Deang Posted October 21, 2015 Share Posted October 21, 2015 Lol. Link to comment Share on other sites More sharing options...
John Warren Posted October 21, 2015 Author Share Posted October 21, 2015 (edited) Even when the sock-puppet requested plots, I provided them Ok... The sock-pupet formally requests the fillowing plots on your network: * Complex impedance in R magnitude and polar phase seen my the amplifier using real driver loads. I can accept R +-jX too if you like. * Group delay and insertion Phase through the mid-range port. * Absolute insertion loss of the midrange channel. (no shelving) * Absolute loss through the tweeter filter. (No problem here - yours doen't provide shelving). BTW: I already gave these plots to Lee a day or so ago in case this came up.. Maybe he will post them for me after you post yours. Al K. So what's going on here... I'm speculating that the ALK has made estimated plots using it's PCFILT software based on what it thinks the net topology and L and C values are. Here, he's publically admitted to making a first attempt at reverse engineering and has also admitted to transferring that result to another. That said, to calibrate the model predictions it is getting from its PCFILT simulator and complete the reverse engineering of the network, the ALK is requesting the complex impedance, preferably with a non-inductive resistive load (the added reactance of actual loads complicates the simulation a tad). If the data is provided by yours truly, it then becomes a matter for the ALK to run the simulation with a few iterations of L and C values until the predicted impedance from PCFILT matches the actual data provided. The group delay of the signal through the network is another simulation parameter. I have interest in providing as much information to customers as is necessary to make an informed decision. Providing data that enables one to reverse engineer is clearly beyond that. I have no interest in having a design I've develop reverse engineered. I don't care if one does so, I'm not going to help however. There's a good reason to show the impedance, I'll show the frequency range from 20-400Hz and that's shown below. It's the impedance of the network over the frequency with the insert operating with the Klipschorn folded bass unit I used to run initial performance plots. The data goes to 30kHz. I also considered being "cheeky" here and contemplated cobbling together an ALK "universal" network and then provide the information requested above as data from my net. It would have been amusing to see the ALK "explode" with that one! It would be easy to do but would take the better part of a full day especially if I wish to provide frequency response plots on a Klipschorn too. Edited October 21, 2015 by John Warren Link to comment Share on other sites More sharing options...
John Warren Posted October 21, 2015 Author Share Posted October 21, 2015 (edited) This was a double post. Sorry about that. Edited October 21, 2015 by John Warren Link to comment Share on other sites More sharing options...
Moderators Travis In Austin Posted October 21, 2015 Moderators Share Posted October 21, 2015 Even when the sock-puppet requested plots, I provided themOk... The sock-pupet formally requests the fillowing plots on your network: * Complex impedance in R magnitude and polar phase seen my the amplifier using real driver loads. I can accept R +-jX too if you like. * Group delay and insertion Phase through the mid-range port. * Absolute insertion loss of the midrange channel. (no shelving) * Absolute loss through the tweeter filter. (No problem here - yours doen't provide shelving). BTW: I already gave these plots to Lee a day or so ago in case this came up.. Maybe he will post them for me after you post yours. Al K. So what's going on here... I'm speculating that the ALK has made estimated plots using it's PCFILT software based on what it thinks the net topology and L and C values are. Here, he's publically admitted to making a first attempt at reverse engineering and has also admitted to transferring that result to another. That said, to calibrate the model predictions it is getting from its PCFILT simulator and complete the reverse engineering of the network, the ALK is requesting the complex impedance, preferably with a non-inductive resistive load (the added reactance of actual loads complicates the simulation a tad). If the data is provided by yours truly, it then becomes a matter for the ALK to run the simulation with a few iterations of L and C values until the predicted impedance from PCFILT matches the actual data provided. The group delay of the signal through the network is another simulation parameter. I have interest in providing as much information to customers as is necessary to make an informed decision. Providing data that enables one to reverse engineer is clearly beyond that. I have no interest in having a design I've develop reverse engineered. I don't care if one does, I'm not going to help however. There's a good reason to show the impedance, I'll show the frequency range from 20-400Hz and that's shown below. It's also the impedance of the Klipschorn folded bass unit I used to run the initial impedance plots. The data goes to 30kHz. I also considered being "cheeky" here and contemplated cobbling together an ALK "universal" network and then provide the information requested above as data from my net. It would have been amusing to see the ALK "explode" with that one! It would be easy to do but would take the better part of a full day especially if I wish to provide frequency response plots on a Klipschorn too. You Sir, are a Real Engineer, and a Real Gentleman. Well said. Travis 1 Link to comment Share on other sites More sharing options...
mark1101 Posted October 21, 2015 Share Posted October 21, 2015 Are you talking about the pair that Roy sabotaged at the Jubilee get together - the ones he ran out of phase with the blown tweeters? "There's something wrong with those Klipschorns..." - Bobby Crites Yes.......you were there too. So, is this a joke or fact? Everyone knew those Khorns sounded very bad. But I wasn't aware of anything purposefully done. I just had my ear tuned to my system, and the second many of us heard those things we all knew something wasn't right. Link to comment Share on other sites More sharing options...
