in speakers having more treble and upper frequencies.

Did anybody confirm that in real life, or is it just hear-say?

I would like to be able to use poly caps due to their long life.

Any thoughts?

Thanks in advance,

Aleksandar

[Carlspeak]

Well, first you say you have used film caps in previous restorations. Then you say you read about them being a problem.

I must ask: Did you notice a problem with your previous restorations? If not, they what you've read may not be true in all cases.

If you're going to recap those AR90's, use the film caps and add a 1/2 ohm audio grade resistor in series and then you may come close to matching the ESR of the old cap. If you have an ESR meter, measure it first and then buy appropriate resistors to match if you don't want any surprises in brightness.

Try recapping without the resistors first and give them a listen. If you don't like the sound, go back in and add the resistors.

But, remember this important property of our listening habits, recapping with film caps brings the ESR back to what it was originally. However, over time, if those AR90's have drifted upward in ESR then surely they will gradually sound less bright - a response you become accustomed to. Then, when the recapping is done they sound bright. But you've forgotten how they sound originally and focus only on the change in brightness which can come as a surprise.

good luck!

[rrcrain]

Hello,

I want to recap crossovers in my "new" AR-90s.

In my previous restorations I used polypropylene caps manufactured by ICEL in Italy.

This is a very popular audiophile brand in my part of the world.

BUT...

I run across some threads saying that modern poly caps are not good, since they will change the voicing of old AR speakers.

The theory is that poly caps have lower ESR, and that using them will result in speakers having more treble and upper frequencies.

Did anybody confirm that in real life, or is it just hear-say?

I would like to be able to use poly caps due to their long life.

Any thoughts?

Thanks in advance,

Aleksandar

Carl had quite a few good points and the fact we don't know what the speakers sounded like when new is a big one. I'm not convinced the type of capacitor makes that big of a difference as long as it's a good quality, after all, the original caps were +- 20% electrolytics.

As for myself, I used Solen 400 VDC caps because they are a good cap without being stupidly priced and will last long enough it's likely I'll never have to replace them. Yes, the speakers will become brighter, probably a lot brighter. As an example, there was a dead upper mid in one of my AR90s and it sounded identical to the other speaker with a good driver. After recapping, the bad driver was glaringly apparent. I had someone locally recommend I remove the attenuator circuit and based upon my experience, I recommend you leave it there, you will probably need it.

[JKent]

You may want to have a look at the KLH forum:

http://www.classicspeakerpages.net/IP.Boar...?showtopic=4111

I recapped my KLH Twelves, first with electrolytics (cheap!) then with mylars (still cheap). I like the electrolytics, and that IS what they used originally. In some of my other restorations I've used polys (Zen, Dayton, Solen). One solution recommended by RoyC is to use mylars for the high frequency and polys for the low. I'm told the advantage of film caps (poly, mylar) over NPE is the films last longer. OTOH--how long do you NEED them to last? If your 30-yr-old NPEs need replacement, maybe new NPEs will last another 2 or 3 decades!

Or you could just use some of the "magic" caps being sold and dump a few hundred bucks in the 90s

Kent

[Diamonds&Rust]

I want to recap crossovers in my "new" AR-90s.

In my previous restorations I used polypropylene caps manufactured by ICEL in Italy.

If you have had good results with these, I would advise you to continue to use them with the AR-90 restoration.

I run across some threads saying that modern poly caps are not good, since they will change the voicing of old AR speakers.

There is no way to know what a 1967 pair of AR-3s sounded like in 1967. We have established that. I would like to add that even if we could, those of us who care what our systems sound like would ever put together a stereo that sounds like a stereo sounded in 1967 (except for nostalgic reasons). We are dealing with new sources, newly recorded and mastered recordings without the RIAA phono curves, etc. If in fact the speakers were "voiced" to sound good with recorded material from 1967, then they were voiced using recordings you and I would find inferior in every respect.

The theory is that poly caps have lower ESR, and that using them will result in speakers having more treble and upper frequencies.

I'm sure this is true. The question we have to answer is, "So what?" In **my opinion** this is a larger issue with the real classics, from the LST backwards, than it is with the ADD "Truth-In-Listening" series, or with the tastelessly named "Vertical Series."

Did anybody confirm that in real life, or is it just hear-say?

It has been confirmed (by listening) many, many times.

The question is / was "compared with what?" I have just last night changed two capacitors in my AR-9s. The capacitors that I removed, measured with my admittedly inexpensive capacitance meter, measured well. They were all not only within their tolerances but measured almost right on their labeled values. Amazing.

I replaced them anyway.

There were two changes; 1) the speaker's tendency to "heaviness" and "ponderous" bass was relieved; perhaps cured entirely, 2) the "smearing" of the bass region also ended. I listened to a lot of drums and will describe the difference as "speed." That is, I think the transients improved a lot.

*After* I listened I did a little analyzing: I've got a real-time analyzer and I admit that it has major limitations and the proper interpretation of its results are problematic. That being said, it's better than nothing.

Before replacing the two bass section capacitors ( a 2500uF and a 470uF) with new Jensen electrolytic non-poalr caps, the analyzer was reporting a huge dip in the 180Hz range (microphone position changed this result very little). After the new capacitors were installed the analyzer reports a small dip in the 125Hz band (also fairly constant and not mic placement sensitive - I ignore results at or below 100Hz, so this is questionable), but the 180Hz band is now flat - this is a *huge* difference (12db or more).

