AR9's - Let's Shop for a Recap

[Stimpy]

41 minutes ago, DavidR said:

Stimpy, the 350uF in the AR90 is a shunt cap and the 1 ohm resistor make it a zobel circuit. If I understand it all correctly all frequencies are 'shunted' to ground and you shouldn't be able to hear any difference between parts.

That's what I thought too.  But, if it can be believed, there's been discussions on forums, to where owners reported hearing positive differences in sound, when changing shunt caps and resistors.  Jeff at Sonic Craft, offered an explanation as to why.  I'll have to l see if I can look up his reply concerning this.

[valkyrie]

I recapped my AR-9s about six years ago.  I was preparing to sell them as they sounded rather ordinary.  Upon recapping I was amazed at the sound change - so good that I sold the "other" speakers and kept the Mighty Nines.

As for caps?  I bought Solens, Jantzens and Daytons - all film for ALL the caps including the 480 uF (which is gigantic in film).  I did NOT pay any attention to the type of film; polyester, polystyrene - whatever.  What I did do was to get +/- 5% tolerance caps.  That way ensuring that I was as close as possible to the design values.

I replaced ALL caps - even the 2500 uF cap in the woofer section was replaced (five 500 uF NPE in parallel - more on this later).  I also rewired the entire speaker - I felt that the stock wire was a bit on the thin side - and I did NOT like the push on connectors - soldered everything were it was supposed to be.

Results?  STUNNING - THE SOUND NO LONGER CAME OUT OF THE CABINET BUT MATERIALIZED ABOUT A FOOT IN FRONT OF THE SPEAKER CABINET.  It was so good that I called a friend over to confirm what I was hearing (avoiding confirmation bias which strikes us all).  Excellent clarity, tons of detail, incredible dynamics - just a new speaker - in fact a speaker that competes, imo, with just about anything available.

Point being?  I know that I spent close to $500 on caps (i did not replace the stock resistors) - but let me assure you that the investment is easily worth the results.  Best $500 I have ever spent - at least on stereo stuff.

Couple of pointers:  There is a "can" made of cardboard tube and a thin plywood top that sits behind the 8" midbass coupler (lower midrange).  That thing was, from the factory, glued to the cabinet.  It will be loose at this point.  Go in there, glue it back on - I used a brace between the backwall and this assembly to ensure it stays put.  This "can" assembly provides the air suspension for the 8" LMR.  Critical assembly and it must be air tight. 

Second; the crossover is on two boards, one on the bottom of the cabinet, one on the back wall.  The one on the bottom will come loose - design feature.  See the picture I included.   On this board (which is for the LMR and bass crossover) reorient the coils so that they are at 90 degrees to each other.  Eliminates cross coupling between the field. 

When rebuilding do ONE crossover at a time - save the other, in original state, for reference.  You will probably have an extra cap in parallel for the UMR crossover - this is NOT on the schematic.  If you have it?  Replace it.

It took me about a week of steady work to rebuild both crossovers.  I hauled the speaker to my bench and humped the thing up there - not easy for an old man - and perhaps beyond me now.  Used a hand cart between the house and the garage - those are LARGE HEAVY SPEAKERS.

Go slow, be extra precise, one part at a time.  Solder well.   Most of your work will be done - at least on the upper crossover, while reaching through a woofer hole - you will burn your hands/fingers at least once.  Get over it. Be PATIENT.  Review every step you take after completion.

[valkyrie]

Here is another picture of where you are going to be working.  I used a "pencil" style iron - you do NOT need a lot of heat to remove/replace parts.  Make sure you have some "solder wick" to suck up old solder.

Second note:  Tom Tyson - the heart and soul of this board - will no doubt shit a brick when I suggest this (the man is a purist through and through), but if you have a powerful MODERN SOLID STATE AMP (NO vintage receivers can do this), just throw out the 2500 uF cap in the lower crossover.  The thing was put there to protect the kind of amps that existed in 1979 - those amps did NOT like < 2 ohm loads (which is what those two 12" woofers look like at 28 Hz).

If you employ tube amps?  Then you are a Luddite and should not make this modification.  It will not work with effete tubes.  

