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

[rrcrain]

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?

Bret,

This is something I have to disagree with. To many people claim to not hear differences between capacitor types for starters and based on my experiences and those of a local shop that's recapped quite a few speakers, it's not true. This gentleman went as far to call Snell and tell them of his customers request to recap using poly caps and wanted their opinion. In a nutshell, they were entheuastic over the idea and told him they would have used poly caps if they had been available at the time the speakers were built. instead, they used ganged NPE caps.

Even the noble idea of using NOS NPE caps is flawed. If left unused, the aluminum oxide layer in an NPE can degrade over time , and that's not even considering the electrolyte will still slowly degrade at near the same rate as a cap in circuit under normal use.

yes, recapping will absolutely make the AR90 brighter, and to be fair, it's needed from what it is with the original caps. I have personal, first hand experience there. Just as well, no one, including me, has a clue what these speakers really, truely sounded like new, and I'm the original owner of my speakers.

Stay with the rated values on the schematic, use whatever cap type you choose that is appropriate and it's almost impossible to go wrong. If they become to bright for your room or personal taste, thats what the switches are there for.

If your concerned with the switches, spray a good contact cleaner into them, work the switches about for a bit and they should be fine if they are still mechanically sound.

As for rewiring the AR90 for biamping, I did that and liked the result, but it only made a difference after I installed an active crossover. Biamping without the active crossover resulted in absolutely no improvement.

[Pete B]

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

Hi Roy,

I seem to recall making a cranky comment about this some time ago, and I don't

want to generalize here because of how things are often blown out of proportion,

not by you by the way. Oh, here's the cranky post:

http://www.classicspeakerpages.net/IP.Boar...myths&st=20

I'll make a highly qualified statement, I don't think I've measured an AR or Advent

cap in the 5 to 16 uF range with an ESR lower than .3 ohms, so .3 would not hurt with a low

ESR film type. You probably know that I verified this in both a listening test and with

measurements for at least one restoration that I did involving Advents. I have measured NPE

caps in other fine speakers that had an ESR that was over 1 ohm. I would say that .3 is better

than zero, and .47 will probably be closer to the originals more of the time when replacing

most NPEs in AR and Advent speakers. This does not apply to computer grade of course,

which I have discussed in the other thread.

You could also computer ESR from the DF data in the spread sheet that Johneio posted some

time ago, I don't seem to have it handy, otherwise I could check a few of his values.

We can discuss this more offline if you'd like.

[Pete B]

Hi Roy,

I seem to recall making a cranky comment about this some time ago, and I don't

want to generalize here because of how things are often blown out of proportion,

not by you by the way. Oh, here's the cranky post:

http://www.classicspeakerpages.net/IP.Boar...myths&st=20

I'll make a highly qualified statement, I don't think I've measured an AR or Advent

cap in the 5 to 16 uF range with an ESR lower than .3 ohms, so .3 would not hurt with a low

ESR film type. You probably know that I verified this in both a listening test and with

measurements for at least one restoration that I did involving Advents. I have measured NPE

caps in other fine speakers that had an ESR that was over 1 ohm. I would say that .3 is better

than zero, and .47 will probably be closer to the originals more of the time when replacing

most NPEs in AR and Advent speakers. This does not apply to computer grade of course,

which I have discussed in the other thread.

You could also computer ESR from the DF data in the spread sheet that Johneio posted some

time ago, I don't seem to have it handy, otherwise I could check a few of his values.

We can discuss this more offline if you'd like.

I cut and pasted that link but it doesn't seem to go to the right post, cranky post is #24, LOL!

[Kuja]

Ok...

Here are the photos of caps that I have in my AR-90 crossovers.

All caps are original.

Some of them (on tweeters and upper midranges) do not seem to be electrolytic.

There are some white caps with yellow ends and some black caps with red ends.

What type are they?

If one would replace like for like, what would be today's closest equivalent?

[Carlspeak]

Ok...

Here are the photos of caps that I have in my AR-90 crossovers.

All caps are original.

Some of them (on tweeters and upper midranges) do not seem to be electrolytic.

