AR9 capacitor questions

[SHOKDU]

Hello all! I am in the last stages of acquiring parts for my upcoming AR9 rebuild. I just have a couple of questions about the two largest capacitors. They are obviously electrolytics, but the question I have is about polarity. A couple of schematics I've come across indicate that they may be bi-polar. Is this so?

I already have these in stock, but they are not bi-polar, I would like to use them if possible.

http://www.digikey.com/scripts/DkSearch/dksus.dll?WT.z_header=search_go〈=en&keywords=604-1097-nd&x=0&y=0&cur=USD

http://www.digikey.com/product-search/en?x=0&y=0〈=en&site=us&KeyWords=604-1133-nd

Thank you!

[SHOKDU]

I have to place an order at PE anyway. Voltage rating on these is right, so I may just use 5 of these in parallel to reach 2500uF and then the same 500uF will also work for the 470uF capacitor as well. That's well within 10% of 470, so it should be fine.

http://www.parts-express.com/pe/pshowdetl.cfm?&Partnumber=027-378&scqty=12

[xmas111]

I bought all (I think all) the caps from Madisound when rebuilding my 9's. You can use the Bennic non-polar electrolytic caps.

Two 1000µf and one 500µf for the 2500µf.

one 140µf and one 330µf for the 470µf cap.

John

[xmas111]

I put together a capacitor parts list for the AR9's if it helps out.

It's a little pricy for everything but certainly worth it.

http://johnsweather.com/web%20posts/AR9%20Capacitors.xlsx

John

[Joel]

When I did my AR9's, I left the 2500µf original. The research I did at the time was controversial as to whether these should be changed or not. What is the current thinking on this?

-Joel

[SHOKDU]

John, thanks for the input. I plan on using mostly daytons from PE, but I do like the idea of using only three caps to reach the required 2500uF, so maybe I'll order those from madisound. The resistors are coming from madisound already, so I'll just add those to my order. The Daytons are just rebranded bennics anyway, so I imagine our speakers will sound very similar.

Joel, all electrolytics are well past service life at 15 years, so at 30 they should certainly be replaced. I guarantee that any modern electrolytic or film cap is better than what AR used back in the day, so pretty much any cap on the market that is geared towards musical use should be an upgrade. Old caps can short out. I'd say that anybody running these with anything original in the crossover other than the inductors is running a risk. You never know when old electrical components are going to short. If I were you I'd open yours back up and replace the 2500's.

[SHOKDU]

All capacitors and resistors ordered. Work will begin this Friday. I expect that if I'm able to put in 8 hours on Friday and a few more hours on Saturday that I will be done by Saturday evening. Restoring the cabs will be an ongoing project over the next few weeks.

[KlausDK]

Joel, all electrolytics are well past service life at 15 years, so at 30 they should certainly be replaced. I guarantee that any modern electrolytic or film cap is better than what AR used back in the day, so pretty much any cap on the market that is geared towards musical use should be an upgrade. Old caps can short out. I'd say that anybody running these with anything original in the crossover other than the inductors is running a risk. You never know when old electrical components are going to short. If I were you I'd open yours back up and replace the 2500's.

I generally agree, but I actually once took a set of 2500 uF Sprague bipolar caps I had removed from a set of AR-10Pi with me to Jensen Capacitors here in Denmark to have it measured out on professionel equipment. They found that they were in perfectly working order and within specs regarding all important parameters like capacitance, ESR, and some other I don't remember anymore. These old Sprague electrolytic caps were manufactured of very high quality.

Just to confuse matters

BRgds Klaus

[Carlspeak]

I generally agree, but I actually once took a set of 2500 uF Sprague bipolar caps I had removed from a set of AR-10Pi with me to Jensen Capacitors here in Denmark to have it measured out on professionel equipment. They found that it was in perfectly working order and within specs regarding all important parameters like capacitance, ESR, and some other I don't remember anymore. These old Sprague electrolytic caps were manufactured of very high quality.

Just to confuse matters

BRgds Klaus

This agrees completely with the findings of my testing of Compulytic caps shown here > http://www.classicspeakerpages.net/IP.Board/index.php?showtopic=6354entry90734

[RoyC]

The Sprague Compulytics have held up very well. Results of extensive testing conducted by John O'Hanlon were consistent with Carl's findings.

Further, "all electrolytics are well past service life at 15 years" is simply not an accurate statement based on my experience..

Roy

[Mach3]

I would very strongly suggest that you wait and ponder BEFORE going any further on your AR-9 cap replacement. Mainly because there is some doubt as to the values recommended in the "official" schematic.

For instance it is widely held that not only does the "official" schematic (available at this site) not contain a shunt coil for the MR crossover - but also that the value of 6 uF that is also in shunt with the MR driver should actually be an 8 uF cap.

