John W's 1" Mini Smith Horns

John W built this pair of very nice "1200Hz" Mini-Smith horns. He decided that he didn't need them and put them up for adoption and Norealtalent bought them. I intruded on their transaction and asked if they could be diverted on their journey eastward by stopping off in SF for a quick measurement session. I'll follow this post with a few measurement plots, a comparison with it's big brother the 2397, some distortion plots and an interesting discovery.

I was hoping to take physical measurements of his horns and post a drawing of it. Unfortunately they are significantly different from the original 2397 and to accurately make a drawing of them I would need to open one up. Relax, Mr. Talent, I didn't do it.:D As it turns out these are very nice little 1" throat horns and I know I would appreciate it if John could share his dimensions with us.


Widget

Here is a plot that compares two 2425s, an LE85 and an LE85 with a 2421 diaphragm all mounted to one of the Mini Smith horns. (All drivers were provided courtesy of Zilch) All measurements were taken at ~1m on axis with great care taken to avoid moving the horn during driver changes. The horn and driver were perched on top of one of my loudspeakers that is ~45" tall and ~24" wide. Since my room has a ceiling height of just over 10', the ideal location would put the center of the driver at 5'... I figured ~4' was good enough. I also put a 24" by 48" sheet of 3" Sonex over my speaker's baffle to reduce any reflected sound.

First I shot the 2421... I wasn't that impressed. Next I shot the stock LE85... it looked pretty damned good. It made me suspect the condition of the first driver. Next I shot one of the 2425s.... whoa??? That's odd. I would expect HF differences, but lower frequencies shouldn't vary that much. I tried the next 2425??? Essentially the same response??? Could they both be damaged somehow? (The two 2425 measurements were in error. I'll explain in a minute.)

Here is a plot comparing the distortion of the LE85 and one of the 2425s. It is difficult to read with all of these plots on one graph, but interestingly both 2nd and 3rd harmonic distortion were higher with the LE85. Giskard may have some JBL Data on this... I'll speculate that this LE85 has passed it's "use by" date.

Now I wanted to compare the LE85 on the Mini Smith and also on a 2397 with a 2327 adapter. But first... the mystery of the harsh midrange measurements on the 2425s is solved. I put the LE85 back on the Mini Smith horn and shot it again as a quality control.... oops. Now it had the same peaks and dip at 1800Hz as the 2425s. I noticed that the Sonex had slipped down the baffle about 4-5" exposing some wood below the lip of the horn. I moved the Sonex back up and the curve straightened right up... it was exactly as the first round. This isn't simply a measurement anomaly... running the MLS noise I could hear the difference. The implications for those of us using this... and most likely most other horns is huge. I will be looking into this more deeply later.

Anyway, back to the comparison. I placed the LE85 on the Mini Smith with the properly adjusted Sonex and generated a plot, and then changed the set up putting the front lip of the much larger 2397 in about the same location as the Mini Smith's was. Obviously with the 2328 and 2327 adapters combined with the much larger horn, the driver was significantly further away. Anyway as can be seen from the curves, the little guy does a much better job from 1600Hz up than the adapted 2397.

I am sorry that I didn't run through the 2425s again, but I am beat and I think we have a pretty good idea how the little horns work. Based on what I saw with the Sonex misplaced on the LE85, the 2425 plots are reasonably accurate above 2.5KHz, and below that I would expect that they would track the first two drivers pretty closely.

Widget

Very impresive "little horn"!@!:D

Nice work guys.

Ian

Thanks Widget. I can't wait to check them out. I have my own theories about the Smith based horns. I think the 2328 adapter is a key factor in getting the smooth response acquired with a large format driver on the 2397. I also think a scaled down version 2328 would be appropriate for the smaller drivers or it may not matter. I love this stuff, there is always something to challenge the mind...:bouncy:

Thanks Mr. Widget ! for the effort to measure and post the response of the 1" Smith-Horn as interpreted by John W. . John W. , certainly does nice work ! Kudos to you John .

Some Questions for Mr. Widget

#1 ; What distortion percentage do these traces represent ?
-
#2 ; What was the voltage drive for the distortion study ?

Quote:

---

Originally Posted by Mr. Widget

(snip) ,,,,, but interestingly both 2nd and 3rd harmonic distortion were higher with the LE85. Giskard may have some JBL Data on this... I'll speculate that this LE85 has passed it's "use by" date.

