Modding the Dayton RS52

[Daveis]

I've seen posts online about using felt discs to alter the RS52's peaks around 13khz.

But what happens if you take the back cup off it?

And has anyone used the RS52 in a Unity Horn Waveguide?

Is there anyway to improve the response of the RS52 down-to 500hz? It seems most people cross over at 700-800hz. Can you go lower? Zaph's THD testing indicates it would start struggling quickly at 500hz.

Any way to horn load and get one more lower octave?

 

[dwk123]

I used the RS52 in a Unity style prototype. It was the 'tweeter', and I used a pair per side dayton RS150-4's in the 'mid' position. It worked 'OK' for a prototype, but you don't get all that much low-end extension from the horn. The setup was good from ~300 up to ~6k or so. I had really hoped for more low end, but I didn't iterate on the design so I'm not sure if it could have been improved.

FWIW, I bought the Yorkville U15's after doing that prototype. That probably says something.

 

[Daveis]

One more thing...

The RS52 sounds clean, but somehow a little metallic. I suspect I am hearing the 13khz peak even though I'm using a 24db/octave electronic crossover at 4000hz. I'll switch to a steeper crossover and see if that tames it.

Any way to tame that acoustically with foam without messing up the response down low? I suspect I will eventually tire of the sound because of this one aspect. (I don't usually like aluminum dome tweets) It's a shame because over 800-3500hz this is a very nice inexpensive driver.

 

[Daveis]

What are the major sound qualities you associate with the unity horn?

 

[angsuman]

Hmm, I use 3rd order passive crossover on my RS52 and I get an effectively 4th order acoustic slope, my peak at 13khz is down about 32dB from the main bandpass area. I haven't really noticed this peak. I have listened to the RS52 by itself and it sounds very clean, but you hear extra sibilance and sizzle when the peak is undamped. Maybe its your tweeter that is contributing to the "metallic" sound?

 

[Daveis]

Hmm, I use 3rd order passive crossover on my RS52 and I get an effectively 4th order acoustic slope, my peak at 13khz is down about 32dB from the main bandpass area. I haven't really noticed this peak. I have listened to the RS52 by itself and it sounds very clean, but you hear extra sibilance and sizzle when the peak is undamped. Maybe its your tweeter that is contributing to the "metallic" sound?

I was using this combination(below) until I swapped out the planar for the RS52. In general it (the RS52) was an improvement in imaging coherence, vocals, bells, drums. Pretty much a win-win all around except for a slight sense of metallic sizzle.

Having dome dispersion for upper midrange is a new experience. It may take awhile to get used to that. I will double-check levels with RTA pink noise. It could be that I have the mid up a few db's. I've found that I'm pretty sensitive to any treble being even a few db's too high.

50-800 hz two Scanspeak 8554 8" woofs
800-4000hz BG RD40 planar
4000-20khz Scanspeak 6600 textile coated tweeter

 

[dlneubec]

Zaph's test of the RS52 show a pretty good sized F3 harmonic distortion peak at around 4200hz. Perhaps you should try crossing a little lower and maybe steeper, if your tweeter can handle it.

 

[DSP_Geek]

Zaph's test of the RS52 show a pretty good sized F3 harmonic distortion peak at around 4200hz. Perhaps you should try crossing a little lower and maybe steeper, if your tweeter can handle it.

The F3 peak is a direct consequence of the 13 kHz FR peak, since the FR peak will amplify motor distortion. The original poster might be on the right track with wanting to clean up the peak instead of hiding it in the crossover.

 

[Daveis]

I've tried crossing over the RS52 at 2000,3000 and 4000 with 24db/octave.

It definitely sounds less metallic at 2000 and 3000 so maybe the 4200 peak should be dealt with if possible acoustically through felt pads.

I haven't figured out how to use a driver for such a small range like 800 to 2000. You have the tweeter and woofer overlapping so much the mid is getting clobbered on both sides.

 

[Hartono]

Hi,

This might be of interest, have a look at Enabl+Mamboni thread. I've applied Enabl pattern on a couple of paper driver, and Soongsc did experiment on fullrange aluminium driver (Jordan).

Hartono

 

[dlneubec]

I've tried crossing over the RS52 at 2000,3000 and 4000 with 24db/octave.

It definitely sounds less metallic at 2000 and 3000 so maybe the 4200 peak should be dealt with if possible acoustically through felt pads.

I haven't figured out how to use a driver for such a small range like 800 to 2000. You have the tweeter and woofer overlapping so much the mid is getting clobbered on both sides.

Perhaps you might try the RS52 a little lower at say 600-650 LR4 and then at then at 3600. That's clsing on 2.5 octaves ane might be doable.

You could also try to go steeper as well on the mids low pass to tweeter high pass to get the 4.2k F3 peak further down and/or try an unbalanced slope and experiment with reversing tweeter polarity if necessary to get a good phase match.