I am posting two files which I found (somewhere on the Internet - perhaps here) that contains the "as measured" values for a set of AR-9 crossovers - and that these as-measured parts from an untouched crossover (factory built) are in many cases far different from those on the "official" schematic. One is for the "UMR" (the upper mid-range) and one is for the LMR (the lower mid-range or 8" driver). The "blue" values are those parts which differ significantly from the official schematic - the "red" values are those parts not even noted on the official schematic.

So proceed with some caution - I am in the process of "rebuilding" my crossover for the second time - and doing so to hear if moving to the larger values in the "as measured" schematic actually change the sound for the better.

BTW: - make sure that the wooden "can" behind the 8" driver is still securely in place - that particular device is usually loose on our 30 year old speakers - and it has to be tightly bound to the cabinet as it forms the air suspension behind the 8" driver.

best of luck

[RoyC]

I am posting two files which I found (somewhere on the Internet - perhaps here) that contains the "as measured" values for a set of AR-9 crossovers - and that these as-measured parts from an untouched crossover (factory built) are in many cases far different from those on the "official" schematic. One is for the "UMR" (the upper mid-range) and one is for the LMR (the lower mid-range or 8" driver). The "blue" values are those parts which differ significantly from the official schematic - the "red" values are those parts not even noted on the official schematic.

It is not wise to be using "as measured" values for any restoration! Original schematic values and component labels are a much safer bet.

Why use measurements of old electrolytic caps found "somewhere on the internet"? How were they measured? Who measured them? It should also be noted that the original caps had a 10% tolerance AND that electrolytic caps drift upward in value as they degrade, so "measurements" showing higher values than those depicted in the schematic would be FAR more suspect.

The values printed on the original caps are the safest to use.

As for missing components, such as inductors, AR made many changes along the way as drivers were modified. It is not always known which modification preceded another, so it is best to replace components in kind with what is found in your cabinets. Inductors, "missing" or otherwise, do not need to be replaced unless they are obviously damaged or have been removed by an aftermarket modification.

Roy

[Carlspeak]

It is not wise to be using "as measured" values for any restoration! Original schematic values and component labels are a much safer bet.

Why use measurements of old electrolytic caps found "somewhere on the internet"? How were they measured? Who measured them? It should also be noted that the original caps had a 10% tolerance AND that electrolytic caps drift upward in value as they degrade, so "measurements" showing higher values than those depicted in the schematic would be FAR more suspect.

The values printed on the original caps are the safest to use.

As for missing components, such as inductors, AR made many changes along the way as drivers were modified. It is not always known which modification preceded another, so it is best to replace components in kind with what is found in your cabinets. Inductors, "missing" or otherwise, do not need to be replaced unless they are obviously damaged or have been removed by an aftermarket modification.

Roy

Boy, do I agree with that!

[Mach3]

The poster from which I got this "as measured" schematic measured the components in two crossovers - from a pair of AR-9 speakers. In BOTH cases the as measured values were identical. So I don't think it was "upward" drift of capacitance. I believe Holl was tweaking and measuring throughout the production life of the speaker. Interestingly enough - if you look at the attached schematics - the coil values were perfectly in line with the "published" schematic.

As for the 6 uF versus 8 uF in the cap in parallel with the UMR (large dome) this has been attested to in a number of articles - and when I removed my "original" it was an 8uF - but since the schematic called for a 6 uF I went with the published value - and to my ears brought some brightness to the region (around 7 KHz I have a 3 dB bump - which means that my upper crossover limit is not falling as steeply as it should - which the larger cap value will/should ameliorate).

If you look even at this posting there are a couple of posts (8 and 9) where the posters are expressing satisfaction with how well the original AR caps hold their value - so which is it? Are the caps indeed holding their value OR are they all uniformly drifting towards identical higher capacitance values? (such uniform drift is somewhat unlikely as if capacitance value change were the culprit I would expect to see some values LOWER than the published schematic). One supposition or the other must be incorrect.

The poster of the "as measured" schematic included this at the top of his post;

Crossover AR9 – Switches have been omitted – Serial 010496

“e” indicates expected value according to AR spec.

Red components are not mentioned in AR spec, but were in my AR9.

Blue values indicate those that are very different from specification.

Values are similar for 2nd speaker and schematic is exactly the same.

In the end while I do respect the work that AR and particularly Terry Holl did with the AR-9 (a wonderful speaker to my ears) I worked long enough in engineering to understand that the documentation is NOT always in line with what is built. This is especially true with a smallish, non-DoD enterprise such as AR.

So how can the OP be hurt by being careful and measuring what he is replacing rather than blindly throwing parts in based on a provably INCORRECT schematic? (provable in the sense that all AR-9s have the extra coil in shunt in the UMR crossover but no schematic reflects this change).