A modern amp - say something from Odyssey Audio, or Pass, can handle a two ohm load and not even break a sweat.  The result of ridding yourself of that ridiculous cap will be bass like you have never heard in your life.   I mean DEEP, ARTICULATE AND TAUT BASS.  But you must have an amp with some ass to it to avail yourself of this approach.

Now the Tom Tysons, Roy C. and the others that know all will no doubt have their panties in a major wad at this heretical suggestion.  My response?  Those are YOUR SPEAKERS - not theirs.  You can try it - it only takes cutting ONE WIRE and if you don't like it or your walls are crumbled by the bass?  Then you can simply resolder the connection.

[DavidDru]

Thanks Valkyrie.

 

Do you have any photos of your completed boards?

[Stimpy]

The idea of removing the 2500uF bass shunt cap confuses me.  It's not that the idea repulses me, since it's a change from stock, but from what I've read, that cap isn't there just for amp protection.  If I recall correctly, it's also part of an EQ circuit, to enhance and extend the bass response of the speaker.  Without it, you actually loose some of the bass extension that the speaker is capable of.  While the bass most certainly will change without it, is the change truly an improvement, or just a change?  Is the bass actually better, or just different?  More mid-bass, and less low bass? 

It makes me happy that I own the 90s and not 9s.  As such, I don't have to worry about it.  I don't have to expend any more effort, than the thought process required to post this reply!     Though, I'm glad that Valkyrie used decent poly caps, and noticed an improvement in the sound of the speakers.  That's encouraging to hear, for those of us that want to try poly recaps too.

[DavidR]

I think you are right Stimpy - it provides bass extension as well as ohm/impedance stabilization (I don't believe its a shunting cap). Even with today's amps it would be a punishing load for many amps. It would have to be stable at 2 ohms (at least).

Unless you really don't want to have to go into the cabinets again electrolytics would be fine in the woofer section. Madisound sells Bennic. You could use (2) 165uF and a 140uF = 470uF and for the 2500uF (2) 1000 + 500 caps.   https://www.madisoundspeakerstore.com/bennic-electrolytic-capacitor-100vdc/

I believe both Bennic and Dayton are owned by Solen.     http://solen.ca/about/brands/

[ar_pro]

Tom Tyson wrote this in 2005:

(1) The AR-9’s cabinet is approximately three-times larger than the volume of an AR-3a, and with two woofers, has double the radiating surface. These two standard 4-ohm 200003-type flat-side woofers are mounted in *parallel,* which gives them 6 dB more output in the deep bass, but at frighteningly low impedance (less-than 2 ohms). The large cabinet also seriously over-damps the woofers at low frequencies.

(2) AR’s solution for this was to design a “tank” or resonant circuit (designed by Alex DeCosta, a senior design engineer working under Engineering Director Tim Holl). This circuit was nicknamed AR’s “Electronic Automatic Transmission,” by AR publicists. The crossover network for the two *parallel* woofers is a choke and capacitor in series, and a resistor across these two elements, on one leg of the parallel-woofer circuit.

(3) At or near system resonance, the choke and capacitor are essentially a short-circuit; the presence of the resistor is irrelevant. As the signal frequency increases, the impedance of the resonant circuit also goes up, and the speaker system output goes down. Thus, the deep bass is pulled up as much as 6 db by the parallel connection of the speakers, but above resonance system output is pushed down to blend with the upper drivers.

(5) A critical amount of polyester cabinet wadding is placed *above* the woofer cavity brace to achieve proper damping. It is therefore extremely important not to add material or to change the density or change the material-type of the filling or to add filling to the bottom cavity. This is critical to proper performance.

The result of all this is that the lowest point in the impedance curve now is now an acceptable 3.2 ohms, and the system is flat at its 35 Hz resonance and down only 3 dB at 28 Hz.! The system is also critically damped with a Q of 0.5. This accounts for much of the AR-9s outstanding bass response, rivaled by few full-range speakers even to this day.

--Tom Tyson

[valkyrie]

Good discussion.  I will address, as best I can, the points and issues raised.

