AR3a Improved tweeter resistor

[dls123]

Hi all,

I am finally getting around to finishing this pair of AR3a Improved's. I downloaded the schematic from here (thanks), and have installed all new poly caps and new speaker terminals. Have also changed a few resistors. They sound pretty good, but they are just a bit hot in the high end. The AR3a Improveds don't have the tweeter level control, but rather a simple A/B DPDT switch. Tweeter is wired directly in position A and has a 1 ohm resistor in series in position B. Even with position B there is a bit of glare that I am pretty sure is due to the tweeters being just a bit hot. So...all you AR3a experts... should I remove the switch and wire it like the regular 3a using the L-pad scheme discussed in the big rebuilding the AR3a document? Or, it seems a simple solution is to either experiment with resistor values greater than the 1 ohm on the B position. Or, buy a simple 4 or 5 pole switch and install the several resistor values on each pole and experiment. Then I would leave the A position alone and when the DPDT was on position B the rotary switch would be in circuit and I could dial in one of a few resistor settings. Any reason not to go this route (a small hole and a good seal of that switch)? Seems to me that you'd then know that each speaker had the exact same resistance for the tweeter. If I go this route, anyone have experience as to what values to try? 1 ohm seems a bit too little. 2 ohms too high? I will have 3 or 4 positions to play with. Any advice appreciated! FYI they are being driven with a rebuilt Citation V (54 watts/ch) and a Citation I. I know they are probably a hard load for a tube amp, but they seem to have plenty of bass and mid-range is good, just a bit hot in the highs. A little "honky"...

cheers,

Don

[Zilch]

I will have 3 or 4 positions to play with. Any advice appreciated!

It's a switched variable resistance divider you need to replicate the original design, not merely different series resistances, which is easily constructed.

See my AR3a study here, which also analyzes RoyC's shunted L-pad attenuation approach:

http://www.audiokarma.org/forums/showthread.php?t=215006

And my earlier work with inexpensive stepped attenuation for AR4x here:

http://www.audiokarma.org/forums/showthrea...693#post1369693

Also on CSP here:

http://www.classicspeakerpages.net/IP.Boar...?showtopic=4681

[dls123]

Thanks Zilch. I read your threads, very interesting. However, the xover in the 3a Improved is quite different and I cannot quite understand how to implement your scheme. Attached is a schematic of my xover. I have replaced all the caps and a few resistors and verified that my speakers are identical to this crossover schematic. Any further help appreciated! I am not a speaker expert, but can read and build most anything from a schematic. Most of my experience is with tube amps/preamps although have rebuilt a few pairs of speakers. thanks!

It's a switched variable resistance divider you need to replicate the original design, not merely different series resistances, which is easily constructed.

See my AR3a study here, which also analyzes RoyC's shunted L-pad attenuation approach:

http://www.audiokarma.org/forums/showthread.php?t=215006

And my earlier work with inexpensive stepped attenuation for AR4x here:

http://www.audiokarma.org/forums/showthrea...693#post1369693

Also on CSP here:

http://www.classicspeakerpages.net/IP.Boar...?showtopic=4681

[Zilch]

Thanks Zilch. I read your threads, very interesting. However, the xover in the 3a Improved is quite different and I cannot quite understand how to implement your scheme. Attached is a schematic of my xover.

Try different values in 0.5-Ohm increments, perhaps. Parts Express stocks Dayton 10W Rs in those low values.

I am unfamiliar with that design, and you can see what's involved in determining what's going on in detail....

[dls123]

Hi again,

Have analyzed the AR3a improveds a bit. It is a hot upper midrange. Not sibilant, but it kind of honks at you in the upper vocal range, particularly female vocals that tend towards brightness. I looked and one of the midranges has a different colored lead wire that someone spliced into with the lead from the xover. So it appears a midrange has been swapped or repaired at some point. It appears identical to the other one though except for the color of that lead. I recall measuring the resistance and the two were the same, but I should do it again. So I wonder if the characteristics of that midrange driver are different even though it appears identical. If you run the tweeter switch on "B" for max attenuation then you lose a lot of the cymbals and shimmer, although it does cut down on that upper midrange honk a bit. Running it on A has decent highs, but then it honks too much on vocals - just edgy. I tried them upstairs on the Cit. II and they do it there too. Interesting. Guess I will have to measure the midrange resistance and look carefully at any markings on the rear of the drivers and see if they are different. They appear identical on the front. I know there is some info on this site about the different midrange drivers. More later...stay tuned.

cheers,

Don

Try different values in 0.5-Ohm increments, perhaps. Parts Express stocks Dayton 10W Rs in those low values.