Whether the analyzer's results are right or not does not change the moral of the story: Things changed substantially

I'm a believer in poly capacitors, but I also believe I have other bills to pay, so buying enough new poly caps to make-up a 2500uF was out of the question. Even the 470uF capacitors would have been about $140 (each) to create with Solen caps, so NPEs were the only practical solution. I wish money were no object, but it is.

I would like to be able to use poly caps due to their long life.

Any thoughts?

Having recapped a pair of 90s and a pair of 9s at different stages with different capacitors, I would recommend that you use **good** poly capacitors everywhere you can afford to use them and stick a small value parallel capacitor across *all* the upper midrange and tweeter capacitors.

I think it might be worth noting that the analyzer reports that the speakers are many db-down at 16,000Hz.

My experience is the same as Rich's. The really badly out-of-spec capacitors are likely to be the ones across the upper midrange drivers. All of them. They are likely to be so bad that you can't even really hear your upper midrange drivers right now. It's an embarrassing truth that I could not tell how bad my upper midrange drivers were until I replaced the capacitors, changed rooms, and changed the drivers as an experiment. I first had to get another pair of speakers against-which to compare. Apparently my hearing is all relative.

Replace the 10uF resistor, too, with an "audiophile" (whatever that means) non-inductive resistor. It absolutely makes a difference. A huge difference? Well, it's significant to me. Anyone who argues otherwise is either deafer than I am or is using a really lousy source and amplification.

The AR-9 and AR-90 share the same crossover except for the woofer. Been there, done that. If you have good electronics and you want your 90s to sing, spend some money on their restoration; capacitors included. If they are going to be your bedroom speakers connected to a questionable receiver playing MP3s from an iPod, save yourself some money and buy cheap components.

I have not compared all the various brands out there, so I cannot recommend a specific capacitor. I'll just tell you that the ultra-inexpensive ones I've tried all sucked. But even the very cheap capacitors are much, much, much, much better than is what is your AR-90s now.

Ignore the urge to restore the ESR with a resistor. Capacitor ESR is a curve, not a resistor. The added resistance might be sonically desirable with an old pair of AR-3s, but it's no more authentic on an AR-90 than leaving it off would be. Will the speakers brighten? Let's hope so.

If you are doing a *full* restoration and are going to replace the resistor bank, I would encourage you to "half" the values and install high-quality, non-inductive resistors. The way it was built originally one switch position to the next is too much difference. You end-up going from too much treble to not-enough with one click.

I make myself unpopular around here with my capacitor opinions, but Klaus and I agree that everything you do does make a difference and it is audible. Whether or not it is audible enough to be important is entirely up to you and your expectations.

[Diamonds&Rust]

That's a 10 ohm resistor. I find it hard to believe I made as many typos as I did. Good grief. It's time for bi-focals.

[dynaco_dan]

That's a 10 ohm resistor. I find it hard to believe I made as many typos as I did. Good grief. It's time for bi-focals.

Hi Bret;

Thank you for the very interesting write-up, as always.

Vern

[Kuja]

WOW! What an excellent write-up!

Thanks a lot!

Things are so much clearer now!

Replace the 10 Ohm resistor, too, with an "audiophile" (whatever that means) non-inductive resistor.

What resistor you are exactly referring to?

Maybe there are some differences between your AR9 and my AR90?

Aside from those resistors connected to 0dB, -3dB, -6dB switches, there is a 6 Ohm resistor in parallel to the upper midrange driver and there is an 1 Ohm resistor in parallel to the woofers.

Here is the AR90 crossover schematic:

[RoyC]

One solution recommended by RoyC is to use mylars for the high frequency and polys for the low.

Actually, I mentioned that speaker companies would use electrolytic capacitors (not polys) for the low frequency drivers, as the larger film caps become quite costly, and higher quality mylar film caps for the high frequencies (as in the AR 303). I have liked mylar film capacitors for restorations because they a have bit more Equivalent Series Resistance than polys, making them act somewhat more like original electrolytic capacitors, with the durability of film capacitors.

We should keep in mind that all of the AR models mentioned in this thread originally used electrolytic capacitors and, according to Ken Kantor, ESR is always taken into consideration in a good crossover design. I recently had the opportunity to examine a crossover from an Avid loudspeaker from the late 70's. These excellent sounding speakers used mylar film capacitors, not electrolytic, that do not need to be replaced. Film capacitors were therefore available to AR at the time of the AR-9 and 90, but were not used. (My guess is that cost, not sound, was the more likely reason, but who knows.) I doubt there is any basis, however, to believe that ESR is somehow more or less important to the AR-3 design than to the AR-9 or 90.

As to whether ESR really NEEDS to be taken into account at all in the restoration of 30+/- year old speaker systems is another discussion altogether, and I largely agree with Bret on that issue. Yes, ESR is a curve, and I personally believe the "differences" that may be heard between caps are mostly attributable to this, but nobody here has the ability to determine the appropriate amount for a given old speaker. Adding small amounts (less than .5 ohms) series resistance is simply an attempt to emulate ESR, and it *seems* to have some merit for those trying to recapture the effects of ESR in an old design. Whatever the case may be, I do completely agree that it is a very good idea to replace very old electrolytic capacitors with new capacitors (of any type).

