Aadams

Edward Does the black wire running from your tweeter pot have a 3.3 cap spliced near the end? if not remove a tweeter and look on the backside negative terminal or the connecting wire. If no cap found it probably wasn’t installed.

"3uf cap in series with the original 6uf cap" "So it appears that the tweets were replaced with an upgraded model" I wouldn't say upgraded just more robust. Edward Don't forget to ask about the tweeter lpad resistor if you are keeping these tweeters.

Those are nice really late model 3as. If both of your tweeters are identical they do not need rebuilding, if they are working. They appear to be the robust tweeter used in the early AR11 which means you may need a resistor on the tweeter lpad. It also may explain why the crossover appears a little different. I think you are good to go except for definite advice on the tweeter resistor which should come from RoyC in the next day or so. Adams

To be clear are you referring to cones or the foam surrounds. Can you post an image of the driver in its current state? Adams

Edward Those look like 3a crossovers but different from what I have seen. Can you please supply a picture of the front of the speaker cabinet showing driver faces. Also what are the serial numbers on the cabinets? Do you know the dates on the original purchase receipts? Does the receipt say which speaker was purchased? 3 or 3a? Pardon the questions, this isn't idle curiosity. There are others here more expert and supplying the requested info can get you sorted out quicker. Also the Parts Express resistor is a low inductance 1%. They no longer stock the 5% version. Also, even if they are working, your tweeters are severely degraded in performance compared to when you purchased them. Now may be the time to consider having them rebuilt in which case you will need a resistor for the tweeter lpad. Photos will help. Adams

Edward Your topic title is AR3a but you mention AR3 woofers. If your woofer was originally foam then foam would be the replacement to maintain correct performance. Your parts list for new NPE caps is good but one recommendation is to use three 50mf instead of a single 150 for the woofer. I think the reason is the smaller value caps in that line have lower variance from spec. I was a pot holdout but I am now an LPad convert. You will be better served by replacing the pots with lpads and resistors. The 25ohm 10 watt resistor is mandatory on the mid range lpad but only recommended for the tweeter if you have the tweeter rebuilt or replaced with a HiVi QR mod. It would help if you post pictures of fronts and backs showing drivers and serial nos as well as interior images of the crossovers so those with more knowledge might have extra helpful tidbits in case you have a model with transitional variations in manufacture. For example I can't remember seeing a 3a with a mix of Spragues and Royalitics. Adams

Wow! That is a big one. 8 ohm 15 watt lpads are 1.5 inch diameter and a snug fit side by side in a 3a box. Threaded stem diameter is near identical to pots. Some members here have used what you are showing as 3a pot replacement so maybe there is a difference in the mounting hole displacement between a 103 and a 3a or they made mods to the masonite holes. If you use Lpads with 1" inch stems you can make use of "air rights" inside the box and have more room for wiring and other pieces to run in the horizontal plan. Setting stem height is easy with supplied spacer washers and nuts.

Mondo I thought someone with 103s would have answered by now. If these pots are larger than lpads then they are large. Are you saying both stem and cylinder are too large? Would not drill out holes. I would find decent used pots before I did that. Lpads are least expensive and most reliable route if they will fit. Adams

The site reffed by ARPro sells the exact Dayton resistor for 2.25. Just click the link and. Scroll down the page.

I made the same mistake but had I known what I was doing at the time I would have found this one. They no longer stock the 5% but there is a 1% for just a bit more. https://www.parts-express.com/dayton-audio-dnr-25-25-ohm-10w-precision-audio-grade-resistor--004-25?gclid=EAIaIQobChMI_-ui6rbT5gIVTfDACh0EGgfoEAQYAiABEgLDxPD_BwE

Even if your pots seem to be exceptional after 40 years LPads with resistors are better path. Look at the March 25 posts in this thread. Because you are working completely outside of the cabinet you can easily copy the tidy version by RoyC.

I have been comparing the AR9 to the AR3a for the last couple of days. In this 4 way configuration it is easy to dial in a 3a to achieve the same spectral balance as an AR9 that is set to 0db attenuation. As I listen I can find differences but one thing that is strikingly missing from the differences list is the fabled AR3a rolled off sound. If you look at the chart below, which is based on a published AR chart supplied by Tom Tyson @tysontom, you can see the sloping integrated response curve of the 3a running below a superimposed AR9 response curve that is much flatter. The chart seems to say the AR9 can easily produce high mid-range and tweeter sounds in areas where the 3a struggles. The practical reality is this is not true. For the AR3a to match the spectral balance of my AR9s the mid-range attenuator must be turned down near halfway with the tweeter at max. The result will be same for any properly working 3a anytime the listening space includes walls, floors and ceilings because there is so much reflected energy. I came to the AR world late and accepted that 3as had a “rolled off” sound but if they are refurbished with Lpads and a rebuilt tweeter or Hivi, they only sound “rolled off” if you make them so. I am guessing by the time digital recording arrived the classic tweeters were already degraded while appearing to work and the legend of the “rolled off” AR sound began. The measured on axis response differences below set expectations for what you would hear in open space with no reflections, but what the you hear in normal listening space is surprisingly different and not rolled off at all.

