My Excellent RS1 Adventure

As anybody who has gotten to this point in the thread will know, I learned a lot about the RS1b during this project. I also had some close knowledge of Infinity starting about fifty years ago and, thus, know what their design goals were. I have seen descriptions of the RS1 design as being either a "quasi line source" or a "pseudo-point source." I think it is something else, and here is what I think it is, and why...

Arnie Nudell started playing with line source techniques at the very beginning of his career. While developing the Servo Statik 1, he worked with RTR to develop a thin, long “ribbon electrostatic” tweeter to improve the dispersion of the high frequencies over what was available from other contemporary electrostatic panels, which were square. I can’t find any photos now, but the original Servo Statik had an electrostatic tweeter that was about an inch and one-half wide, and 30 inches high. This design, and subsequent work with line arrays, was based on page 36 of the 1957 Second Edition of Harry F. Olson’s “Elements of Acoustic Engineering,” where you will see proof that a driver like the one in the original Servo Statik will have very good horizontal dispersion (because of the small width compared to the wavelength of high frequencies), but a lobe-shaped dispersion for vertical dispersion (due to the line array behavior of the 30? tall tweeter).

At the beginning of the 1970s, the best speaker designers, were contemplating the difference between a loudspeaker and a symphony orchestra. I know that sounds glib and superficial, but the basic problem is that an orchestra (and almost everything else in the real world) has constant power output (power into a hemisphere) with frequency, but the power output of a loudspeaker falls off rapidly as frequency increases, because of baffle and driver size limitations (baffles too small, and drivers too big). Bose “solved” the problem by using an active device to boost high frequencies and then, spraying them all around the room by aiming the drivers at the wall. Infinity started experimenting with line arrays as a more elegant way to control power output by frequency.

View attachment 1269049
Referring again to page 36 of Olson, you will see that by the time the length of the array is equal to the wavelength of the frequency, you achieve a nicely directional radiation pattern with little radiation beyond +/- 60 deg of horizontal. Thus, with a 54-inch line (the approximate length in the RS1b EMIM array), you can get a good directional pattern starting at about 250Hz. But with a 54-inch line, you get severe beaming by 2,500Hz (line size 6x the wavelength), and it only gets worse the higher you go. This kind of beaming can be observed by moving your head above and below the midpoint of the array, where you will notice a fall off of the higher frequencies. The solution to this problem is the secret to the design of the RS1b; it is a five-way speaker, utilizing successively smaller line arrays as the frequency increases. The first band is handled by the active crossover and the woofers from about 30Hz to 150Hz. The radiation pattern is essentially omnidirectional. The next band is from 150Hz to 700Hz and is handled by the 54? EMIM array (#s 1, 2, 4, 5, 6, 7). Above that is the upper midrange (700Hz to 3000Hz) array handled by a single 7-inch EMIM driver (#3), then the lower EMITs from 4000Hz to 8000Hz (using “1/2 line source” standard EMITS of about 2 inches, and finally, the upper high frequencies from 8000Hz to about 32kHz, using a “1/4 line source” EMIT (a standard EMIT with one-half of the radiating area taped over, having a line length of about 1 inch). The fifth part of the crossover is the rear tweeter. This was required, as power output for a speaker in a room, is defined as the radiation into a hemisphere. The EMIMs were dipoles and provided plenty of rear radiation. The EMITS only radiated in one direction, thus the need for a rear-facing EMIT.

And that is most of the story. I think the RS1b speaker was one of the most elegant of the classic Infinity speakers, as it used solid engineering to develop a speaker based on solid theory. Perhaps, it is an intellectual speaker. It certainly sounds better than the Bose 901. And while it may not have mid-bass oomph, and the effortless dynamic range of other top high end speakers, it does solve the problem of constant power output very well. Although it doesn’t make much difference, due to the incredible number of drivers in the fabled IRS Vs, if you move your head up and down from the middle of the array, you can hear high frequency beaming, which you will not hear from the RS1bs.

I remember reading an early review of the RS1bs where the writer declared that the tape on the upper tweeter was to tame the highs by reducing output. That, of course, is ridiculous. That wasn’t the point of the tape at all. By shortening the length of the array, the power output at the high frequencies was actually increased. I wonder if Arnie ever sent a note to the reviewer to correct the mistake, or if he just let it go.

Could you elaborate (in as much depth as you wish) on the last paragraph? I just don't 'get it' the taped over portion of the Emit is still driven, right? where does that energy go?
Thanks. I find this fascinating
 
Could you elaborate (in as much depth as you wish) on the last paragraph? I just don't 'get it' the taped over portion of the Emit is still driven, right? where does that energy go?
The tape acts like an acoustical lens, making the radiating area shorter. The energy remains the substantially the same, as there isn't much of anything absorbing energy in this arrangement.
 