Carl's capacitor study, which agrees with my experience, suggests that most caps sold by reputable speaker parts suppliers will probably suffice quite well for most restorations. If electrolytic caps are used in the restoration of the AR-9 or 90, the original quality is achieved! IF you are inclined to experiment, or don't want to recap them again in 20 or so years, then use film caps of whatever cost, quality, color, size, shape or flavor you desire, but don't use something simply because some of us (including myself) say it sounds better...and if you do, then I would urge you NOT to "resist the temptation" to try to emulate ESR with a small resistor or try any other very subjective suggestion (such as "by-passing") that may enhance the enjoyment of this hobby. Be aware that the initial (typically brighter) sounding differences that may be heard by straying from the original design are not always "improvements" in the longer term, and at the very least, won't always be interpreted the same way by different ears.

The problem with contemplating advice resulting from very subjective assessments of electrical components is that it tends to make folks paranoid about what they MIGHT be missing. (Many over-priced, over-sized cables have been sold in this manner.) "Audiophiles" inherently have a need to believe that there is some magical, unmeasured quality to electrical components. It certainly has always been good for debate and, if nothing else, profit margin.

Roy

[Carlspeak]

...........

The problem with contemplating advice resulting from very subjective assessments of electrical components is that it tends to make folks paranoid about what they MIGHT be missing. ...............

Roy

Thanks Roy. I loved that one sentence above!

Unfortunately, there are a lot of DIY'rs out there in cyberspace who are unsure of themselves with respect to restoration. Perhaps because many are doing it for the first time and/or, have done some web research for the first time and got caught up in all the anecdotal reviews and opinions.

[Pete B]

Actually, I mentioned that speaker companies would use electrolytic capacitors (not polys) for the low frequency drivers, as the larger film caps become quite costly, and higher quality mylar film caps for the high frequencies (as in the AR 303). I have liked mylar film capacitors for restorations because they a have bit more Equivalent Series Resistance than polys, making them act somewhat more like original electrolytic capacitors, with the durability of film capacitors.

We should keep in mind that all of the AR models mentioned in this thread originally used electrolytic capacitors and, according to Ken Kantor, ESR is always taken into consideration in a good crossover design. I recently had the opportunity to examine a crossover from an Avid loudspeaker from the late 70's. These excellent sounding speakers used mylar film capacitors, not electrolytic, that do not need to be replaced. Film capacitors were therefore available to AR at the time of the AR-9 and 90, but were not used. (My guess is that cost, not sound, was the more likely reason, but who knows.) I doubt there is any basis, however, to believe that ESR is somehow more or less important to the AR-3 design than to the AR-9 or 90.

As to whether ESR really NEEDS to be taken into account at all in the restoration of 30+/- year old speaker systems is another discussion altogether, and I largely agree with Bret on that issue. Yes, ESR is a curve, and I personally believe the "differences" that may be heard between caps are mostly attributable to this, but nobody here has the ability to determine the appropriate amount for a given old speaker. Adding small amounts (less than .5 ohms) series resistance is simply an attempt to emulate ESR, and it *seems* to have some merit for those trying to recapture the effects of ESR in an old design. Whatever the case may be, I do completely agree that it is a very good idea to replace very old electrolytic capacitors with new capacitors (of any type).

Carl's capacitor study, which agrees with my experience, suggests that most caps sold by reputable speaker parts suppliers will probably suffice quite well for most restorations. If electrolytic caps are used in the restoration of the AR-9 or 90, the original quality is achieved! IF you are inclined to experiment, or don't want to recap them again in 20 or so years, then use film caps of whatever cost, quality, color, size, shape or flavor you desire, but don't use something simply because some of us (including myself) say it sounds better...and if you do, then I would urge you NOT to "resist the temptation" to try to emulate ESR with a small resistor or try any other very subjective suggestion (such as "by-passing") that may enhance the enjoyment of this hobby. Be aware that the initial (typically brighter) sounding differences that may be heard by straying from the original design are not always "improvements" in the longer term, and at the very least, won't always be interpreted the same way by different ears.

The problem with contemplating advice resulting from very subjective assessments of electrical components is that it tends to make folks paranoid about what they MIGHT be missing. (Many over-priced, over-sized cables have been sold in this manner.) "Audiophiles" inherently have a need to believe that there is some magical, unmeasured quality to electrical components. It certainly has always been good for debate and, if nothing else, profit margin.

Roy

You make some good points, especially that we should keep in mind that the original design incorporated low cost non-polar electrolytic capacitors. The claim that because ESR is a curve, it is not worth including with a film cap is not based in sound engineering. Trained engineers perform sensitivity analysis, where it is possible to determine the change in transfer function as related to deviations in each component in a circuit and with regard to frequency. Understanding sensitivity analysis and matching ESR where it is most critical provides an excellent solution. Film caps, even low cost types, with added resistance is the best solution in my opinion, and good quality NP electrolytic caps of similar DF to the originals is also a good low cost solution for larger values. There have been bad chemical formulas used in some recent electrolytic capacitors which leads to premature failure, so one must be cautious when selecting them. I've not seen this in any NPEs - yet, but there have been some power supply caps with problems.

[RoyC]

Film caps, even low cost types, with added resistance is the best solution in my opinion, and good quality NP electrolytic caps of similar DF to the originals is also a good low cost solution for larger values.

Hi Pete,

Glad you weighed in...

I have had luck placing small resistors in series with film caps (mostly on tweeters) when replacing electrolytic caps...and have never heard a downside to doing so. Bearing in mind that the typical speaker restorer does not have any special testing equipment, would the routine addition of, say, a .3-ish ohm series resistor to new film caps be a generally safe recommendation for old speakers originally equipped with electrolytic capacitors?