There are some white caps with yellow ends and some black caps with red ends.

What type are they?

If one would replace like for like, what would be today's closest equivalent?

The white and yellow ones and black and red ones appear to be film/foil caps. They may be in okay condition. However, only testing with an in-circuit ESR meter will answer that question. Without doing the testing, you have a choice; either just replace the NPE's or, replace all. Suggest Dayton FF as a roughly equivalent, cost effective replacement of the caps described in my first sentence.

[Kuja]

Suggest Dayton FF as a roughly equivalent, cost effective replacement of the caps described in my first sentence.

Thanks Carl,

I live in Europe, so I'm not familiar with Dayton brand.

What type exactly those "FF" caps are?

Best regards,

Aleksandar

[Pete B]

Thanks Carl,

I live in Europe, so I'm not familiar with Dayton brand.

What type exactly those "FF" caps are?

Best regards,

Aleksandar

That looks to be a 6 uF cap, and it is most likely just a low

cost film type. FF caps use a thick foil and it is not possible to get

large uF values in a package that size.

[KlausDK]

Ok...

Here are the photos of caps that I have in my AR-90 crossovers.

All caps are original.

Some of them (on tweeters and upper midranges) do not seem to be electrolytic.

There are some white caps with yellow ends and some black caps with red ends.

What type are they?

If one would replace like for like, what would be today's closest equivalent?

Hi Alexander

This looks like the same caps I had in my AR-90, for the tweeter and midrange drivers. I agree with Pete, they are low cost not worth saving. I upgraded one speaker completely first, left the other as is, and compared the sound. The speakers were like night and day, the original being the night, with a darker soundpicture, canned voices, you wouldn't beleive it was the same speaker.

I agree with rrcrain, that you need electronic crossover for the full effect of bi-amping, remember to divide 1 octave above and below the 200 Hz for the woofer and upper section respectively (100 Hz for midrange/tweeter circuit, 300 Hz for woofer)

BRgds Klaus

[mluong303]

Hi Aleksandar

The made in France Axon and Solen polypropylene capacitors are very reliable with consistence quality control. 250VDC rating units are better due to small physical dimension to jammed onto the fixed small space of the original AR crossover board.

Minh Luong

[Kuja]

Hi Alexander

This looks like the same caps I had in my AR-90, for the tweeter and midrange drivers. I agree with Pete, they are low cost not worth saving. I upgraded one speaker completely first, left the other as is, and compared the sound. The speakers were like night and day, the original being the night, with a darker soundpicture, canned voices, you wouldn't beleive it was the same speaker.

I agree with rrcrain, that you need electronic crossover for the full effect of bi-amping, remember to divide 1 octave above and below the 200 Hz for the woofer and upper section respectively (100 Hz for midrange/tweeter circuit, 300 Hz for woofer)

BRgds Klaus

Thanks Klaus!

Can you be more specific about details of your AR-90 restoration?

What caps did you use? Any resistors in series due to different ESR?

What do you think of Mundorf from Germany?

A this moment, it seems that it is the only brand readily available in my country.

I'm thinking of getting caps from their MCap series:

http://www.mundorf.com/english/bauteile/frkondensator.htm

What about the AR-90 internal wiring? Original seems barely adequate to me. I'm thinking of changing it to something better.

Here I can get some quite nice Neotech UPOCC solid core wire.

http://www.neotechcable.com/upocc.htm

What did you do with those levels switches on the back? Cleaned them, or bypassed them?

Best regards,

Aleksandar

[Carlspeak]

Aleksandar

I think the Mundorf's are not a cost effective choice. They are very expensive and used in VERY high end speakers. Perhaps you can find a dealer for Clarity caps on the continent. They are made in the UK and sold in the US by Madisound.

As recommended earlier, just replace the caps (without ESR compensating resistors) and listen for a while. If the sound is a bit brighter, perhaps the attenuation controls can compensate for this. Don't rush to put in resistors in series with your new caps. Get used the speaker's sound over a period of weeks. You may find it quite pleasing after a while.

Can't help you with the wire question, however.