Maybe you can experiment with strips of 1/4" or thicker wool felt, maybe 1" wide at various distances from the top and bottom of the driver to see if you can effect the peak at 13kkhz as well.

 

[Daveis]

Perhaps you might try the RS52 a little lower at say 600-650 LR4 and then at then at 3600. That's clsing on 2.5 octaves ane might be doable.

You could also try to go steeper as well on the mids low pass to tweeter high pass to get the 4.2k F3 peak further down and/or try an unbalanced slope and experiment with reversing tweeter polarity if necessary to get a good phase match.

Maybe you can experiment with strips of 1/4" or thicker wool felt, maybe 1" wide at various distances from the top and bottom of the driver to see if you can effect the peak at 13kkhz as well.

Thank you all for the ideas...

I did some listening tests today with some Jimmy Buffet tracks.

On the low end, I tried 500,600,800 hz. I settled for 800hz. Male vocals were too thin sounding at 500-600 hz even with 12db/18db crossover. The sound is clear and coherent, but just doesn't have enough weight on male vocals.

On the high end, I tried crossovers of 2000,3000,4000. It seems crossing over anywhere from 3000 to 4000hz is workable. Above 4000hz, the RS52 is beaten soundly by the Scanspeak 6600. I tried 12db,18db,24db,42db crossovers. 24db/ocatave sounds remarkably similar to 42db/octave. Both 18db/oct and 24db/oct are pleasant.

Only heavy guitar distortion songs (eg. Tool) really have too much bite up high. I can't hear any problems with mellow music like Jimmy Buffet.

I think I'm going to use a steep crossover on the mid-to-treble and call it good.

If I read this site http://www.speakerdesign.net/dayton_rs/midrange/rs52/dayton_rs52.html

I think I can conclude that a felt disc in the middle plus taking the back off will help a little bit.

 

[dlr]

I wouldn't take the back off

If I read this site http://www.speakerdesign.net/dayton_rs/midrange/rs52/dayton_rs52.html

I think I can conclude that a felt disc in the middle plus taking the back off will help a little bit.

I have measured the driver with the back removed, though I didn't do any testing beyond that. I just had too much else going on, I'll revisit that one day. If you remove the back you're going to alter the Fs considerably since it acts as a closed box for the driver. In addition, it will then have to be in its own chamber of your box, otherwise a woofer will back-modulate it and possibly damage it. It won't sound good in any case unless isolated.

Dave

 

[Daveis]

Re: I wouldn't take the back off

I have measured the driver with the back removed, though I didn't do any testing beyond that. I just had too much else going on, I'll revisit that one day. If you remove the back you're going to alter the Fs considerably since it acts as a closed box for the driver. In addition, it will then have to be in its own chamber of your box, otherwise a woofer will back-modulate it and possibly damage it. It won't sound good in any case unless isolated.

Dave

I currently have the driver in its own separate box along with a tweeter. I've been impressed recently with the sound of upper midrange drivers (BG planar and Tang Band) run dipole.

Well, the distortion specs are good even with the back cup on so I guess it can stay.

I'll try various amounts of felt at the apex of the cone. I will also try a plastic 12-oz coke bottle cap also for fun.

Another driver I heard recently is the Tangband-W4-1337SA . It's been measured in the same Zaph test group as the RS52. Its distortion is higher that the RS52(at least in the 800-400hz range), but it's frequency response has no nasty peaks.

One thing I'm wondering, if one tries to add a phase plug to the RS52, what's the appropriate size, shape and material to do it with. It seems felt is just an approximation of what the manufacturers would normally do.

 

[Patrick Bateman]

What are the major sound qualities you associate with the unity horn?

This was addressed to Doug, but since he didn't reply maybe I'll take a stab at it?

The one thing that I noticed was clarity and coherency. There is something special about having the sound from all drivers "in sync." I can see why the "real" Unity is used so much for PA; the clarity of speech just gives me goosebumps. I was listening to Johnny Cash and I could barely tear myself away from the speakers.

I dig into these observations further in my thread about my Unity clones here:
http://www.audiogroupforum.com/csforum/showthread.php?t=62789

 

[Daveis]

Spent more time with the RS52.

I've concluded that that RS52 is a winner when used from 800-4khz with LR4.

Switching from cone to dome midrange takes some getting used to.
It's this difference in sound that at first sounds like you are hearing
the peak at 13khz. That's not the case. It's a matter of getting used
to a dome mid.

I would like to thank Zaph and others for putting me onto this driver.

I've recently compared a

1. Seas Froy
2. Scanspeak Solist
3. 3-way system with Tangband-W4-1337SA crossed around 400.

to a RS52/Scanspeak 6600 tweet combination

and I prefer the RS52 in upper midrange position.