Thanks,

Roy

[ar_pro]

An important factor that's often missed in the discussion of re-capping/restoring the AR-9, is the effect that the original level switches have on performance.

These switches were of questionable quality when the speaker was new, and are close to worthless, 30 years later. Most of them will exihibit intermittant behavior when moved through their actions, and that's *got* to have an effect on the speaker's sound - new caps, or not.

Debating ESR and the relative merits of bi-pass capacitors is useful, but seriously - where else in your audio system would you tolerate this kind of poor connection?

[RoyC]

An important factor that's often missed in the discussion of re-capping/restoring the AR-9, is the effect that the original level switches have on performance.

These switches were of questionable quality when the speaker was new, and are close to worthless, 30 years later. Most of them will exihibit intermittant behavior when moved through their actions, and that's *got* to have an effect on the speaker's sound - new caps, or not.

Debating ESR and the relative merits of bi-pass capacitors is useful, but seriously - where else in your audio system would you tolerate this kind of poor connection?

Timely reality check, ar_pro...and it gets even crazier with the old pots and decomposing drivers of the earlier models .

Replacement cap differences probably sit at the very bottom of a very long list in any restoration or comparison of these old beasts.

Point well taken.

Roy

[JKent]

Actually, I mentioned that speaker companies would use electrolytic capacitors (not polys) for the low frequency drivers, as the larger film caps become quite costly, and higher quality mylar film caps for the high frequencies (as in the AR 303).

Sorry. I'm soooo confused!

That IS what Roy said and it is the advice I followed for my KLH Twelves.

[RoyC]

Sorry. I'm soooo confused!

That IS what Roy said and it is the advice I followed for my KLH Twelves.

Kent,

The photo of your finished crossover board over in KLH land shows mylars and electrolytic capacitors, just as we discussed...no polys. You didn't want to use polys due to their large size and cost.

Your new board looks good from here.

Roy

[roberto.tione]

Hello,

I want to recap crossovers in my "new" AR-90s.

In my previous restorations I used polypropylene caps manufactured by ICEL in Italy.

This is a very popular audiophile brand in my part of the world.

BUT...

I run across some threads saying that modern poly caps are not good, since they will change the voicing of old AR speakers.

The theory is that poly caps have lower ESR, and that using them will result in speakers having more treble and upper frequencies.

Did anybody confirm that in real life, or is it just hear-say?

I would like to be able to use poly caps due to their long life.

Any thoughts?

Thanks in advance,

Aleksandar

I have just restored a pair of Allison One and a pair of Allison Two.

For big values I used Audyn Cap bipolar electrolitic capacitor in parallel to MKT poly caps to achieve the original values (Royalitic Cap values used by Allison are today almost out of standard). Audyn Cap have very low ESR and very high precision: they are characterised @ 5% but I measured with a high precision RLCmeter FLUKE PM6306 and I found them far better than 1%.

MKT poly are available on RS-Components: again, I measured them and they were far better than 1%, loss-angle-tan negligible or even close to null. You could match them to achieve the expected value without measuing individually.

The sound is incredibly detailed, not boring. The speakers are 30 years old and I don't remember as they sounded at that time, but I would not change the present sound.

Do your recapping without regrets, and enjoy the result.

Warm regards, Roberto

[Diamonds&Rust]

What resistor you are exactly referring to?

I was stupidly referring to a resistor that isn't in your speakers. Visions of other AR crossovers were dancing in my head. Still, I think it would be worthwhile to replace all of them. Why? It costs next-to-nothing, it's not much trouble, and I've had the sand-cast resistors crumble to dust when I touched a couple of them (admittedly, not in a 9 or 90).

Maybe there are some differences between your AR9 and my AR90?

Yes, the woofer sections are different, the crossovers to the lower mid from the woofers are different.

You know, I have no problem with doing what Roy is talking about. A little resistance here and there can be a good thing. I do not know, and can't get anyone to tell me, how "voicing" is done (exactly) before a speaker goes into production. Here's why I question the whole ESR thing:

* I am a manufacturer of speaker systems. I have an engineering staff. They sit-down at the drawing board to design a new speaker system. They decide what they are trying to achieve and why. They draw it up and calculate it out and run simulations on their programmable HP calculators or use a slide rule and an Apple "Lisa" or maybe write the code and have Ross Perot run it on his mainframe at EDS for $100,000 / minute of CPU time.

They order prototype drivers to be built. They order inductors to be wound. Do they also put in an order for a series of capacitors custom to the speaker? ( "I know we have a parts bin with 10,000 8uF Callins capacitors in it, but dammit, I need one with a slightly higher ESR, and a 4uF with slightly lower ESR, and I need a 6uF with an ESR that goes up here, down here, and sideways in this range!") I doubt that, seriously.

So, maybe they add and subtract these effects using inductors? No, doesn't look like it.... I mean, all the AR inductors are 17 gauge wire. Tim Holls calls that "low resistance." If low resistance was what he was after, wouldn't he maybe have used 12 gauge wire? Why not? Because AR wound its inductors using 17 gauge wire, that's why.

"I've got an idea, we'll compensate for the ESR I want in this capacitor by adding a resistor to it in the AR-xyz, and we'll drop resistance by paralleling a resistor in the AR-abc. " If anyone sees any evidence that **any** fiddling with ESR by adding and subtracting parallel or series resistance went-on at AR, please point me to it. I have not seen any evidence at all. We would be looking at tiny value resistors attached to a capacitor somehow. Anyone? Any evidence?