Lets start with AR_Pro's quote from Tom Tyson (and please remember - without Mr. Tyson we would NOT have this board - or at least not as information rich as it is).  That said lets look at (2).   THERE IS NO RESISTOR IN THE LOW BASS CROSSOVER CIRCUIT.  The infamous drawing showing the resistor was provided by Holl to illustrate a point - THERE IS NO RESISTOR ACROSS THE COIL AND CAP.

Next lets look at the concluding point at the bottom of Mr. Tyson's comment, and I quote; "The result is that the lowest point in the impedance curve is now an acceptable 3.2 ohms".   Hey fellows - the point of that cap/coil is to raise the resistance of the circuit so that solid state amps from the late seventies can drive the Mighty Nines.  Which is EXACTLY WHAT MR. TYSON IS SAYING. 

We don't have to engage in that exercise any longer - we now have modern solid state amps that can drive loads as low as 0.5 ohms.  All day (and night) long.

Now in regards the coil/cap providing a better blending between the two 12" drivers and the LMR (8" driver)?  Maybe - maybe indeed.  However I ran an F/R plot before and after I cut that cap loose (I left the 2500 uF in the crossover but it is disconnected) and did NOT see any boost in that region.  Caveat: I ran a room plot and rooms are notorious in what they do to bass response.  So my experience is not world wide applicable - it is possible that cutting the cap from the circuit - COULD - in fact cause a lower midrange boost from flat.  BUT IT DID NOT FOR ME IN MY ROOM.

I am looking for this F/R plot - and i will post it when I find it.  I can find the "before" but not the "after" (however I have posted both on the forum before this date).

As for subjective impressions (and these with all humans must be taken with a grain of salt - for we are ALL prone to confirmation bias, et cetera) - the bass is now deeper, tighter, more articulate and far more powerful - with the cap cut loose.  Needless to say - I enjoy it a great deal more.   There is an album by one Zach Hemsey - play track 6 which has some bass strokes that sound like "rifle cracks" - good test of bass response.  Same with Gladiator or Thin Red Line - the bass, subjectively, seems louder, deeper and tighter.  (the F/R plot confirms this).

A good friend - the one who owns the test equipment - tells me that my bass is boosted by 5 dB from about 200 Hz down.  FINE BY ME.  I listen for MY ENJOYMENT AND NOT TO SATISFY THE EXPECTATIONS OF OTHERS.   Also my bass goes somewhat deeper - down to about 25 Hz relatively flat.

I did notice that when we were doing the F/R tests that as we moved the microphone around the room the bass response varied wildly - in some places it was incredibly flat down to 24 Hz - in other places there were peaks and valleys. The F/R tests we kept were measured in my listening spot.

More on this later......

[valkyrie]

More responses to comments.

There seems to be some feeling that based on the few comments left by Tim Holl that the 2500 uF cap is not simply there for amp protection (though I contend that such is the sole purpose of it being in circuit) but that the cap/coil smooths and makes more linear the response of the Nine from 200 Hz down.  I do not agree with this - I think that cap was put there for only ONE REASON - to make driving two 4 ohm speakers in parallel POSSIBLE - and nothing else.

Even many of our modern amps could not handle such a load (at times < 2 ohms).

Stimpy; the point you raise is valid - but I contend that my bass is deeper, tighter and more articulate - and the F/R measurement seems to confirm this.  However - you may still be correct.  Perhaps in my room at my seat things are "better" - but such does not hold for all rooms and all seats.  Holl's comments on the addition of the cap/coil are at least a little obtuse - I have read them many times and am still not certain what exactly those parts do - other than raise the impedance of the speaker in the lower bass region.

Bottom line: do change the caps in the crossover to film - I am not much of a believer in high dollar film caps - I think that such are about on the level of power cords - but others disagree.  I do KNOW that moving the crossover to all film caps - with very tight tolerances - transformed my Nines from being rather cloudy sounding and ordinary to EXTRAORDINARY.

[Stimpy]

Mach3 - Is that you...???  

[DavidR]

27 minutes ago, Stimpy said:

Mach3 - Is that you...???  

