Thanks for bumping this back up.  Bob, thanks for all the good work.

There is a Dope from Hope about the change over from the spring clip (one piece phase plug) to the two-port phase plug (solder lug).  It shows extended response up to 6 kHz on the two port.

But, it shows that the glitch (peak) above 6 kHz is there on both.  It just moves a bit.  I've read some posts here and there claiming that the peak is characteristic of only one model.  That does not seems to be the case at all.  There seems to be anomalies above 6 kHz on all drivers of this type including the Electro Voice units.

- - -

I don't have any personal experience with the P-trap.   It has a lot of supporters though.  As intended, and as Bob shows, it filters out the peak.

It was first suggested to me by a good fellow (Mr. B.L.) who is a big Klipsch fan.  Mr. P. set up his system in general and installed the P-trap.  Mr. B.L. told me that his daughter remarked very favorably on the change without being told about it.  "What did you do to them?" 

Therefore, even if some ears don't hear a difference, others probably do.  I don't know if it is young ears versus old ears, or female ears versus male ears, etc.

- - -

The Vacuum Tube Valley article on the K-Horn investigated the P-trap and said it cured a "shouty" horn sound. 

- - -

The Heyer Audio review reported some issues attributed to the crossover zone to the tweeter.  Just maybe, the real problem was the glitch.  Pure speculation on my part.

- - - -

In my view, part of the problem is that the K-400 has on axis gain up above 3 kHz and this may still be climbing above 6 kHz.  Hence the glitch-peak from the driver come through.  It could be that other horns which people substitute do not have the same gain characteristics.  Therefore the glitch-peak is less pronounced.  So it could be that the alternate horns "solve" the glitch-peak by not amplifying it.  But the P-trap would do the same at far lower cost.

Edit addition:

- - -

I'll let Al K comment.  It is my understanding that Al.K's extreme slope crossovers also filter out the glitch-peak.  People have good things to say about those crossovers. 

Gil

Gil,

All of Al's networks filter out the glitches mentioned.  You don't even need to go to the extreme slope to get that benefit.  As you move up the line the slopes get steeper and eventually virtually eliminate driver overlap (the ES nets).

The hallmark of Al's designs are the bandpass on the squawker which makes all of this glitch business irrelevant.  Pretty smart.

I've used just about every one of his network designs now and they are very very good.

As intended, and as Bob shows, it filters out the peak.

Bob also noted that the P-trap knocks almost 2dB off the output across the range of the driver, which is almost as much attenuation as you get by dropping a tap on the autoformer. Bob believes this is what people are hearing -- it's hard not to agree with him (even though I'm always looking for a reason not to agree with him:)

Guys,

If the "P-trap" attenuates the squawker level in it's usefull range, then you have it wired wrong or you have an extremely poor inductor. It should have no impact on the sound at all!

BTW: my original universal "ALK" network attenuates the 9 KHz glitch by 7-8 dB. My ES networks and the "P-Trap" kill it DEAD!

AL K.

Bob,

Here's the analysis of the P-trap.

Loss is abouy 0.33 db in the squawker range and impedance is equal to the assumes termination of 13 Ohms.

At 9170 Hz the loss is about 18 dB and present 169 Ohms to the source. The notch deapth (18 dB here) is a direct function of the inductor loss which is assumed to be .173 Ohms and DC and DCR+ Q of 200 at 9 KHz.

BTW: I tried a P-trap in my Cornwall network long ago before I realized it didn't need one. The loss was measured and was negligable. The coil I used was .105 mHy. Q=12 at 10 Khz and a DCR of about .52 Ohms.

Al K

Bob,

I think the key phrase is "acoustic output". Acoustic measurements is a real bag of worms subject to all sorts of errors. If you are measuring the effects of a change in a network measure the output of the network directl, not the acustic output. I would be extremely surprised if the difference between trap in versus trap out will be different acoustically than electronically using a resistive 13 Ohm load in place of the K55.

AL K.

Bob,

If you are going to do it by acoustic output the way would be to have the trap in for one sweep than short it with a clip lead and run the test again. You would also need to run many traces to make sure each trace repeats exacty. You might also try ploting the voltage across the driver using it as a load. Another thought: replace the trap with a resistor equal to the DCR of the inductor and plot that.

Al K.

Al Klappenberger:

Bob,

If you are going to do it by acoustic output the way would be to have the trap in for one sweep than short it with a clip lead and run the test again. You would also need to run many traces to make sure each trace repeats exacty. You might also try ploting the voltage across the driver using it as a load. Another thought: replace the trap with a resistor equal to the DCR of the inductor and plot that.

Al K.

 

Bob,

The inductive reactance of .1 mHy at 1 KHz is about .63 Ohms. That plus the DCR of the .1 mHy inductor is probably the difference. You should be able to measure the same loss change in the network directly as voltage change across a 13 Ohm load. In any case, the loss is insignificant expecailly conidering the very idea of the P-trap is just a jury-rig fix for the glitch problem anyhow. Anything you add will give you some loss. The question becomes a judgment of how important the added loss is. Chances are the total loss will actually be less once you put a better quality cap in place of the 13 uF of the AA network anyhow. Another way to look at it is that I think the squawker is set too hot on the AA network anyhow. The bottom line is that it's really a moot point. There are too may other things going on to worry about it!

Al K.

BEC:

Michael, I would say that every driver tested here looks good for a Lascala or Khorn driver with the exception of the K-52H which does not go low enough. It looks good for a Cornwall or a Heresy. As far as standouts, the K-55V with 2 piece phase plug has the cleanest trace and is very flat to 6.5 Khz. But, the particular one I tested at least gives up a couple of dB in output compared to the K-55V single piece phase plug and the K-55M. The K-51V extends as high as the K-55V 2 piece phase plus and has a bit higher output than the others. Not quite as flat a curve as the K-55V 2 piece phase plug. I am just surprised how really good these drivers are. I think the first four are pretty much interchangable. Once I got the test method down to very few test induced errors, the drivers look great. Now, another surprise is the peak at 7.5 khz on the K-55M. If someone wants to get upset about the little peak on the K-55V 1 piece phase plug at 9 Khz, they should go nuts over the one for the K-55M. Also notice the one the K-52H has at 9.5 Khz. Bob Crites

This K55M peak, I believe, was what PWK was referring to when he talked to me about the AK network for the Khorn. Jim Hunter had done the design and removed the Zener clamp from the tweeter and went to a steeper slope. But the main reason they did it was to compensate for that peak. PWK said that as far as they were concerned, Atlas was out of the driver business and the M driver was made by EV with a ceramic magnet instead of the Alnicos used in the Atlas driver. This was early August of 1985.

Claude