What are the odds that Callins capacitors had just the right ESR so that the same capacitor could be used on five different tweeters and satisfy the engineers? If the resistance of the inductors and the ESR curves of the capacitors were used in calculating the effects of the crossover, then they MUST have had all the drivers built to co-exist nicely with the ESR and inductance of the caps and the resistance and capacitance of the inductors, right?

So now we are building a speaker. Let's use an AR-11 as an example. We have *Sprague* capacitors in the warehouse and 17 gauge inductors from our own winder so we design, voice, and manufacture the speaker perfectly by designing the drivers, themselves, around the parts we already have. Oooops, our tweeters are burning-out. Let's change it to a new ferrofluid model built to exactly co-exist with the capacitors and inductors.

Oooops, Sprague loses the contract or goes out-of-business. Well, no matter, we'll use our stock until it is gone. When we run out of 120uF capacitors, we'll parallel two 60uF caps. (this changes the ESR by half)

Oooops, we're out of Sprague "Compulytic" capacitors (not cheap NPEs, btw), our new supplier is Callins. The capacitors are much different than the Spragues. They are smaller with much higher ESR and DF and are made from different materials. We tell Callins - "Hey, we need you to match the Sprague caps *exactly* because all of our drivers were built to play-nice with the Spragues. Assume that Callins does this (it's a far-flung fantasy, but assume it for now). The speakers are now built using "updated" tweeters and have gone through two capacitor changes and nobody has thought to add a tiny resistor here or there to compensate for anything. Evidently, professional engineers taking all of this into account think that there is no difference, or that their calculations show that the difference is not meaningful, or the finance department is deaf and won't spend the money for the engineering time.

So let us assume, for the moment, that ESR may be a portion of the engineer's calculations, but it is a "throw-away" figure. It's taken into account, but it is NOT used to spec *anything*. AR used "Compulytic" capacitors in an attempt to minimize ESR and or DF or microphonics or failure or whatever. "Compulytic" was the result of building the cap a specific way, NOT something AR had stamped on their capacitors. It wouldn't surprise me a bit to find that AR was buying what Sprague was selling, not directing Sprague to manufacture some caps in a new and specifically custom way just for AR.

I flogged this dead horse for years and it never did get up and take me for a ride. Yes, Roy is absolutely correct, adding and subtracting resistance does change the way the speaker sounds. Yes, Pete is correct, a professional engineer doing professional engineering in a sound manner trying to exactly match the old capacitors would take X, Y, Z, into account and precisely design a.... But the ORIGINAL values that engineer would be working to are RANDOM side-effects.

It's like trying to rebuild a house after it burns to the ground by buying exactly the same wood from exactly the same trees in exactly the same forest, then putting nails *exactly* where they were in the old house, but compensating for new nail technology by varying the size of the nail to exactly match the sheer strength of the original nails. It's a completely, absolutely, useless, pointless, bound-to-fail endeavor.

Because we *prefer* the slightly higher resistance and can *justify* it by pointing to a random variable does NOT make it "sound engineering."

At one point it was shown that pipe smokers actually lived longer than non-smokers. Citing that statistic does not make pipe smoking "sound medical advise."

We have ample evidence that if AR used ESR measures in the crossover design, they didn't think it important enough to *ever* compensate for it in ANY way, ever, ever, ever. Never ever. So professional sound engineering or not, it (engineering) wasn't done, or if it WERE done it was roundly and completely ignored. Other than appeals to authority for what could have been done in a perfect world, please someone show me any evidence that ESR was ever compensated-for. Taken into account? Sure - they dropped it wherever they had to - when it got out-of-hand - by using coke-can sized caps.

I suspect it was "You pays your money and takes your chances," when they ordered capacitors. They certainly did NOT order different 8uF capacitors depending on if it went "here" or "there" in the crossover.

Think, guys - carefully - about what you are saying. You're saying that AR built drivers TO the WWII surplus caps, and the Chicago Industrial, and ______, and Sprague, and Callins, and, and, and, and... I'm tellin' ya, that did not happen. We have evidence that it did NOT happen.

Can we please, please, put this one to bed as a matter of "precision" and "correct information" and "professional engineering?" Can we? Please?! Pretty please?

Now, onto ar_pro's observation about switches. The mini or micro switches used in the AR9 are inadequate in every way. I have to go into mine with De-Ox-It all the time (which ain't easy) and I suspect they are just falling apart inside. I am debating whether to just by-pass them entirely and use an EQ if I find it necessary to turn-down some band. I'll probably end-up replacing them with like parts (available from Digikey). The switches in the AR90 are somewhat more robust, but they probably are not up to the task some 30 years later, either.

He makes an *excellent* point about our putting-up with such a poor connection elsewhere in our speakers. Pay attention to your slip-on connectors (to the drivers, themselves) and be sure they are tight and making more than a one-point loose connection, but don't make them so tight that they can't be removed.

I'll make my last unpopular point and I'll go away and let you enjoy the restoration process however you choose to do it:

If you are going to power these speakers with a modern 100w/channel receiver, you will never hear what they can do. The AR90 may be one of the most difficult loads to an amplifier that I have ever been around. They need a big, heavy, high-current amplifier to sound their best. If you intend to run them with a 120w/channel $250 receiver, you might *really* consider going back with NPEs and saving the hundreds you will spend on high-quality crossover parts. You will not get full benefit of expensive parts if you are going to drive them with a marginal amplifier and listen to marginal sources.