I was going to post the same query. LOL

You beat me to it.

[DavidDru]

SHould I do the resistors while I am at it?

[Stimpy]

20 minutes ago, DavidDru said:

SHould I do the resistors while I am at it?

I'm thinking about it.  I'm at least going to change the resistor in my AR90 woofer circuit. 

As to the adjustment switches, I'm not sure, as I usually leave mine set to flat.  But, if you do change the toggle switch db resistors, you could consider changing the resistance values, to make for less heavy adjustments.  The switches are at -3 and -6 db now.  Maybe change it to -1.5 and -3 db.  Or whichever values you'd like to try?

[owlsplace]

1 hour ago, valkyrie said:

There seems to be some feeling that based on the few comments left by Tim Holl that the 2500 uF cap is not simply there for amp protection (though I contend that such is the sole purpose of it being in circuit) but that the cap/coil smooths and makes more linear the response of the Nine from 200 Hz down.  I do not agree with this - I think that cap was put there for only ONE REASON - to make driving two 4 ohm speakers in parallel POSSIBLE - and nothing else.

Even many of our modern amps could not handle such a load (at times < 2 ohms).

Would be nice if someone posts the schematic here for reference. I'm too lazy to go find it.

This discussion kind of reminds me of the LST schematic variations though.

Roger

[valkyrie]

I think I used to be "Mach 3" - hence the picture of the B-70 "Valkyrie" bomber.  That thing CRUISED at Mach 3 - not burst speed - sustained cruise speed.  Wonderful aircraft, especially when you consider that it was designed in the late fifties, early sixties using mostly slide rules.  Six huge engines - thing could reach 80k feet (confirmed) but if it used some toxic fuel it could go to 100k feet.  Since the fuel (some Boron salt I believe) would have caused poisoning of most of the planet ;-)) they decided to not employ that.

The outer portions of the delta wing would fold down at speeds > Mach 1 - thus trapping the trans sonic shock wave beneath the wing and creating lift.  Looks like a Klingon space craft with the wings folded.

How I lost my Mach 3 handle and became Valkyrie I am not certain - it was not intentional that much is for sure. I think it occurred when they changed the board software.   Whatever.

As for the adjustment switches?  I dispensed with those - which was a mistake.  My thinking was that the crummy switches were not good for sonic reasons - so out they went.  However I later found that I had to go into the crossover three times to "pad" down the UMR and tweeter units.  As I recall I ended up with about 4 ohms on the UMR and 2 on the tweeter.  It would have been easier with the switches in place - but those switches did not appear to be a sonic boom - looked more like a sonic degradation point than anything else.

If I had to do the crossover again?  I would have built it outboard in a separate box and then been able to tweak to my hearts content.

[valkyrie]

Here is the woofer portion of the AR-9 crossover.  The red line illustrates where I cut the 2500 uF cap and its coil loose from the circuit.  The next illustration is the entire crossover - be wary the value on the cap in the UMR should be 8 uF not 6 uF.  Also in my version (stock from the factory) there was an extra cap in shunt with the UMR driver. I think it was an 8 uF - I replaced it.

 

[valkyrie]

OK - after titanic struggles, bizarre conversions (from pdf to jpg) and rotating the image here is the F/R of my system, in my seat AFTER I CUT THE 2500 uF CAP LOOSE FROM THE CIRCUIT.   Needless to say - it doesn't appear that the 2500 uF cap does anything for the lower midrange (161 Hz to 320 Hz) or the upper bass (81 Hz to 160 Hz).  

Before there is another tempest in a teapot I divide the octaves starting at 20 Hz and doubling for each - so the bass goes three octaves from 20 Hz to 160 Hz.  The midrange goes from 161 Hz to 1280 Hz. Some people argue that the midrange should include another half octave up to around 2000 Hz.  I don't believe so but make up your own minds.

Please reference to the 1KHz level - this is something of an industry standard.  My bass from 90 Hz to 25 Hz increased substantially.   There is a peak around 130 Hz - but I attribute that to a room resonance (many rooms show a similar resonance - see John Atkinson's plots in Stereopile).  There is a slight dip around 200 Hz - about 3 dB and another at 80 Hz or so.  But I do NOT believe that these valleys had ANYTHING TO DO WITH THE REMOVED CAPACITOR.

