Dayton Audio RST28F-4 W/ SEOS Wave Guide.

[xmax]

I never claimed to be sane and yes I turned up the gain to get about the same output at 2K.
I will do more controlled measurements in a few. Again I was mainly only looking at FR
or anything crazy in the THD curve. For people who like good news.

[xmax]

Blue=RST @ 1600HZ 24LR (active), Red=RST with 6.2uf (only), Green= Satori (NO SEOS WG) TW29R all at same gain settings.
I used a test baffle and B&K mic this time around, these should be much more accurate. The RST with no cap
hit 103dB. As you can see the Satori does not like a low XO 1.6K 24LR (active).

[KEtheredge87]

With only a 6.8 uF cap, does that present a usable XO for the RST tweeter? Can't say I've seen a 1st order XO done for any tweeters lately. Most folks opt for 2nd order at least... but this FR looks pretty darn good. Just trying to take these results and interpret / predict the performance possibilities with this arrangement.

[xmax]

The distortion certainly is very low for a 6dB filter, often there are other issues with first order,
the overlap can be very wide and the tweet and mid being in phase at the XO with the proper
acoustic center, the 2 drivers are likely to be out of phase above and below the XO making
for some funky vertical dispersion. That said I will still try it! Also this time around it was a
6.2 uf cap and the crappy high distortion curves from the other night had the active filter
combined with a 6.8uf. Todays curves had either the 6.2 OR the active LR4.

[dcibel]

I never claimed to be sane and yes I turned up the gain to get about the same output at 2K.
I will do more controlled measurements in a few. Again I was mainly only looking at FR
or anything crazy in the THD curve. For people who like good news.

Haha I don't know xmax, this post appears far too sane to me ;)

I'm sure you can see then, that when posting two sets of measurements side by side, without providing the extra context of "I change the gain to level match at 2kHz", that someone might assume that the gain isn't changed, therefore the data is questionable due to the results. I'm really not trying to beat you down, I'm just not one to accept things blindly, with the technical details I will question what doesn't line up with my understanding of reality. Let's not blame the capacitor on this one, but the equipment operator that couldn't keep his fingers off the gain control ;)

Anyway, you don't need to take any more controlled measurements on my part, though speaking of measurements I recall some time ago you were going to compare several nice midrange drivers, the new RS52FN, the big expensive Volt dome, Scan Discovery 5.25", and maybe a couple others.. Did you ever get around to those drivers?

As far as the single cap filter goes, if you recall some years back Zaph did a waveguide TMM (on his site) that he did a crossover design for using only a single cap, then another design getting a bit more complex for a better response. It can certainly be done, but the driver's natural response is very important for it to be effective.

[Mike220]

I got away with crossing the old rs28a in that guide and a rs270p-4 with 9 parts 5 for the tweeter and 4 for the 10" woofer and 3 were resistors.

[xmax]

Haha I don't know xmax, this post appears far too sane to me ;)

I'm sure you can see then, that when posting two sets of measurements side by side, without providing the extra context of "I change the gain to level match at 2kHz", that someone might assume that the gain isn't changed, therefore the data is questionable due to the results. I'm really not trying to beat you down, I'm just not one to accept things blindly, with the technical details I will question what doesn't line up with my understanding of reality. Let's not blame the capacitor on this one, but the equipment operator that couldn't keep his fingers off the gain control ;)

Anyway, you don't need to take any more controlled measurements on my part, though speaking of measurements I recall some time ago you were going to compare several nice midrange drivers, the new RS52FN, the big expensive Volt dome, Scan Discovery 5.25", and maybe a couple others.. Did you ever get around to those drivers?

As far as the single cap filter goes, if you recall some years back Zaph did a waveguide TMM (on his site) that he did a crossover design for using only a single cap, then another design getting a bit more complex for a better response. It can certainly be done, but the driver's natural response is very important for it to be effective.

Don't worry I will lose my mind by the 5th page.

I regrettably assumed one could gather the gain had been monkeyed with because of the horn gain and obvious transfer function of the 6dB filter @ approximately
5K via cap. In this case you were right in doubting the distortion curves of those first measurements because they were F%&*ed.

We would never blame a capacitor (or resistor) of reasonable quality of anything. Jerk Face Xmax might say a resistor could cause a suspicious transfer function,
but he is sleeping in. He will have his fingers all over the place soon enough.


RS52FN and friends are waiting for nice Xmax to mount them.

