AudioXpress article

0 Members and 1 Guest are viewing this topic. Read 2206 times.

Jazzman53

Re: AudioXpress article
« Reply #20 on: 23 Aug 2017, 05:04 pm »
Why did you go to the stretched wire, then? Wasn't the welding rod easier to use?

The only advantage to using welding rods is that you don’t have to build a stretching jig, as they are perfectly straight right out of the box.   
And now the downside:   
If using use light diffuser grids to support the rods (as I did), it’s a pain to run a glue line across every rung of the grid and the glue line tends to stand a proud of the rods’ surface rather than laying perfectly flush, as desired.  Also, welding rods are only 36” in long so if you want panels taller than 36” you would either have to butt/solder them (very difficult) or butt but not solder them, in which case you would then have to connect power leads at both ends.   And you would have to over-coat them too if you want an extra measure of arc resistance.

I opted for wire/oak lattice stators because I wanted my panels to look as good as they sound and also match the nice oak woofer box & frame assembly.  The wire panels are also more finely segmented to give even smoother dispersion (beats any curved panel, hands down).  The biggest headache with wire panels is designing a proper stetching jig—and mine works perfectly.  After agonizing for a considerable time over the jig design, building it was fairly easy. 

The interlocking oak stator lattices were a bit time consuming but not all that hard to build—I just cut and profiled some boards on the table saw and router table and then sliced out the individual pieces (from the profiled boards) on the table saw. Some woodworking experience is required but you don't need to be an expert (I’m not).

I would be happy to provide dimensioned CAD drawings for my stretching jig and stator lattices to anyone who wants them.   

I can tell you that I would not build welding rod / light diffuser stators again.  In fact I still have the ones shown on my website which I would sell for less than it cost me to build them -- but then the buyer would lose out on the biggest reward of all, which is the DIY experience.   

josh358

  • Full Member
  • Posts: 1050
Re: AudioXpress article
« Reply #21 on: 27 Aug 2017, 12:12 am »
Yeah, the jig is the scary part, but the dimensioned CAD drawings would be a real help.

Agree that segmentation is the way to go -- I've been waiting for someone to make a commercial segmented line source stat, but unfortunately, the market for large panels isn't very good right now . . .

Jazzman53

Re: AudioXpress article
« Reply #22 on: 27 Aug 2017, 12:05 pm »
Yeah, it's a shame that no commercial source that I know of makes a segmented line source ESL.  Technically, ER Audio in Australia has a kit speaker that qualifies but it has only 2 segments-- a wider bass-mid panel and a separate narrow tweeter panel, which would not provide the same smooth trending dispersion of a multi-segmented wire panel.

I think you can build a better ESL than you can buy right now.

JohnR

Re: AudioXpress article
« Reply #23 on: 28 Aug 2017, 12:08 am »

Jazzman53

Re: AudioXpress article
« Reply #24 on: 28 Aug 2017, 02:58 am »
I stand corrected   :oops:

josh358

  • Full Member
  • Posts: 1050
Re: AudioXpress article
« Reply #25 on: 28 Aug 2017, 02:53 pm »
Yeah, it's a shame that no commercial source that I know of makes a segmented line source ESL.  Technically, ER Audio in Australia has a kit speaker that qualifies but it has only 2 segments-- a wider bass-mid panel and a separate narrow tweeter panel, which would not provide the same smooth trending dispersion of a multi-segmented wire panel.

I think you can build a better ESL than you can buy right now.
I suspect you're right.

There have actually been a number of segmented ESL's but they typically still beam and can have less than uniform dispersion with frequency.

Also, there's a difference between the crossover/segmentation approach and the time delay approach pioneered in the ESL-63. The time delay approach is better. It produces a better polar pattern, avoids the lobing issues in a crossover, and also allows you to achieve higher output with lower distortion because all frequencies play across the entire panel (although you might get more IM).

Jazzman53

Re: AudioXpress article
« Reply #26 on: 29 Aug 2017, 01:58 am »
My panels are finely segmented symmetrically about the centerline using 15 discrete wire groups. The RC line functions as a series of low-pass filters between the wire groups which creates a wider and smoother trending dispersion pattern than a curved panel of similar dimensions.  The RC filters attenuate frequencies predominantly but also impart a slight delay (to the extent that the charging time for the inter-wire capacitance component is greater than zero).  I haven't calculated the delay component but, in any case, the net effect of the segmentation scheme is remarkably apparent by merely walking around the speaker while it's playing and it's also evident in directivity sonograms comparing unsegmented panels, curved panels, and multi-segmented panels like mine:

http://www.diyaudio.com/forums/planars-and-exotics/246846-time-esl-builder-14.html#post4163636

Johnny2Bad

Re: AudioXpress article
« Reply #27 on: 29 Aug 2017, 03:21 am »
Very impressive build, you are very much correct to be proud of the build.

I was wondering if you could give some more information about the amplifiers you are using (power output, impedance rating), and if there are any special considerations when choosing amplifiers.

Jazzman53

Re: AudioXpress article
« Reply #28 on: 29 Aug 2017, 08:30 am »
I was wondering if you could give some more information about the amplifiers you are using (power output, impedance rating), and if there are any special considerations when choosing amplifiers.

The system includes the hybrid ESLs and a pair of Ripol subs in a trip-amp arrangement consisting of a Behringer DCX2496 crossover feeding three vintage Carver TFM-25 amps, each amp rated at 225W/channel into 8 Ohms.  The crossover frequencies are 50Hz & 228Hz using the Behringer's 48db/octave Linkwitz-Riley filter slopes.

Whereas conventional stat panels are predominantly a capacitive load, my multi-segment RC wire panels are predominantly a resistive load.  With the resistors in circuit, the capacitance reflected back to the amps is roughly equivalent that of first 2 wire groups in the series.  So, practically any quality amplifier could drive my panels, whereas less stable amps could get smoked driving conventional stat panels.  Even so, I would recommend at least 100 w/channel.

The Carvers had no problems driving my old perf metal panels either, BTW. 
   
« Last Edit: 29 Aug 2017, 12:58 pm by Jazzman53 »

Skeeboingen

  • Jr. Member
  • Posts: 18
Re: AudioXpress article
« Reply #29 on: 8 Oct 2017, 03:05 pm »
I was wondering about using direct servo subs in the ripole.  What's your opinion?

~S

Jazzman53

Re: AudioXpress article
« Reply #30 on: 10 Oct 2017, 09:49 am »
I was wondering about using direct servo subs in the ripole.  What's your opinion?

~S

I have no experience with Servo subs so I will not offer an opinion on this.