Shokdu - if you have the time, equipment and inclination would you please share with us what the values of the caps that you replace actually are? This would be very helpful to all future work with the "Mighty Nines".

Thanks

[Carlspeak]

"

Shokdu - if you have the time, equipment and inclination would you please share with us what the values of the caps that you replace actually are? This would be very helpful to all future work with the "Mighty Nines".

I presume you mean the PRINTED values on the caps?

[Mach3]

I would prefer - if such is not a big bother - both the "printed" value AND the measured value. Since most DVMs in this era also include capacitance measurement capabilities this should not be too burdensome. That would go a long ways for preventing another "tempest in a teapot" on this forum. More importantly this would be some really great information for all rebuilders, tweakers and modifiers.

But the "printed" value will be great and gratefully thanked if that is all you can do.

so thanking you in advance,

D

[Pete B]

"Computer" power supply grade caps and Sprague Compulytics are built with

multiple attachment points from the connectors to the plates. This provides

much lower ESR than the typical electrolytic capacitor. If you replace them with

a typical crossover cap, I'd use 5 in parallel, if I had to guess at a number, since

the multiple units will be similar to multiple attachment points. I don't know exactly

how many attachment points are used but I'm fairly certain it is considerably higher

than 5. Back to back modern power supply polarized caps might be better but they

are hard to find in 10% tolerance.

The shelf life for electrolytics caps specified in US Military requirements is 10 years

IIRC. Certainly, they usually do better than that, but 20 to 30 is really pushing it.

We are lucky that Sprague used a long lasting formula for the electrolyte and I'm

not sure what I'd do if I had a pair of 9s. I'd probably measure the caps and see

what it would take to do as good or better for ESR with multiple standard caps.

[RoyC]

n the end while I do respect the work that AR and particularly Terry Holl did with the AR-9 (a wonderful speaker to my ears) I worked long enough in engineering to understand that the documentation is NOT always in line with what is built. This is especially true with a smallish, non-DoD enterprise such as AR.

So how can the OP be hurt by being careful and measuring what he is replacing rather than blindly throwing parts in based on a provably INCORRECT schematic? (provable in the sense that all AR-9s have the extra coil in shunt in the UMR crossover but no schematic reflects this change).

Nobody is recommending the purchase of a pile of caps based on the schematic alone. AR changed many models on the fly, including the AR-3 (several times), AR-3a/2ax/4x/5 and 6. The schematics in the forum library do not always reflect these changes. The best advice is to replace the capacitors with the same value caps original to the speakers being restored. It is a simple matter of reading what is printed on the original capacitors!

The crossover changes may not reflect improvements. Most AR crossover modifications I have seen appear to be adjustments for alterations made to the response of later drivers.

I would never trust someone's "measurements" enough to modify a crossover original to the speaker I am working on. Measuring caps is always of interest, but will only tell us if they are still within the values and tolerances printed on the casing.

Roy

[SHOKDU]

I have enough capacitors in stock now to follow either the schematic or the "as measured" values. In my experience, it is always best to use the same capacitance AS MARKED on the capacitor you are replacing. This is especially true in amplifiers, receivers, and any old PCB requiring re-work. As a general rule you should never follow a schematic, you should always replace with like values, printed on the capacitor you are replacing. This is only true when replacing original parts though. If the particular item you are refurbishing has passed hands before, and underwent work where capacitors were changed from the factory originals, then you really have no way to tell if what is now on the board, or what is on the schematic is correct. For example, I've had odd instances while rebuilding amplifiers where the schematic will show a polarized capacitor, and in reality the board came from the factory with a non-polarized capacitor. Reasons like this justify always following what is currently installed, as long as it was a factory part. If it's been replaced before and you have no idea which it should be, then God be with you as you make the decision to follow what is being used, or what is on the schematic.

I will certainly pay attention to, and replace with values close (within 10%) to what was originally installed in the speakers. I will gladly record what is printed on the capacitors, and what I replace them with. With the other request regarding testing actual capacitance...I don't know what DMM you're referring to, but I don't know of any handheld units that have a way to test capacitance value. I use fluke 87's and the only way to really test a that a capacitor works is to set it to read ohms, basically charge the capacitor, then switch to the VDC rating and watch the discharge of the capacitor.

When I said 15 years is typical service life, that is generally the "rule". As somebody earlier mentioned, mil-spec is 10 year shelf or usage life. Mil-spec is a bit hardcore, but they are critical applications that kind of require that level of dedication for replacement. Of course, I did not mean to say that at or past 15 years your capacitor is completely shot, but they are deteriorating and will require attention at some point. I prefer to be safe, and at the age of these capacitors, I prefer to not take a chance. You never know when one could start leaking, and worst case short out.