Last edited:
I know this thread is a couple of weeks old, but I'm still catching up after a week's vacation.

The idea of tape making the radiating area shorter had me thinking about the Kappa 9. This uses what they call a SEMIT (basically a half-length EMIT, same as the EMIT used in some versions of the RS3000, but with a different faceplate). Even though the Kappas aren't line arrays, could that have been for the same reason? It certainly isn't for high-frequency extension.
 
I’m sure it’s the same reason. An EMIT is a line radiator. If the frequency is high enough, it will start to beam. The SEMIT, having a shorter length, will beam less at the highest frequencies.
 
A smaller aperture improves dispersion. So, by taping off Emit openings, or by using a Semit, dispersion improves at higher frequencies. Which again, lessons beaming at high frequencies. But, playing with toe in does pretty much the same thing. Slightly more toe in aims the tweeters at the ears, and improves the high frequency content. Tape or toe in, both easy enough to experiment with.
 
But they aren’t the same. Shortening the radiating length of an EMIT increases vertical dispersion but leaves horizontal dispersion the same. Toe in re-aims existing dispersion lobes, and on a RS1 or RS2 can increase the illusion of a wide soundstage by pointing the rear radiation outward.
 
It's time for the next chapter: the Reference Standard Servo Crossover.
[...]
This fixed the buzz in the midrange/tweeter towers, but there is still some minor hum in the woofers (there must’ve been some reason that Infinity added a hum balance adjustment pot to later versions of the crossover). I have decided to live with it, and I'll now get down to some serious listening.

As it turned out, the hum wasn't minor, or it was major enough that I couldn't deal with it, so I bought a miniDSP HD 4x10 digital signal processor to use as a crossover. I got a model with enough channels that I could eventually get rid of most of the passive crossovers in the speakers and quad-amp the speakers (woofer towers / lower EMIMs / upper EMIM / tweeters).

I've been listening to it for months, the hum is completely gone, and it sounds great to me. Here are some screen shots of the miniDSP control app set up in my system:

Here is the main screen showing the eight analog input channels
Capture-1.PNG

Here is the main screen showing the eight output channels
Capture-3.PNG

Here is the low-pass crossover...
Capture-4.PNG

And here is the high-pass crossover...
Capture-5.PNG

And, here is the parametric filter on the left input (it's the same on the right input). Since the RS1 woofer towers are in a closed box with a resonant frequency of about 54 Hz, bass response will fall off at 12db/octave below that point unless you do something. The standard Infinity Server/Crossover box uses servo feedback to extend the bass below 54 Hz. I'm using a special filter with a gain of 16 db at 23 Hz which makes the bass flat to about 23 Hz.
Capture-2.PNG

As people are always complaining about the integration of the RS1/a/b woofer towers with the mid/high towers, the next project will be adding a mid-bass coupler between the woofers and the mid/high towers. This will allow the woofers to cross over lower, and the mids to cross over higher. I have a pair of KEF B200 woofers in classic Bailey Wireless World transmission line cabinets that I am going to use for the experiment. I'm going to try crossing over at 80Hz and 220Hz. Obviously KEF drivers aren't vintage Infinity approved but I know for a fact that Infinity experimented with KEF B139 woofers, but never made a product out of them, so I'm telling myself that this is an experiment that might have happened. (Wilson Audio put KEF B139 woofers in the original WAMM, showing that KEF has a good pedigree).
 
Last edited:
Great write-up Paul. Thanks for sharing the info and setup. I’ve managed to get good sound out of my controller with much effort. I’m on my third unit which has hum balance. I have replaced the electrolytics and socketed opamps. The dsp is definitely on my roadmap though. Thanks again for the info. I look forward to results using the dsp to replace the passive components of the tower too.
 
A quick upgrade that was absolutely worth the $20 and one-half hour it required to accomplish. The stock banana plug terminals are cheap and loose with many plugs. These replacement terminals from Parts Express are a big upgrade. (Though maybe not as much as “tube connectors” ha ha ha).

5AE85EC2-E9AB-4661-9334-CF21C1E08126.jpeg

6755282F-7FE9-454E-88C6-5200CA9F7958.jpeg
Original on left, replacement on right.
 
Last edited:
Those are the same ones I used on the 4.5’s, huge difference in the connection stability ! It’s nice to have a good solid connection with modern cables.