Good luck with your project.

[rrcrain]

What about the AR-90 internal wiring? Original seems barely adequate to me. I'm thinking of changing it to something better.

Here I can get some quite nice Neotech UPOCC solid core wire.

http://www.neotechcable.com/upocc.htm

My wo cents worth on replacing the wire is simple. If you want to replace it, feel free to do so, but realistically, don't expect miracles. Replacing good stranded copper wire for good solid solid wire isn't cost effective, no bang for the buck. Stranded wire has more surface area and will conduct more current for this reason alone.

The wire AR used is quite adequate since we aren't talking high current loads, but it is old. If you suspect there is or might be corrosion, then by all means replace it, but IMHO, stay away from the expensive stuff.

If your wondering, I did replace all the wire in my AR90s with Kimber Teflon coated wire and it didn't make a difference anyone could hear.

[Robert_S]

Perhaps of interest, some previous discussion on the subject of the AR90:

• AR90 Crossover rebuild

http://www.classicsp...p?showtopic=850

• New member! Just bought a pair of AR-90's, need advice

http://www.classicsp...?showtopic=1158

• AR90 crossover re-cap: Remove board or work in situ?

http://www.classicsp...?showtopic=1554

• AR-90 refinishing update

http://www.classicsp...?showtopic=1279

Regarding capacitors, I can't speak from personal experience, but the brand ClarityCap may also be worth considering (supposedly very good value)?

http://www.claritycap.co.uk/

http://www.hificolle...capacitors.html

Additionally, the Blueprints (page 1&2) and the Schematic that we presently have in the library are quite low resolution. Here are the more detailed versions:

And for visual reference, some images (from a recent online auction) that give a clear overall view of the crossover:

Hope that helps!

Robert_S

Edit: See also the attached PDF, prepared by Rlowe, which shows a schematic of a crossover version that does not employ a circuit board for the level resistors.

ar90-schematic2009_Riowe.pdf

[Kuja]

Aleksandar

I think the Mundorf's are not a cost effective choice. They are very expensive and used in VERY high end speakers.

Thanks Carl!

Mundorf has different product ranges.

I was thinking of getting caps from their lower mid offer - the MCap series.

Their prices seem quite OK to me.

I attached the MCap page from Mundorf brochure with specs and prices.

Prices are in EUROs.

What do you think of them?

[Carlspeak]

Thanks Carl!

Mundorf has different product ranges.

I was thinking of getting caps from their lower mid offer - the MCap series.

Their prices seem quite OK to me.

I attached the MCap page from Mundorf brochure with specs and prices.

Prices are in EUROs.

What do you think of them?

Those prices for that line look quite reasonable. I don't that line is offered here in the US. Give them a try. They should be quite good considering the brand's high standards.

[rrcrain]

Holy Pete,

one of those threads was started by me... 4 years ago when I recapped my AR90s. I didn't realize it was that long ago!

[Diamonds&Rust]

Ok...

Here are the photos of caps that I have in my AR-90 crossovers.

All caps are original.

Some of them (on tweeters and upper midranges) do not seem to be electrolytic.

There are some white caps with yellow ends and some black caps with red ends.

What type are they?

If one would replace like for like, what would be today's closest equivalent?

The photographs are useful.

The black capacitors with the red caps are, unquestionably, Callins brand non-polar electrolytics.

A close equivalent does not exist in this country. Low ESR, high quality, non-polar electrolytic capacitors of whatever brand would be the closest replacement. I looked for some and months later gave-up. Apparently there simply are none available.

Contrary to intuition, the larger can-type capacitors have lower ESR than their smaller counterparts **of the same era**. It seems that if you leave the plates the same thickness and make them larger, by increasing the surface area you drive ESR down. And there are a lot of other differences <he says in teeth-gritting frustration>. If there is a modern little electrolytic capacitor that is acceptable, neither the "Bennics" or the ones with MDF printed on them, nor the Nichicon, nor the...starts with a J... made me happy.

I have never seen the white 100v capacitor in an AR crossover before. I have no idea what it is.