Zaph was onto something with his WG project mainly taking advantage of the huge mid peak created by the WG
dealing with it in the electrical domain made for a useable response whilst letting the tweet creep low in a
leisurely fashion.

Everybody knows drivers natural response needs to be man handled electronically with 37 parts to sound good.

[Pete Schumacher]

don't worry i will lose my mind by the 5th page.

I regrettably assumed one could gather the gain had been monkeyed with because of the horn gain and obvious transfer function of the 6db filter @ approximately
5k via cap. In this case you were right in doubting the distortion curves of those first measurements because they were f%&*ed.

We would never blame a capacitor (or resistor) of reasonable quality of anything. Jerk face xmax might say a resistor could cause a suspicious transfer function,
but he is sleeping in. He will have his fingers all over the place soon enough.


Rs52fn and friends are waiting for nice xmax to mount them.

Zaph was onto something with his wg project mainly taking advantage of the huge mid peak created by the wg
dealing with it in the electrical domain made for a useable response whilst letting the tweet creep low in a
leisurely fashion.

Everybody knows drivers natural response needs to be man handled electronically with 37 parts to sound good.

lol!!!!

[itschris]

xmax, I always enjoy your posts.

I'm curious if anyone has thoughts on this combo vs the mini XT25 also in the SEOS-8. I'm currently using the XT25 with a very similar crossover to what Pete suggested in his XT25+SEOS thread. I'm planning on doing a build for a friend, and am deciding between sticking with that, or using the RST. Both look like very low distortion + flat responses, although it looks like the RST can cross lower. Input would be appreciated.

[xmax]

I will try the mini XT tomorrow against the RST, I have some somewhere and I guess they glue right on.

[TN Allen]

"...RS52FN and friends are waiting for nice Xmax to mount them. ..."

Just curious, what do you mean by the above? I've probably overlooked a previous post and also misinterpreted the above, however, are you planning to "mount" the RS52 in a WG?

[rpb]

Don't worry I will lose my mind by the 5th page.

I regrettably assumed one could gather the gain had been monkeyed with because of the horn gain and obvious transfer function of the 6dB filter @ approximately
5K via cap. In this case you were right in doubting the distortion curves of those first measurements because they were F%&*ed.

We would never blame a capacitor (or resistor) of reasonable quality of anything. Jerk Face Xmax might say a resistor could cause a suspicious transfer function,
but he is sleeping in. He will have his fingers all over the place soon enough.


RS52FN and friends are waiting for nice Xmax to mount them.

Zaph was onto something with his WG project mainly taking advantage of the huge mid peak created by the WG
dealing with it in the electrical domain made for a useable response whilst letting the tweet creep low in a
leisurely fashion.

Everybody knows drivers natural response needs to be man handled electronically with 37 parts to sound good.

What I like to do to compare distortion, is to match the responses closely by filtering the responses. If you can get the responses within 2 or 3 dB in the range of interest, then the distortion measurements are easier to interpret. If one tweeter is 8dB louder at 2k than another, the results at 2k are not comparable. If your interest is in crossing near 1k, then the same applies. You could even filter each driver for a LR4 at 1k, and go from there. If you look at your results in post #17, the red, blue and green all cross 900hz at the same spl. If you compare the distortion at 900hz, red, and blue are the same, and better than the green. Green has higher 2nd order, but is equal to red, and blue in 3rd order distortion. If you were to match the curves from say 800hz to 3k, then you could directly compare the distortions over that entire range.

[xmax]

"...RS52FN and friends are waiting for nice Xmax to mount them. ..."

Just curious, what do you mean by the above? I've probably overlooked a previous post and also misinterpreted the above, however, are you planning to "mount" the RS52 in a WG?

Exactly, and testing in a standard flat baffle.

[raiderone]

The distortion certainly is very low for a 6dB filter, often there are other issues with first order,
the overlap can be very wide and the tweet and mid being in phase at the XO with the proper
acoustic center, the 2 drivers are likely to be out of phase above and below the XO making
for some funky vertical dispersion. That said I will still try it! Also this time around it was a
6.2 uf cap and the crappy high distortion curves from the other night had the active filter
combined with a 6.8uf. Todays curves had either the 6.2 OR the active LR4.

I really like that you added a 6.8uf cap for comparison and so one could see the effect of the waveguide and the potential simplicity of a 1st order crossover. Please post your results of your attempt at a first order crossover. One of my favorite JBL speakers L100 (or 4312) used only caps on the mid and tweet and it seemed to work well (at the time).

[xmax]

Again, the second set of measurements used a 6.2uf just to be clear.