[soundminded]

I recapped my AR9s four years ago while the woofers and LMRs were out for refoaming. I did not change the two big metal can caps. I used Dayton NPCs for the rest. I checked the values on the caps themselves and compared them to the schematic in the library. There was in fact an error on the schematic and I used the values on the original factory caps. The results were entirely satisfactory. Changing the old caps is very important to prevent damage to the UMR and especially the tweeter. As was once pointed out, as the dielectric deteriorates over time, the value of the cap actually increases. This lowers the crossover frequency for series wired caps. This would not only create the possibility of distortion but of damage to these next to impossible to replace drivers. I measured the values of the caps that came out. Some were within spec, some fairly far off. Even for those that measured good, I changed them anyway.

[Mach3]

Several posters have stated that as electrolytic caps change over time their capacitance tends to "increase". This to me is an interesting obvservation which I have never heard before - which means nothing as the concepts I don't know far exceed those that I do know (but I would speculate that such is true for all men).

Does anybody have a definitive study or report on this phenomena of age based "increasing capacitance". Or is this just one of those "common knowledge" types of phenomena? For if it is "common knowledge" then for myself it goes into the bin of "urban myth" - much like power cords changing the sound of your amplifier or any of the other urban myths (uber rigid cabinets, pistonic drivers, et cetera) that are so popular in audiophile style discussions.

Science and engineering do not rely on "common knowledge" or on "urban myths" - instead these disciplines are based on exhaustive studies generating data supporting well formulated theories. Work like Dr. Toole does.

So please - a link or some other reference to this phenomena would be much appreciated.

[Mach3]

'SHOKDU'

I will gladly record what is printed on the capacitors, and what I replace them with. With the other request regarding testing actual capacitance...I don't know what DMM you're referring to, but I don't know of any handheld units that have a way to test capacitance value.

Thanks a lot for your efforts Mr. Shokdu - you are providing valuable information for every person who ever attempts an AR-9 crossover rebuild - and for almost everybody else as well.

In re the DVM that does the capacitance measurement - I just bought a Triplett Model 9007 (about USD $30 or so) that has built in capacitance measurement capability. As to how the meter does this I have no idea - since micro-processors are so inexpensive in the current era (small 16 bit style go for approximately $2 in bulk) this model may very well time the charge/discharge cycle and use that to determine capacitance. I have no information on the accuracy of such - though new capacitors I have measured using this meter seem to indicate that it is very accurate (within +/- 5% per the "Dayton Audio Precision Caps" I bought from Parts Excess.) To establish an absolute I would need some "known" capacitor and then compare.

Interestingly enough EVERY DVM (DMM is actually more appropriate I guess) that I looked at had a capacitance measurement capability. So such seems rather common place in the hear and now. I was surprised as well.

[Carlspeak]

Several posters have stated that as electrolytic caps change over time their capacitance tends to "increase". This to me is an interesting obvservation which I have never heard before - which means nothing as the concepts I don't know far exceed those that I do know (but I would speculate that such is true for all men).

Does anybody have a definitive study or report on this phenomena of age based "increasing capacitance". Or is this just one of those "common knowledge" types of phenomena? For if it is "common knowledge" then for myself it goes into the bin of "urban myth" - much like power cords changing the sound of your amplifier or any of the other urban myths (uber rigid cabinets, pistonic drivers, et cetera) that are so popular in audiophile style discussions.

Science and engineering do not rely on "common knowledge" or on "urban myths" - instead these disciplines are based on exhaustive studies generating data supporting well formulated theories. Work like Dr. Toole does.

So please - a link or some other reference to this phenomena would be much appreciated.

I can't point you to a difinitive study, but I can say that while writing an article for audioXpress mag. (May 2011 issue) on testing capacitors, I tested many caps then and have done so for some years now for customers for whom I do xover upgrades. I can say that for the most part, old NPE's do drift upward in their uF value from target and also ESR. Some more than others. The latter statement is covered in this thread http://www.classicspeakerpages.net/IP.Board/index.php?showtopic=6354

[Mach3]

Carl - excellent and very high quality work. Thank you very much.

[fast_eddie_72]

I happened to have the caps from my 9s in a box.

Values as marked

4, 6, 24, 30, 40, 80 and 470

I have not yet done the 2500uF caps.

I'll say this - I've recapped quite a few speakers and it made the least difference I've heard on the AR9s. My LST2s had the blue and purple Compulytic caps and many told me that they should be fine, but it was a night and day difference. Not the case with my 9s. I did what I normally do and redid one then listened to the pair. Normally it's a clear difference. On the 9s, they still sounded like a *pair* of speakers. Imaged well and sounded good.

I haven't measured the capacitance on the old caps. I may do that later. I don't have an ESR meter. For what it's worth, my LST2 caps tested in spec for capacitance but they were clearly not doing well. I suspect ESR was very hight.