When I first replied to you about what capacitors to use I was trying to keep it simple. And look at all the trouble that caused.

When you start reading all those threads, please note that some of us have become older and wiser.

There has been something worth hearing from everyone, even people I disagreed-with almost violently. I have learned to have less doubt of other people's experiences.

Pete B =-evidently-= got good results with the capacitors he used once he added a little resistance. I don't want to be accused of misrepresenting his experience, so if I'm remembering incorrectly what Pete ultimately said I wish he would come and correct me.

Roy has had good experience with what he has used; sometimes that means mylar (which is what Ken used on the tweeter of the 303 he designed), sometimes it means a film cap with a resister, sometimes it was just a higher-priced film resistor.

And others have had success with different things. Some people like their Solen capacitors with no by-pass and no resistence, for that matter.

I've had bad experience with a relatively inexpensive film capacitor until I "jumped" it with a "by-pass" capacitor - but I might have liked it a whole lot better if instead of "by-passing" it I had added a little resistance. The "by-pass" made it easier to live-with, but the resistor might have made it perfect. I don't know. I didn't try it. It would have been a great deal cheaper.

I've had good experience with a more expensive film capacitor, but I will tell you honestly, if I could achieve the same thing by buying cheaper (but not garbage) film caps and applying a little resistance I would much rather do that next time. It's cheaper by a mile.

The only thing I have tried that was simply not acceptable was buying inexpensive NPEs as replacements. There was essentially no difference in them and a still-within-spec 25 year old (at that time) capacitor which I think we all agree needed to be replaced. The difference between them and film capacitors wasn't strictly in the "amount" of signal they passed.

I have not heard any difference, at all, in internal speaker wiring. If all that expensive silver tweeter wire and concrete reinforcing steel-bar-thick woofer wire does anything at all I can't hear it in an A/B situation - I think it's a waste of money for the lengths we use.

How do you plan on reaching the large values? If you plan to "build-up" the value by paralleling smaller capacitors, I think several of us would tell you to replace some of the resistence you will be missing. Alternatively you should look for an old-style (but newly manufactured) can-type NPE to use. You could run into problems by using a large NPE as the base capacitor and adding a smaller film cap to it. Don't ask me why, I have no idea, I just know it can be a problem. Use a NPE by itself or use a film (or two, or three) and add a resistor.

Bret

[Diamonds&Rust]

Perhaps of interest, some previous discussion on the subject of the AR90:

It may be worth noting that "nathanso," the skeptic who used inexpensive NPEs when he recapped his 90s got rid of them not too very long after recapping them.

Why his 90s were so easily defeated in listening tests I will leave to the imagination of the readers.

Bret

[Robert_S]

It may be worth noting that "nathanso," the skeptic who used inexpensive NPEs when he recapped his 90s got rid of them not too very long after recapping them.

Why his 90s were so easily defeated in listening tests I will leave to the imagination of the readers.

Bret

An astute observation! The possibility of said events bearing a linear relationship, is undeniable.

(Indeed, the devil resides in detail . . . and so, to recap , where were we?)

Robert_S

[KlausDK]

Can you be more specific about details of your AR-90 restoration?

1. What caps did you use?
   
2. Any resistors in series due to different ESR?
   
3. What do you think of Mundorf from Germany?
   
4. What about the AR-90 internal wiring?
   
5. What did you do with those levels switches on the back? Cleaned them, or bypassed them?
   

Hi Alexander

1: I used Mundorf M-Cap Supreme for the tweeter caps bypassed with small 0,1 myF polystyrene caps, an SCR (SOCIETIE DES CONDENSATEURS RECORD)=Solen for the upper midrange (they happened to have a 24 myF) and Jantzen Cross Cap 82 myF for the lower midrange. The M-Cap supreme sounds better than the standard polypropylene capacitors (like SCR and Jantzen), however they are very expensive, and I am not sure I would buy them today. Today I would I probably use the Jantzen Cross Cap capacitors (wound on the Finnish Tervakoski foil) for all three drivers, as I find that they sound brilliant for a very moderate price. I find them marginally better than the Solen/SCR caps.