I am unfamiliar with that design, and you can see what's involved in determining what's going on in detail....

[Zilch]

Link us to the thread describing this design, please.

[dls123]

Hi

I can't seem to find the thread, but I posted the xover schematic that I pulled from this site above in my earlier post. My speaker is exactly like that schematic. I have put solens caps in, but used the same values. Resistors are all within 5%.

Don

Link us to the thread describing this design, please.

[Zilch]

I can't seem to find the thread....

That's the problem I'm having; I have no clue what and "Improved" AR3a is....

[genek]

That's the problem I'm having; I have no clue what and "Improved" AR3a is....

From what we have been able to discern by looking at photos of different "AR-3a Improved" versions, the early pre-1976 version mostly just modified the crossover to to eliminate the level control pots (though I think we've identified at least one set of "Improveds" that still have the pots) and perhaps bump the highs up a bit. Later, the tweeters were replaced with the ferrofluid-cooled units from the 10pi and 11, bumping the highs up quite a bit more and essentially turning the model into an AR-11 in AR-3a clothing. Whether this really represented an effort to "improve" the AR-3a or just consolidated inventory and enabled AR to sell speakers that used the same components as the ADDs but looked like classics because that was what their intended market preferred is just guessing. Most of the post-classic era 3a variants ("Improved" and "Limited") we've seen seem to have originally been built for sale overseas.

[Robert_S]

Hello,

Link us to the thread describing this design, please.

The topic/post dis123 had in mind is perhaps the following:

http://www.classicspeakerpages.net/IP.Boar...ost&p=73706

From what we have been able to discern by looking at photos of different "AR-3a Improved" versions, the early pre-1976 version mostly just modified the crossover to to eliminate the level control pots (though I think we've identified at least one set of "Improveds" that still have the pots) and perhaps bump the highs up a bit. Later, the tweeters were replaced with the ferrofluid-cooled units from the 10pi and 11, bumping the highs up quite a bit more and essentially turning the model into an AR-11 in AR-3a clothing.

My understanding is that, to date, we have only seen one version of the AR-3a Improved. The original factory-installed tweeter is the same type as found in the AR-3a (allowing for front- or back-wired design variations). The AR-3a Improved crossover remained essentially the same throughout production. The only part variation we've seen has been the wadding type.

Please correct me if I am mistaken?

Robert_S

[soundminded]

Hi all,

I am finally getting around to finishing this pair of AR3a Improved's. I downloaded the schematic from here (thanks), and have installed all new poly caps and new speaker terminals. Have also changed a few resistors. They sound pretty good, but they are just a bit hot in the high end. The AR3a Improveds don't have the tweeter level control, but rather a simple A/B DPDT switch. Tweeter is wired directly in position A and has a 1 ohm resistor in series in position B. Even with position B there is a bit of glare that I am pretty sure is due to the tweeters being just a bit hot. So...all you AR3a experts... should I remove the switch and wire it like the regular 3a using the L-pad scheme discussed in the big rebuilding the AR3a document? Or, it seems a simple solution is to either experiment with resistor values greater than the 1 ohm on the B position. Or, buy a simple 4 or 5 pole switch and install the several resistor values on each pole and experiment. Then I would leave the A position alone and when the DPDT was on position B the rotary switch would be in circuit and I could dial in one of a few resistor settings. Any reason not to go this route (a small hole and a good seal of that switch)? Seems to me that you'd then know that each speaker had the exact same resistance for the tweeter. If I go this route, anyone have experience as to what values to try? 1 ohm seems a bit too little. 2 ohms too high? I will have 3 or 4 positions to play with. Any advice appreciated! FYI they are being driven with a rebuilt Citation V (54 watts/ch) and a Citation I. I know they are probably a hard load for a tube amp, but they seem to have plenty of bass and mid-range is good, just a bit hot in the highs. A little "honky"...