And do, really, please notice that I never, ever, ever said "Go buy the $100 magic-liquid-filled capacitors from "Caps R Us" that can be found in the 'I'm an idiot' section of Aisle 5." All I said was, "If you are happy with what you have used before..." and "...buy good...," meaning "don't buy the cheapest thing you can find."

That's it. I'm done. Enjoy your speakers!

Bret

[KlausDK]

Hi Alexander (and who-else reads the thread)

I very much agree with Bret about the value/nessecity of ESR compensation, and for the same reasons plus one: Every component you put in between the source and the driver changes sound. I have used quite some time to experiment with resistors in the price range betwee 1½ $ and up to 30$ (per resistor), and there are big differences. It started when my usual supplier of 1½ $ resistors suddenly changed brand and what a difference in sound (for the better). After some experimentation, I have settled for the 4 $ RSS metal film resistors from ThyOhm as an acceptable compromise bewteen sound and price, eventhough for sure the 30 $ Duelund resistor did sound better!

So cut in short, don't put a resistor in unless you are sure that it is nessecary!

I use the Jantzen Cross Cap capacitors (wound on the Finnish Tervakoski foil) for most of my restorations, and they sound brilliant for a very moderate price. I find that they are marginally better than the Solen/SCR caps. If I occasionally use more expensive capacitors, it is only for the tweeter. I would advice you not to use Mylar/Polyester (mkt) capacitors in the signal path, to me at least they give a harsh sound.

As power requirements to the AR-90, I agree with Bret as well, if you want your woofers to give you a tight and well controlled bass, you need something with muscle. If you are going to restore the cross over anyway, you could prepare them for bi-amping as the AR-9 is from the factory.

BRgds Klaus

[Carlspeak]

Bret, your long post regarding ESR appeared to me to be focused on the hypothetical design and production of a new loudspeaker and what, if any, consideration developers of early AR speakers took into account with regard to this capacitor property.

I think your absolutely right, I'd be VERY surprised if they did.

ESR, DF and uF (not listed any any particular order of importance) are the three primary attributes of a capacitor. I'm guessing, designers probably initially specified a uF value and type (NPE, mylar, film, etc.) and purchasing went out and got the best price. Then again, maybe during development and 'voicing' designers found they liked a particular brand of cap used in the prototypes and told purchasing to specify that brand in adidtion to the type and uF value.

With all of this, it had to be understood by all experienced in real world manufacturing that there would be some level of variance in all three characsteristics I listed previously. The capacitance value obviously carried a tolerance that was printed on most caps. However, the other properties may not have been as important or, specified as to their tolerance. AR llived with this reality and accordingly, the buying public had to live with it as well.

So, what I'm trying to say is, back then, it was obvious to all there would be some mild variance in the sound of individual speakers coming off a new speaker line due to these normal variances in properties of the caps, resistors and inductors - just to focus on the xover components for now. Perhaps that's why audio consumers back then (and now) are most happy when they see serial numbers on a pair either consecutive or, very close together becasue they believe the components are probably from similar lots and are most likely to be matched better than two serial numbers 10's of thousands apart and thus, the sound out of each would not be significantly different.

But now we come to the the 40 year old speaker with 40 year old caps which, in my experience with the classic speakers from NE manufactures, are most likely NPE's. There ESR's have drifted from relatively low values of < 0.25 ohms to >> 1 ohm due to the age related degradation of electrolyte paste encapsulated inside and perhaps suffering from some leakage as well. Someone like Aleksandar comes along and buys these speakers and wants to replace the caps. In doing so (sans resistors) he will most likely create a STEP CHANGE in the tweeter's output due to the step change (down) in ESR brought about by the new cap. It's that step change most recappers get nervous about - not the normal variance the original production models had which was much-much more subtle. And therin lies the concern most posters here and elswhere have when approaching a recapping project. Many don't know what to expect and seek advice.

Therein lies the difference between the need to be concerned about ESR when recapping old speakers vs the acceptance of mild variances that come about in the normal course of original production.

IMO, Ideally, the recapper should test his old caps with an ESR meter so he'll know exactly how badly the old cap has drifted and make a better informed decision on how to approach the recapping project to suit his or her tolerance for this potential step change in performance. This is why I answered Aleksandar's original post in the way I did.

[Diamonds&Rust]

It's that step change most recappers get nervous about - not the normal variance the original production models had which was much-much more subtle. And therin lies the concern most posters here and elswhere have when approaching a recapping project. Many don't know what to expect and seek advice.

Your point is well-taken, but it has *nothing* to do with my observations. I'm saying that the .25ohm that was accidently there, which step changed to 1 ohm, was the product of chance. A step change from the original .25 ohms to .0025 ohms is random, also.

Therein lies the difference between the need to be concerned about ESR when recapping old speakers vs the acceptance of mild variances that come about in the normal course of original production.

I'm not meaning to be argumentative, but I don't see your point. If AR dropped ESR by half by compounding capacitors to use what was in stock, that would seem to represent a step change wouldn't it?

But absolutely nowhere did I say that we should be concerned about small variations in component values, or that I was, or that the engineers would be.

I guess maybe I'm being dense; but it seems to me if ESR was "whatever it is" and wasn't adjusted for, ever, then not-adjusting it up .2 ohms is no more valid than leaving it alone.