I agree great write up Paul, I’ve read it over a few times now. So the DSP has now replaced the passive crossover’s if I’m reading this correctly ?

It will probably be around May for me to secure the gear needed to bring my 4.5’s back to life and the Mini DSP interest me to the point I think I’m going to give it a try with these guys. So do you physically remove the original crossover.
 
Last edited:
Those are the same ones I used on the 4.5’s, huge difference in the connection stability ! It’s nice to have a good solid connection with modern cables.
I agree great write up Paul, I’ve read it over a few times now. So the DSP has now replaced the passive crossover’s if I’m reading this correctly ?
[...] So do you physically remove the original crossover.
Those are the same ones you used, because I got the link from a post you made here. Thanks for the pointer! It's a big improvement.

So far, I've only replaced the active crossover with the DSP (between the woofer towers and the mid/tweeter towers). Plus, at the moment, my system is tri-amped, as I've added a mid-bass coupler from 80 to 220 Hz so the bass towers don't go as high and the mid/tweeter towers don't go as low.

However, I intend on removing the passive crossovers for the lower and upper EMIMs and driving them directly with amps being fed from the miniDSP. I don't have enough DSP channels or amplifiers lying around to sext-amp the speakers and get rid of all of the passive crossovers, but I would certainly like to do that.
 
That’s right huh, I forgot about that. Needing to have an amp per bank of speakers ? So we would need 6 amps, one for each woofer set, one for each Emim set and one for the emits. Or would it be like a 6 channel amp for each speaker set with the same power output. I wonder how that would work with the Watkins Woofers. How would you figure out adequate power for each channel going separates.

I don’t think the 3 emits would need the same power as 8 Emims so on and so forth.

I’ve got some reading and rethinking to do now, hope you don’t mind me posting this Paul, not trying to hijack your thread just thinking out loud. And while I’m doing that I’ll be in the Bay Area next month :).
 
For the Watkins woofers, you're still using the internal bass crossover; you can't drive each coil with a separate amp. You'll probably want as much power for the woofers as you had for the entire system when using a passive crossover, since bass uses the most power. Many people prefer tubes for the top end on the big Infinities. If you decide to go that way, you'd probably be OK with 30-60 watts per channel for the EMIMs, and 15 watts for the EMITs, maybe a KT88/6550 or 6CA7/EL34 amp for the EMIMs and a 6BQ5/EL84 amp for the EMITs.
 
Meh....it's a overrated speaker in every sense of the word really. I just scored a set, for cheap as this one needed work (we found out once it arrived) the owner gave me back a good chunk of the purchase.

Had two dead emits (I replaced those already thanks Bob douglas) I am replacing them all but by bit.

2 emims slap in one of the mid-range panels, at higher volumes. These have been sitting, you can see the diaphragms are sagging a bit. So replacing all those too bit by bit. As bob will release the emim diaphragms soon I'll be ordering !

The servo controller, isn't even a servo thank God I measured it before I hooked anything up. Two resistors burned up , and seems like a store. Repaired it by replacing the step pot on the input impedance. To a variable brush pot of some sorts, guranteed it doesn't reflect the proper resistive values .


I don't even bother repairing it, as a miniDSP is far superior.


So far ? Not a bad speaker, but for its size?? Not that impressive. It doesn't outperform my magnepan tympani as of yet. It might if I get it all dialed in and proper .
 
Last edited:
I know this thread is a couple of weeks old, but I'm still catching up after a week's vacation.

The idea of tape making the radiating area shorter had me thinking about the Kappa 9. This uses what they call a SEMIT (basically a half-length EMIT, same as the EMIT used in some versions of the RS3000, but with a different faceplate). Even though the Kappas aren't line arrays, could that have been for the same reason? It certainly isn't for high-frequency extension.

The rs1 isn't line array either common mistake ....they are most certainly not line array, those emims play different from one other 3 sections it's divided in.
 
Last edited:
The rs1 isn't line array either common mistake ....they are most certainly not line array, those emims play different from one other 3 sections it's divided in.
Post #80 in this thread explains what theory the RS1b is based on, and how they work. The front EMIMs and EMITs are divided into four frequency ranges.
 
Last edited:
Even though a member is replying to your post directly, it's likely they are just trying to more generically advise the rest of us as well.


audiokarma.

all audio, no attitude.
 
Even though a member is replying to your post directly, it's likely they are just trying to more generically advise the rest of us as well.


audiokarma.

all audio, no attitude.
Doesn't take away it's not a true line array because it simply isn't. Doesn't matter how many crossover points it has.

There is misinformation here about them.
 
Back
Top Bottom