cheers,

Don

I'd go with the L pads. That's what makes by far the most sense to me. It isn't authentic. It's not the way the factory did it. But it's the way I would have done it, many others did it, and I'd do it now. If you are concerned about an exact match, they come with plates and I'd check the impedence with an ohmeter at a few points on each one and mark the plates to be sure the attenuation settings exactly correspond. But personally, I'd trust my ears. Most likely they will all track pretty much the same.

[genek]

My understanding is that, to date, we have only seen one version of the AR-3a Improved. The original factory-installed tweeter is the same type as found in the AR-3a (allowing for front- or back-wired design variations). The AR-3a Improved crossover remained essentially the same throughout production. The only part variation we've seen has been the wadding type.

Please correct me if I am mistaken?

We've seen several "versions." The problem is, we don't know for sure whether the ones we've seen all had their original tweeters. But at least one of the versions we've seen had the same 0.10mH tweeter inductor in its crossover that the 10pi and 11 have, which can't help but make you wonder.

[soundminded]

We've seen several "versions." The problem is, we don't know for sure whether the ones we've seen all had their original tweeters. But at least one of the versions we've seen had the same 0.10mH tweeter inductor in its crossover that the 10pi and 11 have, which can't help but make you wonder.

I'll guess.......that they added the inductor because too many of them were coming back with burned up tweeters. The inductor turns it from a first order to a second order filter increasing the slope from 6db per octave to 12. This prevents upper midrange energy from overdriving the tweeter. It also probably hardly went noticed soundwise.

[genek]

I'll guess.......that they added the inductor because too many of them were coming back with burned up tweeters. The inductor turns it from a first order to a second order filter increasing the slope from 6db per octave to 12. This prevents upper midrange energy from overdriving the tweeter. It also probably hardly went noticed soundwise.

Could be. And could also be that if they came back with burned up tweeters after AR ran out of 3a drivers they were shipped back from the factory with the ferrofluid 10pi/11 tweeters installed...would that make them another official "version...?"

[dls123]

Could be. And could also be that if they came back with burned up tweeters after AR ran out of 3a drivers they were shipped back from the factory with the ferrofluid 10pi/11 tweeters installed...would that make them another official "version...?"

Hi

OK, getting back to original topic. I am not sure how many official AR3a improved versions there were, but looking through the excellent rebuilding document on the home page, my tweeters do match one of them in there. I can't remember which, but it was one of the later ones. At any rate, the xover schematic I posted is exactly what is in my speakers and so is the wadding. So, the problem is that they just have too much energy in the upper mids and perhaps the tweeter. If you use switch position B it tames them somewhat, but then the highs lose some of the shimmer. I really think it is more of an upper mid-range glare I am hearing. I am sure these drivers are capable of better. So we have one vote for modifying it to the L-pad xover. Any other suggestions? I have two Citation II's on the bench now and a Sherwood S-8000 that I am finishing up, so it may be a few weeks until I get back to the AR's. The other amps are for other folks who actually pay me a bit to work on the so they sort of have higher priority! Still, I wouldn't mind ordering some parts if I can agree on a design. Any suggestions will be considered! Thanks all for an excellent forum!

Don

[Zilch]

Any suggestions will be considered!

The best way to figure it out is with response measurements, but that's not an option.

The switch alters both the mids and highs simultaneously, and I can't figure that out without putting it in Spice. I'll need to know the mid and tweeter impedances, nominal, at least. Do we assume 4 Ohms for the tweeter?

Do the two speakers behave differently? If not, why are you questioning one of the mids?

If they behave the same, Solens would be a likely suspect; many find them too bright, sometimes "harsh" in vintage recaps.

[dls123]

The best way to figure it out is with response measurements, but that's not an option.