For the capacitors parallel to the drivers, I would use cheap film or bipolar electrolytic caps.

2: I have never compensated for ESR changes.

3: I think Mundorf make sexcellent caps, if you have easy access to the standard line at a reasonable price, I would use them.

4: I use 2½ square mm OFC cable, nothing special or expensive.

5: I left the switch box for now but changed the switches to a more solid version. I will take the upper crossover board outside the speakers, and replace with a peice of masonite. When I have listened in the speakers and found that a level correction of the tweeter or midrange might be needed, I will propably bypass the switchbox with resistors or wires as needed.

BRgds Klaus

[rrcrain]

OK,

Here's what I did 4 years ago and my thoughts on what I did.

I replaced all of the caps with Solen 400VDC poly, the inductors with perfect lay coils and rewired using Kimber project wire. I also split the crossover to enable bi-amping and routed out a proper hole for a new connector cup for the new connectors.

What I would not do again.

I'm absolutely positive the new wire didn't do a thing detectable. Didn't hurt, but it was a waste of money as far as I'm concerned performance wise.

The new inductors probably had an impact on sound, but cost wise, I'd also rate it as money not well spent. The advantages of the new inductors are each one is exactly a known, repeatable resistance and inductance. Most inductors vary in resistance due to minor differences in wire length. Probably not enough to worry about in the real world.

I would not remove the attenuator circuit.

My opinion of the speakers before recapping was they were incredibly bass heavy, and why the heck didn't they give me the option to attenuate the bass? Over the years, I'd forgotten what they sounded like new and was absolutely convinced AR made a horrendous design blunder, so I removed the cirtuit and disposed of the switches and associate resistors.

In retrospect, AR was very wise to install the three switches. Depending on the room acoustics, the speakers can now be over bright. In our current home, they were fine in the living room with a knotty pine vaulted ceiling, but downstairs in the now finished basement, they are a bit to detailed for my taste now and then. Sarah Brightman can almost pierce ear drums.

What I'd do again

As for the Solen caps. I would reuse them but would probably investigate bypassing a bit more to soften their edge. New, the highs were painful but they settled down after a couple hours.

I would wire for bi-amping again.

If you use a decent high end active crossover with the crossover points set properly, this single option has the potential of making a wonderful improvement. without the active crossover, my opinion is it's worthless but feel free to experiment.

The active crossover does allow me to attenuate the high end. (Marchand XM26) and I do have an equalizer if I get desperate.

I would spike te speakers.

My speakers generally sit on reasonable deep carpet and the cabinet is at best, unstable. Spiking connects the speakers to the subfloor making them very stable and incredibly, tightened up the bass.

Overall tho, my AR90s compare extremely well to some very expensive, well made Itallian and German speakers currently on the market so I can't complain to loudly.

And Sarah Brightman pierces ear drums on those speakers as well...

Richard.

[blastinbill1]

I am recapping my AR90's (originally purchased in1979). I've already soldered the 4, 6, and 8 microfarad capacitors in place (the 6 and the 8 mylar from Madisound, the 4 a poly from PE). I have checked capacitance values on all original 16 capacitors in the two speakers and have discovered that 6 out of 16 varied more than 10% from stated value, and 3 out of 16 varied more than 20% from stated value. Most of the worst ones, as expected, were the electrolytics, and the black and red ones (I don't think they are electrolytics) were all less than 10% off except for one of the 40's in the upper midrange, which measured 46.6 (17% high). The original tweeter cap values were all between 1% and 9% high.The two 80's in the lower midrange measured 91 and 98 (14% and 23% high, respectively) and the two 30's in the lower midrange measured 30.6 and 39.1 (2% and 34% high, respectively). The big aluminum can woofer electrolytics measured 353, 356 (less than 2% off) so I can't justify replacing those two.

I've decided to bypass all three level switches (I never used them in 34 years) which made it easier to solder in the 6 microfarad capacitors without all the level switch resistors being in the way. I noticed that all the electrolytics I've ordered from PE seem to measure high, and the new replacement 80's from PE in particular measured 87.3 and 87.6 out of the box (almost 10% high). I have ordered combinations of electrolytic caps from Madisound and PE to add up them up in parallel to get closer to the correct measured values for remaining caps (24, 30, 40, and even the 80's again).