By the way, I thought of an example of where it was adjusted-for. Evidently the 2500uF caps in the 10pi (and maybe the LST??) were paralleled by a 10ohm resistor. That's the only time I can remember seeing anything like that in a schematic. Did they use the capacitor's ESR and 10 ohms as a precise value or were they just doing whatever they could to minimize ESR?

IMO, Ideally, the recapper should test his old caps with an ESR meter so he'll know exactly how badly the old cap has drifted and make a better informed decision on how to approach the recapping project to suit his or her tolerance for this potential step change in performance. This is why I answered Aleksandar's original post in the way I did.

We'd have to have the spec sheets from Sprague, Callins, etc to know what we started with. IF we had those all of this debate would become extra-moot. But yes, the ideal recapper would come complete with a lab and an encylopedic reference section.

Because we cannot know what those original values were, or how our drivers have changed what those values should ideally be, or how much attention was every given to ESR by the engineering staff is why I answered Aleksandar's post the way I did.

How can we possibly suggest adding resistence to the capacitors he has used with success in the past? We don't know what the ESR of those caps is. Just because we **assume** the ESR is neglegible due to the materials doesn't mean that it is.

I wish you could be here right now listening to the tweeters in my 9s. I sincerely doubt you'd suggest adding a resistor to the film and foil caps in series with the tweeter. You might suggest adding a super-tweeter, but a resistor? Never.

...and one of my tweeters is a barely used NOS I got from Tom. It's not possible to tell which one without looking for a sticker on the back.

I wouldn't have "argued" back with you, but I felt as though you were confusing my dismissal of the ESR values with dismissal of small manufacturing variants, and I was being much more broad than your post implied.

Oh, lastly, for anyone reading-along that maybe doesn't know this:

If Roy told me to spit on the woofers in my AR3a's because it made them sound better, I'd be horking-up a good wad-o-spit. If Roy says, "use a resistor on the 3a tweeter to bring it back into balance with the cabinet," I'm all over that. I've had "modern" NPEs in my 9s and in some 90s - they sucked. Stepping through the tulips or not.

Bret

[RoyC]

If Roy told me to spit on the woofers in my AR3a's because it made them sound better, I'd be horking-up a good wad-o-spit. If Roy says, "use a resistor on the 3a tweeter to bring it back into balance with the cabinet," I'm all over that. I've had "modern" NPEs in my 9s and in some 90s - they sucked. Stepping through the tulips or not.

Bret

Yikes, I had to go back and read what I wrote to see how we got here !

Bret,

You never fail to make me smile! Even if I thought a wad of spit on the 3a woofer would make it sound better, I probably would not suggest that to anyone (other than you ) unless I could somehow measure the positive results...and I certainly have NO experiences to share regarding the AR-9 or AR-90. You know way more about those beasts than I do.

Although I am intrigued by the possible effects esr may have on the scheme of things, and have played around with small series resistors to see if that is what we may be "hearing" in terms of cap differences, I was not suggesting it was necessary to include series resistors in restorations. Added series resistance is simply something I've been experimenting with. Replacing old, worn caps is the only thing that is absolutely necessary. When Pete, an engineer, posted his response I was hoping he would share more insight on the subject. (As an aside, upon reading ar_pro's comments in your AR-9 thread, I couldn't help but think that the issues he shared regarding building up caps in the woofer circuit has something to do with esr. Paralleling caps to build capacitance lowers esr just as paralleling resistors lowers total resistance.)

As has been stated here many times before, we really have no frame of reference when "restoring" old speakers. Even if the new electrolytic or film caps sound especially "good" or "bad" to the restorer, there is no way of truly knowing whether the speaker sounds better or worse than it did many years ago. Additionally, it seems AR speakers were eventually somewhat compromised as original drivers were discontinued or modified for later models. Changing cap brands or types were probably the least of the company's concerns after a model had already been established.

Roy

[Kuja]

Thanks guys!

Things are much clearer now.

I'll use quality "for audio" poly caps, but I will not go crazy spending too much money on some audiophool voodoo stuff.

What theory says? What differences in ESR can be expected between old electrolytic and modern poly caps in general?

Looking at schematic and those lower/upper mid and tweeter level switches, I see that resistor values for -3dB position are between 1.5 and 3 Ohms.

I doubt that differences in resistance between various types of capacitors can be that big.

Is difference in some driver otput due to different cap ESR smaller or bigger than let's say 1dB?

Can we actually hear that difference?

I would be very happy if frequency response of our speakers were that linear (+/- 1dB)

About bypassing with smaller caps, or using multiple caps with smaller values in parallel...

...is any "real" manufacturer using this technique, or is it exclusively used by DIY enthusiasts who "know better"?

I'm also worried about those level switches. How many AR-90 owners actually use these to tailor their sound?

Can they be cleaned/restored so that they don't have any negative influence on sound quality?

Right now, I'm thinking of bypassing them.

On the other side, I'm affraid that I might regret that decision later, since I've never owned a loudspeaker system that has this feature and I don't really know how useful can it be.

They seem more robust than those found in AR-9s.

If you are going to power these speakers with a modern 100w/channel receiver, you will never hear what they can do. The AR90 may be one of the most difficult loads to an amplifier that I have ever been around. They need a big, heavy, high-current amplifier to sound their best. If you intend to run them with a 120w/channel $250 receiver, you might *really* consider going back with NPEs and saving the hundreds you will spend on high-quality crossover parts. You will not get full benefit of expensive parts if you are going to drive them with a marginal amplifier and listen to marginal sources.