The switch alters both the mids and highs simultaneously, and I can't figure that out without putting it in Spice. I'll need to know the mid and tweeter impedances, nominal, at least. Do we assume 4 Ohms for the tweeter?

Do the two speakers behave differently? If not, why are you questioning one of the mids?

If they behave the same, Solens would be a likely suspect; many find them too bright, sometimes "harsh" in vintage recaps.

Hi,

You are right in that I need to hook them up again, put the preamp in mono mode and compare channels to see if one speaker is the culprit. Assuming they are both the same it could be caps, but I have used Solens before in my Advents and they sounded way better. Albeit, a completely different design. I could grab some Dayton caps to try, a cheap experiment. Anyway, the Solens have about 25 - 30 hours on them so they should have settled down. Impedances..I will have to look at the drivers. Will experiment a bit shortly if I get the chance. Thanks.

[KlausDK]

I have a set of AR-3a improved, and taking all aspects into consideration, they sound, as the name says: Improved. This is probably partly due to the lack of the rheostats; the midrange and the highend is even clearer and more open, with a rock steady soundstage, amazing. They are a bit to the bright side, but I have bypassed the swithch and while upgrading the cross-over, I fixed it in the "high/A" position, ie bypassing the switch and without resistor in the signal path to the tweeter. So there is no posibility to regulate sound.

I have used a 35 year old Capacitor for the tweeter, from the highly regarded Bosch MP series (Metalized paper/oil), I happened to have a 6 myF capacitor, I have bypassed it with a 100 nF polystyren cap. After having restored several AR-3a's, LST, 5's etc, I have come to the conclusion, that the normal metallized paper polypropylene capacitors (like eg the Solens), don't work as well with the paper 3a/LST/5/2ax dome tweeter. Film and foil (polypropylen) like the Zen Crescendo cap or Paper Oil (like Jensen or Bosch MP) works better with that specific tweeter, in my opinion.

I have had "miss-sounding" midranges due to dried out magnetic oil (10Pi's and 11's), but I am not sure if there is oil in the 3a Improved midrange, there is not in the tweeter, assuming all drivers are original. AR-3a improved was originally mounted with AR 3a drivers with a new cross-over, there has been lots of rumors about 3a-Improved with ferro fluid 10Pi/11 drivers ex factory, but as far as I know this was not the case, and nobody has been able to produce pictures or documentation to back up that. They came with both front and back-wired tweeter and midranges.

http://www.classicspeakerpages.net/IP.Boar...c=4396&st=0

BRgds Klaus

[Robert_S]

We've seen several "versions." The problem is, we don't know for sure whether the ones we've seen all had their original tweeters. But at least one of the versions we've seen had the same 0.10mH tweeter inductor in its crossover that the 10pi and 11 have, which can't help but make you wonder.

Please show some evidence of an AR-3a Improved crossover that is without a coil across the tweeter!? I think we've questioned this already, and so far, no crossover design variation throughout the production period has come to light.

Robert_S

[genek]

Please show some evidence of an AR-3a Improved crossover that is without a coil across the tweeter!? I think we've questioned this already, and so far, no crossover design variation throughout the production period has come to light.

IIRC, so far we've seen the innards of 2 or 3 "Improveds." They've all had the coil, but they haven't all had the same tweeters and none of them has had an unbroken chain of ownership that would ensure they are original and unmodded. With a sample that small and uncertain and no factory info, pretty much everything we think we "know" about the model is guesswork.

[dls123]

I have a set of AR-3a improved, and taking all aspects into consideration, they sound, as the name says: Improved. This is probably partly due to the lack of the rheostats; the midrange and the highend is even clearer and more open, with a rock steady soundstage, amazing. They are a bit to the bright side, but I have bypassed the swithch and while upgrading the cross-over, I fixed it in the "high/A" position, ie bypassing the switch and without resistor in the signal path to the tweeter. So there is no posibility to regulate sound.