Blastinbill

[rrcrain]

I am recapping my AR90's (originally purchased in1979). I've already soldered the 4, 6, and 8 microfarad capacitors in place (the 6 and the 8 mylar from Madisound, the 4 a poly from PE). I have checked capacitance values on all original 16 capacitors in the two speakers and have discovered that 6 out of 16 varied more than 10% from stated value, and 3 out of 16 varied more than 20% from stated value. Most of the worst ones, as expected, were the electrolytics, and the black and red ones (I don't think they are electrolytics) were all less than 10% off except for one of the 40's in the upper midrange, which measured 46.6 (17% high). The original tweeter cap values were all between 1% and 9% high.The two 80's in the lower midrange measured 91 and 98 (14% and 23% high, respectively) and the two 30's in the lower midrange measured 30.6 and 39.1 (2% and 34% high, respectively). The big aluminum can woofer electrolytics measured 353, 356 (less than 2% off) so I can't justify replacing those two.

I've decided to bypass all three level switches (I never used them in 34 years) which made it easier to solder in the 6 microfarad capacitors without all the level switch resistors being in the way. I noticed that all the electrolytics I've ordered from PE seem to measure high, and the new replacement 80's from PE in particular measured 87.3 and 87.6 out of the box (almost 10% high). I have ordered combinations of electrolytic caps from Madisound and PE to add up them up in parallel to get closer to the correct measured values for remaining caps (24, 30, 40, and even the 80's again).

Blastinbill

Bill,

I hadn't used the attenuator switches in years either, which is what prompted me to remove them. What I realized after the fact was the caps in my crossover were so poor that the tweeters were effectively dead as were the upper midrange speakers. In fact, I discovered one of the upper midrange drivers was bad, and had been bad for some time; The voice coil had popped out of the magnet groove. before rebuilding the crossovers, the speaker with a good upper midrange sounded identical tothe speaker with the bad upper midrange.

You will be amazed at how bright the AR90 can be after installing new caps and may regret the decision to bypass the switches.

As for the 350uf cap in the bass section, use a 200 and 150 in parallel.

[blastinbill1]

Rich,

Thanks for the reply. I have healthy output from all drivers already...I suspect the crossover frequencies will be better locked on after this. I wasn't going to even replace the 350uf capacitors because the capacitance values were so good. How bad could the ESR be on those anyway? Looking at the schematic, there are three wire disconnects to the lower board, and those cap values may be easily checked by disconnecting those wires at any time. So I was going to leave the 350 in the woofer crossover for now and check it periodically. I could always wire a bypass capacitor in parallel with it if ESR was a problem.

[rrcrain]

Most of the caps I removed were technically within spec as far as capacitance was concerned. Leakage and ESR on the other hand is another story.

Aluminum foil electrolytics caps do fail as they grow old. The electrolyte dries out, the oxide layer thins and the worst that could happen is an internal short taking out an amplifier.

With absolute certainty, I can tell you I've no real idea if replacing the 350uf caps in my speakers made a difference or not, but I felt if I was going to spend the time and money to rebuild the crossovers, I'd do all of it. Solen caps aren't that horribly expensive and will do very well in these speakers.

Are there better, or at least more expensive caps? Without a doubt, yes. Will you "really" here a difference with 30+ year old drivers? My guess is no but I wasn't willing to spend the money to find out.

From my own experience, I do recommend the following:

1: replace all of the caps. You will not regret this and will be amazed at the improvement.

2: leave the coils. The originals are decent quality. Better can be had, but the gain, if any, will be marginal.

3. leave the switches. At lease until after you've heard the speaker with the rebuilt crossovers.

4. I'm not a big proponent when it comes to wire. If the original isn't green, leave it. If you want to replace it, that's fine. I'm not one of those that believes in the magic some claim expensive "oxygen free" or silver wire offers.