Do you have PCB or Masonite board. The mod is resistors and tweeter polarity. No caps

Der Here is a link that should help http://www.classicspeakerpages.net/IP.Board/index.php?/topic/9817-hivi-q1r-install/

This is my preference. I like the feeling that I am there. Regarding AR9s They can image well but it sounds to me like "spooky imaging" involves a ratio of direct and reflected sound. The AR9 series verticals don't do this until well back in the listening area. Earlier ARs can be made to do it in the near field and the LSI models can do it out of the box. Of course you still have to find the spot.

This thread may be destined for the Other Speakers Forum but right now I have one last question about "spooky imaging". Is "spooky imaging" a near field or far field phenomenon? Adams

How would you characterize "spooky imaging"? Is it in the category of "you are there" or "they are here"?

About all that can be said for certain about the speakers mentioned, is they are all capable of 10 octave performance and they are towers. The only meaningful way to compare their performance is by ear in the same room in the same position with the same source. A properly operating AR9 or 90 can image just fine in the near field if you are willing to compromise flat bass performance by toeing them in away from walls. If Paradigm is truly comparable there will little difference between it and the the 9s. The big variable here is subjective taste and the human ear. Adams

You are ok. The white glue is way more forgiving. Longer set time allows for fine tuning the fit.

It will look better than 18 gauge but Cat5 is solid copper single strand. Multiple strands might be easier to work with and better for the application. OTOH the originals were single copper strands without a plastic sheath so go for it. Adams

Ligs Hello, For the last few years, I have wondered why the LST is always touted has having a better quality sound than an AR3a even though it uses identical drivers and crossover points. The LST should be able to make more sound in bigger spaces than an optimally setup 3a but why would the same recording have a better sound quality? I believe it is because the 4 LST mid range units are fixed at identical output and cannot be varied. A stereo pair of LSTs, operating at factory tolerances, should be a matched pair in the mid range, always, without regard to how the spectral balance control is set. I would further generalize and say that all AR dome midranges from the 3a through the 58s should sound identical when stripped of their faceplates, which means the Classic dome sound should sound virtually the same on axis, in the near field, to anything that followed that used the same family of mid domes. I must give @frankmarsi credit for dropping the LST clue over 2 years ago, in this thread, but at the time I had not traveled enough on this experimentation path to use the information. I must also credit you, because I would not have bothered with lpads and mid range balancing had you not questioned my preference for the AR5 top over the 58s top, some time back. Adams

The AR classics with potentiometer or lpad attenuators, as opposed to later series with 0db attenuation settings, are difficult to balance. If you want to hear a classic at it’s absolute best the attenuators must be adjusted to bring the corresponding drivers of the stereo pair into exact loudness balance. In the case of the 3a and 5 the mid-range attenuator is the predominate factor in determining whether the sound is focused and centered as opposed to sounding vaguely centered and with a shifting soundstage. I have found an easier way to balance attenuators that employs a 10 band equalizer. For a 3a or 5 set the sliders to full minimum on every band except the 2k band which should be set to maximum. This provides max emphasis at 2k which is in the middle of the mid range driver bandwidth. 1. Assuming you already have the attenuators at a loudness setting that pleases you: 2. Play a mono recording with vocals, I like Glenn Miller or early Beatles for this, and listen within 3-4 feet to judge if the image is centered between the speakers. Toe them in if you can. Adjust the attenuators until the image is centered in the near field or at the listening position or both. 3. Return the equalizer to its normal setting and reposition the speakers. You are done. Adams

If you are switching drivers it suggests you can solder and use a screw driver at least. You are halfway there. Also, you might solve a lot of your problem with a ten band graphic equalizer. Check your PMs

If the QR1 mod was done correctly it will not scream in the mid range. You should be able to turn it down to a setting that pleases you. Does the QR1 have the recommended inductor installed on the back of the unit? If not it will scream in the midrange. If the attenuators are bypassed then even with a correct QR1 I can see how one might say the mid range is too bright especially if they were bypassed incorrectly. The best solution for pots is lpads, as described in the restoration document, but if your pots are working at all then you can probably fix your tweeter problems now and think about lpads later. You need to talk to Roy about QR1 vs rebuild. Do not toss any of your broken tweeters. Broken tweeters can be repaired nowadays if the bodies are not deformed. PS I can guarantee none of the your original AR tweeters are the source of a midrange problem. In fact, none of them are working anywhere close to their original output specification. Adams

I would like to see a picture of that "screamy" tweeter. I can't imagine how an AR tweeter could be "screamy" at 5khz, which is the bottom of their range. Rebuilding is an option that has been available for about 2 years now and is preferred by some of us. If cost is the prime driver the Hivi QR1 has a proven record when installed correctly. If you are a DIY type you should download the restoration document. If not DIY you could send a private message to RoyC who is an authority on both options. Adams