I have used a 35 year old Capacitor for the tweeter, from the highly regarded Bosch MP series (Metalized paper/oil), I happened to have a 6 myF capacitor, I have bypassed it with a 100 nF polystyren cap. After having restored several AR-3a's, LST, 5's etc, I have come to the conclusion, that the normal metallized paper polypropylene capacitors (like eg the Solens), don't work as well with the paper 3a/LST/5/2ax dome tweeter. Film and foil (polypropylen) like the Zen Crescendo cap or Paper Oil (like Jensen or Bosch MP) works better with that specific tweeter, in my opinion.

I have had "miss-sounding" midranges due to dried out magnetic oil (10Pi's and 11's), but I am not sure if there is oil in the 3a Improved midrange, there is not in the tweeter, assuming all drivers are original. AR-3a improved was originally mounted with AR 3a drivers with a new cross-over, there has been lots of rumors about 3a-Improved with ferro fluid 10Pi/11 drivers ex factory, but as far as I know this was not the case, and nobody has been able to produce pictures or documentation to back up that. They came with both front and back-wired tweeter and midranges.

http://www.classicspeakerpages.net/IP.Boar...c=4396&st=0

BRgds Klaus

Thanks Klaus. I rebuild tube amplifiers and have a number of caps laying around that would be good for bypass. I have a whole drawer of K40 PIO's. I do recall that I looked at the drivers and compared them to the info in the rebuilding the AR3a document when I picked these up a year ago. I am pretty sure I could identify both the tweeters and the mids as late model ones in that document. One tweeter measured just a bit low in resistance compared to the other and both mids were spot on. However both tweeters are working and the speaker does image well. I will try some warm sounding caps in both the mid and tweeter positions and maybe even bypass with a small K40 as an experiment and see if that tones things down a bit. A very cheap experiment. You can hear their potential. I don't know how much a cap change will tame them, but worth a try. A nice vintage speaker. They can't compete with my big Josephs upstairs, but that is not a fair comparison. I think my Citation V will push them very nicely if I can tame that upper midrange. Will identify all drivers, test sound of one speaker vs. the other and report back soon.

cheers,

Don

[dls123]

IIRC, so far we've seen the innards of 2 or 3 "Improveds." They've all had the coil, but they haven't all had the same tweeters and none of them has had an unbroken chain of ownership that would ensure they are original and unmodded. With a sample that small and uncertain and no factory info, pretty much everything we think we "know" about the model is guesswork.

Hi again,

this might be a new topic...but has anyone tried the big Russian PIO caps in the xover for the 3a? Here is an interesting cap:

http://cgi.ebay.ca/6-uF-400-V-Paper-Capaci...93%3A2|294%3A30

Think these would be too slow an syrupy? Anyone with any experience? You could also make it work with K-73's or use a combo of the two. I have a lot of experience with Russian PIO's in amps, but none in speakers. They are so cheap you can buy 10 and match them with a meter, but perhaps would sound terrible in speakers. Anyone with experience please chime in.

thanks

don

[Zilch]

Voltage drives:

A = top pair

B = bottom pair

[Kuja]

...I have had "miss-sounding" midranges due to dried out magnetic oil (10Pi's and 11's)...

Klaus,

Can you describe the sound of drivers with dried out ferrofluid?

My AR-90s sometime have some coldish overtone.

I don't know hoe to describe this - music can sound "glassy", especially higher pitched voices.

I don't know if this is due to poor recording, or it has to do with my speakers.

They sat unused in some basement for 10 years or more before I restored them.

I did not tell my "speaker guy" (Sinisa from Sinevoice - you know him) to open midranges and tweeters to check them.

Now I wonder if I should have done that....

How common is ferrofluid drying out?

Big thanks,

Aleksandar

.

[KlausDK]

Can you describe the sound of drivers with dried out ferrofluid?

Hi Aleksandar

The experience I have with dried out midranges comes from the magnetic particle sediments in the magnetic gap, making "mis-sounds" and distorsion of the sound, if you make a frequency sweep, you can find the midrange resonating at different frequencies within the spectre of the driver.

I don't know how common it is, but I have experienced two with the above problems, and have asked Sinisa/Sinevoice, and he can clean and refill the midrange drivers with magnetic oil.

BRgds Klaus