Deang Posted October 21, 2015 Share Posted October 21, 2015 I think the tweeters were blown, and no, I don't believe Roy sabotaged the system, lol. Link to comment Share on other sites More sharing options...
jwc Posted October 21, 2015 Share Posted October 21, 2015 I actually stepped out of the audition room once that demo started. Immediately sounded bad. Instead I went back to the main lab area and ate some of dwilawyer's redneck nachos. I do recall though that the "jubscala" sounded good though. However, an active network was in that mix. Link to comment Share on other sites More sharing options...
Al Klappenberger Posted October 21, 2015 Share Posted October 21, 2015 I have interest in providing as much information to customers as is necessary to make an informed decision. Providing data that enables one to reverse engineer is clearly beyond that. I have no interest in having a design I've develop reverse engineered. OH BROTHER!! I have a lot to do today, so my socket puppet will be posting the plots I sent him a couple days ago for me. You are free to reverse engineer my ES5800 squawker / tweeter network if you think you can. As to reverse engineering the ES500, just look around. I already released the design for it and the Universal long ago. No need to reverse engineer that. As to reverse engineering yours, no problem, I did that in a few minutes to make the early computer study of your shelving adjustment without a swamping resistor claim! Lee even drew the schematic from your pictures and a typical crossover schematic I drew for him to use as a guide. AL K. Link to comment Share on other sites More sharing options...
John Warren Posted October 21, 2015 Author Share Posted October 21, 2015 (edited) The green plot is ALK Universal midrange using one of Dean's 3619 -6dB tap. I then took the network for the B&C-Selenium setup and put it on a tap close to -6dB as Dean's I got the blue response. I also added a resistor (but not a shunting resistor). Not bad! Above 6kHz is attenuated and the peak at 5.4kHz is knocked back. I would use the new network with the Crites B&C tweeter. The low end needs a bit more trimming but overall a respectable response. I use a -5.6dB tap in the setup for the insert. For this I used -6.1dB (or so). Edited October 21, 2015 by John Warren Link to comment Share on other sites More sharing options...
Arkytype Posted October 21, 2015 Share Posted October 21, 2015 John, I was able to "reverse engineer" your network schematic in less than a half hour. Anyone can look at the images you posted, draw out the components and follow the connecting wires. With the exception of an unswamped autoformer, it's a pretty generic design. The schematic was then sent to Al who used his software to compute probable component values. If other forum members would like to see the probable component values, I can post it. The reason Al asked you to provide those electrical plots was, (a) to determine if you knew how to measure the parameters in question and, ( if you haven't done the measurements yet, what the heck are you doing in the crossover network business? So, now that we've reverse engineered your "autotransformer based crossover network" to our complete satisfaction, what excuse will you now proffer not to publish them? Attached is a plot of two midrange ES sections showing how closely each pair of networks Al builds is matched. Do you measure each pair of network you build to insure similar matching? Wanna share a plot? Another plot is the complex impedance of an ALK network as seen by the amplifier. The third plot is the shelving range available for the ES500 high filter output and ES5800 high output. Not bad for a SOCK PUPPET!! SmartWool Sock Puppet Link to comment Share on other sites More sharing options...
Arkytype Posted October 21, 2015 Share Posted October 21, 2015 The emoticon should have been a (. Thanks Obama. Lee Link to comment Share on other sites More sharing options...
Arkytype Posted October 21, 2015 Share Posted October 21, 2015 Diddled by the dangling dong of Destiny!! OK the emoticon should be the lower case second letter of the English alphabet surrounded by parentheses. Lee Link to comment Share on other sites More sharing options...
Al Klappenberger Posted October 21, 2015 Share Posted October 21, 2015 I didn't send the plots of insertion loss through the squawker port of the ES500 + ES5800 netwrok set to Lee earlier so I am trying to upload and include it now myself. It is simply one the earlier plots magnified though. I haven't been on here since the forum software was upgrade, so how to do it is a mystery! I hope it works! Anyhow, the plot was taken of voltage across the swamping resistor and transformer input. The marker shows about 0.8 dB in the center of the squawker range (2 KHz). BTW: I am considering writing a step - by - step procedure on how to reverse engineer as crossover network. "Somebody" is going to need it! It sholdn't take long considering the one in question is simply a hacked Klipsch network. While using it as an example, I certainly would not reveal any element values in fear of someone actually trying to build it! Al K. 1 Link to comment Share on other sites More sharing options...