Enjoy - and if you are fortunate enough to; 1) have a set of Mighty Nines and 2) to have an amp with some real ass to it, then cut that stupid capacitor LOOSE. ;-))).

 

[owlsplace]

The cap removal looks doable to me especially after seeing your response curve. The LST and 10pi schematics show those large caps. The first version LST crossover used four large caps amounting to a DC blocking circuit which apparently proved to be a bit challenging for contemporary amplifiers and was changed to two large caps with small cap and resistor shunts.

Onplane recommended passive bi-amping for the LST in an earlier post to eliminate the large caps and autotransformer with a tweek to the tweeter circuit to allow balancing the mid/high ratios.

Not sure why 20mh of inductance is required in the AR-9 woofer circuit when 2.85mh or less are found in AR cabinets using single 12" woofers.

Roger

 

[DavidR]

On 11/2/2016 at 0:55 PM, DavidDru said:

SHould I do the resistors while I am at it?

If you won't be using them I'd say no. It's a big expense and trying to match the values AND wattage would be pricey.

You could use the 12 watt Mills and parallel them to get 24 watts handling BUT you cut the resistance in HALF. If you want a 2 ohm 24W resistor you'd need (2) 4 ohm 12W resistors. Not a cheap way to go.

[DavidR]

On 11/1/2016 at 11:25 AM, Stimpy said:

That's what I thought too.  But, if it can be believed, there's been discussions on forums, to where owners reported hearing positive differences in sound, when changing shunt caps and resistors.  Jeff at Sonic Craft, offered an explanation as to why.  I'll have to l see if I can look up his reply concerning this.

Stimpy, you should also consider doing the 6 ohm resistor on the UMR

[Stimpy]

16 minutes ago, DavidR said:

Stimpy, you should also consider doing the 6 ohm resistor on the UMR

Thanks for pointing out my omission.  I had completely forgotten about the 6 ohm on the upper dome.    I didn't realize it until I looked at the schematic last night.  So, it's definitely going to be changed out to a Mills resistor, along with the 1 ohm resistor in the bass circuit.  After a recap, I'll address the switched resistors, and if I need to adjust those values, for tuning.  

[valkyrie]

There is NO RESISTOR IN THE BASS CIRCUIT OF THE AR-9 CROSSOVER.    Therefore do NOT replace it.

6 ohm on the tweeter? Perhaps you mean the 6 ohm on the UMR (big dome) crossover?   I did not replace that unit.  I did remove it and glue it back using RTV (silicon sealant) - so that it was vibration isolated.  As far as I know resistors do not go bad with time. 

 

[DavidR]

Mach3, Stimpy and I have AR90s. It is the UMR we are referring to with the 6 ohm resistor.

I have one on the dB switches that's gone funky. Its like reverb sound.

[Stimpy]

2 hours ago, valkyrie said:

There is NO RESISTOR IN THE BASS CIRCUIT OF THE AR-9 CROSSOVER.    Therefore do NOT replace it.

6 ohm on the tweeter? Perhaps you mean the 6 ohm on the UMR (big dome) crossover?   I did not replace that unit.  I did remove it and glue it back using RTV (silicon sealant) - so that it was vibration isolated.  As far as I know resistors do not go bad with time. 

 

You are definitely correct about the lack of a resistor, in the bass circuit, of the AR9.  It wasn't my intention to get DavidDru's original thread side-tracked.  It's just that the AR9 & AR90 are more similar than different, and discussions about recapping the 9 also relate, in a great extent, to the 90.  I'll try to limit my thread-within-a-thread mention of 90's crossover parts, to only those that relate to the 9s in future.  Sorry about the confusion.

[Stimpy]

1 hour ago, DavidR said:

I have one on the dB switches that's gone funky. Its like reverb sound.

That's odd!  I wonder if a resistor is cracked?  Maybe bass notes are making it vibrate and ring?  After all, early reverb units used simple springs.