As power requirements to the AR-90, I agree with Bret as well, if you want your woofers to give you a tight and well controlled bass, you need something with muscle.

I hope that I'll have no problem driving them.

This vintage monster integrated amp is driving effortlessly my present pair of AR98-LSI:

http://www.thevintageknob.org/KENWOOD/KA907/KA907.html

I'm a big vintage enthusiast and here you can see some pics of my recent restoration of this amp:

http://www.kujucev.com/KA907restoration/907.htm

http://www.kujucev.com/audio/01_Kuja_sistem_1.jpg

[rrcrain]

I'm also worried about those level switches. How many AR-90 owners actually use these to tailor their sound?

Can they be cleaned/restored so that they don't have any negative influence on sound quality?

Right now, I'm thinking of bypassing them.

On the other side, I'm affraid that I might regret that decision later, since I've never owned a loudspeaker system that has this feature and I don't really know how useful can it be.

They seem more robust than those found in AR-9s.

I strongly recommend not bypassing the switches. Depending on your room, they might suddenly become useful after recapping. I have a few well done CDs of afemale vocalists that I'd swear were trying to pierce my ear drums. Thankfuly, they do the same to every decent system I've heard

[Diamonds&Rust]

Things are much clearer now.

Give us time. We'll fix that. Nothing will be clear in a matter of days.

What theory says? What differences in ESR can be expected between old electrolytic and modern poly caps in general?

Huge. Poly caps will be, for all practical purposes, zero. Even .1 Ohms is a big percentage gain from 0 Ohms --- let's see.... carry the 9.... cross-multiply and divide by 2.... That's an infinite percent increase if I did my math right. Can it be heard? You betcha it can.

By using poly film capacitors instead of NPEs you absolutely will "tilt" the voice of the cabinet upward. No question. No question at all. You might even cause some resonance problems and create crossover resonant peaks. You will, absolutely, be defeating the original "voice" of the speaker. Of course, you don't now have the original voice of the speaker, either. What do you want to achieve?

I hesitate to ask, but is this going to be a vinyl-only system or will you play CDs?

Looking at schematic and those lower/upper mid and tweeter level switches, I see that resistor values for -3dB position are between 1.5 and 3 Ohms.

I doubt that differences in resistance between various types of capacitors can be that big.

Is difference in some driver otput due to different cap ESR smaller or bigger than let's say 1dB?

Can we actually hear that difference?

You are correct. The various types of capacitors are not as dramatic as the attenuators. See my comment about the steps in the attenuators being too large. You can hear a change with far, far less than 1.5 Ohms.

I would be very happy if frequency response of our speakers were that linear (+/- 1dB)

Yes... that would be nice and still there would be differences between our + /- 1db speakers and someone else's +/- 1db speakers.

About bypassing with smaller caps, or using multiple caps with smaller values in parallel...

...is any "real" manufacturer using this technique, or is it exclusively used by DIY enthusiasts who "know better"?

Uh oh... Do you remember back in school that there would always be one teacher who would fail you on a math exam because, "You didn't show your work!" Every time someone here says "by-pass capacitor" someone else here jumps up, pulls out another clump of hair, stomps their foot, picks their lip until it bleeds and after lighting 13 black candles begins incantations to curse the opinioneer with warts and long-term mother-in-law visitation.

But to answer your question: Some "real" manufacturers did.

"The woofer and lower midrange networks _____ high quality electrolytic capacitors bypassed with metalized film capacitors. The upper frequency networks use low-loss metalized film capacitors exclusively."

That's a quote from the AR-MGC brochure.

That doesn't mean you should do to your 90s what was done to the MGCs. Let someone's actual experience work for you, or, and I'd do this if I had a spare pair of 90s, which I've tried to get, which I haven't been able to get... externalize your crossovers and knock yourself out trying different things. I, for one, would be really interested in your results and preferences.

I'm also worried about those level switches. How many AR-90 owners actually use these to tailor their sound? Can they be cleaned/restored so that they don't have any negative influence on sound quality? Right now, I'm thinking of bypassing them. On the other side, I'm affraid that I might regret that decision later, since I've never owned a loudspeaker system that has this feature and I don't really know how useful can it be. They seem more robust than those found in AR-9s.

They are easily by-passed if you choose to do it later. Remember that these were not inexpensive speakers in their day and AR felt these switches were adequate. You are correct. The 90s switches, by virtue of not needing to be pretty, were "bigger" than those in the 9. Mine gave me no trouble for many years and I've never had any trouble with a switch like the ones in the 90s which were used on other models, too. It would be easier to justify by-passing the tone controls if you are willing to use your tone controls on your amplifier - but I'd say leave them in to start. You can easily "customize" them away, later.

I hope that I'll have no problem driving them.

This vintage monster integrated amp is driving effortlessly my present pair of AR98-LSI:

http://www.thevintageknob.org/KENWOOD/KA907/KA907.html

I'm a big vintage enthusiast and here you can see some pics of my recent restoration of this amp:

http://www.kujucev.com/KA907restoration/907.htm

http://www.kujucev.com/audio/01_Kuja_sistem_1.jpg

Yeah, you should be fine. Nice amp.

If you have access to some really high quality non-polar electrolytic capacitors, you could try them. We really don't have good ones available to us in the U.S. anymore, so using poly film caps is really all we can do.

Bret