Al Klappenberger Posted October 21, 2015 Share Posted October 21, 2015 It looks like the other plot went up.. Lee forgot to post the plot of complex impedance of the network as seen by the amplifier. Here it is along with the test setup. The woofer is a real K33 in a closed box with a slot to emulate the loading of a horn. It's a nearly flat 6 Ohms, the actual impedance of most vintage K33 drivers. AL K. Link to comment Share on other sites More sharing options...
Al Klappenberger Posted October 21, 2015 Share Posted October 21, 2015 AHHhhhh , what the heck. In for a pennie in for a pound. Here's the computed response of the ES500 and ES5800 along with the actual measured response of group delay, insertion phase and forware attenuation of the ES500 highpass channel. AL K. Link to comment Share on other sites More sharing options...
Schu Posted October 21, 2015 Share Posted October 21, 2015 I suspect the thread has now Peaked ^... all down hill from here Link to comment Share on other sites More sharing options...
Al Klappenberger Posted October 21, 2015 Share Posted October 21, 2015 The green plot is ALK Universal midrange using one of Dean's 3619 -6dB tap. If the thread has peaked, I suppose it's time to move on the the next item: First off, the 3619 is not "Dean's", it's MINE! The 3619 and 3619-ET are both exclusive only to me. The only transformer Dean has is that oversized wax-dipped budget buster! I forgot Universal Transformer company's designation number for i t. Secondly, If that plot is through my Universal network, somebody must have picked a bad horn too. The B&C DCM50 doesn't look that bad! Al K. Link to comment Share on other sites More sharing options...
mikebse2a3 Posted October 21, 2015 Share Posted October 21, 2015 (edited) Last I checked it was the complete Loudspeaker System that we hear and experience and the Final Acoustical Performance of the complete system (ie: Drivers/Horns, Networks and Physical Layout of the components all must be designed to work in synergy).Obviously real world Drivers/Horns, Crossover/Balancing Networks and Physical Layout of the Cabinet/Driver/Horn components aren't perfect and this fact means the Crossover/Balancing Network must have the ability to take all the imperfections and minimize them for an Optimum Loudspeaker Design.My Points Are:(1) The Ultimate Measurements of a Loudspeaker System is the Acoustical Performance Measurements of the Total Loudspeaker System (ie: Frequency, Polar, Phase, Dynamic Linearity and Transient Performance)with of course Controlled Listening Test as the Final Judgment of Performance Quality and Success of the design.(2 )Crossover/Balancing Networks "Must Be Designed" not with "resistors as terminations in a simulator" but with consideration of the actual components(Drivers/Horns) and their physical location relative to each other in the physical cabinet/layout of the loudspeaker system.(3) Any Crossover/Balancing Network simulation/measurement that doesn't take into account these factors from Point(2) above may look good/great in simulations/measurements but fail in Acoustical Performance and Listening test.miketn Edited October 21, 2015 by mikebse2a3 Link to comment Share on other sites More sharing options...
Deang Posted October 21, 2015 Share Posted October 21, 2015 I'm not sure what I sent to John. The unit they build for me is the UT-3670. It sells for $10 more. If you have a problem with the way the unit is built, take it up with UT. I asked them to build it the way they thought it should be built, and the 3670 is what I ended up with. I figure they know what they're doing. Everything is built on something, isn't the 3619-ET just a 3636 with an extra input tap and a few additional output taps. Yeesh, you sound like Daffy Duck, "MINE, MINE, MINE!" So where did you get the idea to ask to have extra input and output taps added to accommodate the next generation of compression drivers? How about 1dB steps on the Universal and your other builds? The labeling? How about Gentle Slope Networks, which I just called, "low order types", when I suggested designing some for people who might prefer them over the ESNs. How about the ditching of the kits, who saved you from that nightmare? What do you prefer, gluing your parts or strapping them (btw, thank you Dave Harris for the boards, horns, and the half dozen other things you probably do). Does anyone ever get credit for anything with you? You misrepresented my builds and work earlier, to include criticisms regarding the way I wire and solder my parts together. I create a mechanical connection, and then use the solder to "paint" the connection. I normally connect the lead of a capacitor or resistor to the much stronger coil wire. Sometimes I'm forced to use standoffs, but it's not my preference. At any rate, what I do works, and apparently works well -- I haven't had a failure since I started doing this. Link to comment Share on other sites More sharing options...
John Warren Posted October 21, 2015 Author Share Posted October 21, 2015 The responses shown in the plot are the K401 and PD5VH. Link to comment Share on other sites More sharing options...
Recommended Posts