---

- My speculation would be, that it's the tangential surround ( as implimented by JBL ) that give those elevated distortion specs for the le85.

- Here's why I don't think that the distortion is age related aluminum fatigue ( or work hardening ) . The following are some pics of two different 288-8K ( diaphragms ) on the same driver mounted to 803 multicells . One is obviously "pooched" while the other was newly installed after delivery from GPA. The other CLIO shot of the 288-8K ( on a leCleach horn ) is included as a more detailed reference of a 288-8K response .

- These pics were "borrowed" from the Altec forum and belong to sfogg ( shawn is a member here also ) & vuki . The original thread is located Here !

- One can handily "see" the FR differences that a worn out aluminum diaphram ( sporting a tangential suspension ) displays. As you can see, I feel "age-related" aluminum fatigue shows up quite handily as a reduction in frequency reponse ( at both ends of the spectrum ).

:)

EDIT : Ignore the response spikes above 20K in the first two pics . They are definately some sort of measurement anomaly .

Widget - 'any chance of trying that 1" horn with one of Zilch's JBL 2407 neodymium drivers? They appear to go quite a bit higher than the drivers you tested with, and might be the almost perfect match.

John

Quote:

---

Originally Posted by Earl K

Some Questions for Mr. Widget

#1 ; What distortion percentage do these traces represent ?

---

We've been around the block on that one before, Earl. As far as I know CLIO doesn't present a simple % number and one probably is about as valid as saying that a driver has a 16 ohm impedance... these things are really frequency dependent. I don't have the math chops to figure it out, but all of the data you need is before you in that plot. The distortion is raised by 30dB as stated on the plot. You should be able to calculate the percent distortion at any given frequency and then use what ever method of averaging best suits your needs.

Quote:

---

Originally Posted by Earl K

Some Questions for Mr. Widget

#2 ; What was the voltage drive for the distortion study ?

---

I didn't keep track. I would say that if we assume that these drivers on the Mini Smith are 3dB less sensitive than they are on the 2397, see the last plot, and we know that JBL claims a sensitivity when mounted on the 2397 of 108dB/watt, I'd guess that they were being powered by ~10dB less than one watt or, ~0.1watt. Again, I am basing this entirely on what you see before you. You can interpret it differently.

Quote:

---

Originally Posted by johnaec

Widget - 'any chance of trying that 1" horn with one of Zilch's JBL 2407 neodymium drivers? They appear to go quite a bit higher than the drivers you tested with, and might be the almost perfect match.

---

Highly unlikely. I am triple booked right now. Sorry.


Widget

Thanks Mr. Widget for taking the time to put these to the test. I am a little relieved that they seem to pass muster.
I am a little curious about the baffle diffraction in your test set up. It would be nice to see a picture of how that came about.
As far as dimensions are concerned, I followed the plans that where nicely drawn up in the original 1200hz horn thread here:
http://audioheritage.org/vbulletin/showthread.php?t=5454
The only deviation I made was to pull the throat opening back even with the arc made by the fins and carve a smooth transition from round to square.

John,

Very nice work.

Pending the measurement conditions and mounting of the horn within a baffle the response may well be smoother than it appears.

For comparison here is a scan of an IMP measurement done on my JBL 2397 horn with an JBL adapter using my JBL 2420 driver with a JBL DR82525 titanium diaphram. A 077 slot extends the response above 10,000 hz and small cone driver below 1000 hz. The purpose of these measurements was to develop a passive crossover network. The time interval seperating the drivers can be seen on the time displacement trace.

The detail is in the devil. The IMP had a calibrated mic file. The measurement was done at 2 metres with the 2397 resting on a table with the front lip extending over the edge. The a professional measurement mic was mounted on a camera tripod. Heavy open cell foam was placed underneath the horn and above attached to a light fitting.

As you can see the response is very smooth apart from the dip of about 4.5db at 1000 hz which was a crossover issue. The divisions are 5 db .

These measurements were arranged by an engineer, Doug Tipping in late 1997 (as can be seen from the date)

I quite liked this horn although I no longer have the 2397 or the adapter.
They were sold and shipped to someone in the USA at great expense!

Based on your work I maybe tempted to build a mini smith horn!

Ian