Deang Posted February 9, 2017 Share Posted February 9, 2017 Pretty cool project. http://www.northreadingeng.com/Forums/viewtopic.php?f=7&t=21 1 Quote Link to comment Share on other sites More sharing options...
Honeybadger Posted February 10, 2017 Share Posted February 10, 2017 Too technical for me. But cool project. HB Quote Link to comment Share on other sites More sharing options...
SiliconTi Posted February 10, 2017 Share Posted February 10, 2017 Interesting. Too bad there is that 12db valley at the crossover point, but I doubt you can hear it. I wonder if it was worth the effort... Quote Link to comment Share on other sites More sharing options...
John Warren Posted February 10, 2017 Share Posted February 10, 2017 1 hour ago, SiliconTi said: Interesting. Too bad there is that 12db valley at the crossover point, but I doubt you can hear it. I wonder if it was worth the effort... There is not -12dB valley(?). The two individual responses cross at -6dB but the actual response must be measured. The green plot in the graph is what the reposnes would be if the two drivers had their acoustic centers located the same distance from the listener. (Hey Dean!) Quote Link to comment Share on other sites More sharing options...
SiliconTi Posted February 10, 2017 Share Posted February 10, 2017 13 hours ago, John Warren said: There is not -12dB valley(?). The two individual responses cross at -6dB but the actual response must be measured. The green plot in the graph is what the reposnes would be if the two drivers had their acoustic centers located the same distance from the listener. (Hey Dean!) Ah, yes, I was not paying attention. I am just starting out in the Klipsch Heritage world, and after listening to the La Scalas I have, I can see the power of being able to adjust each driver individually. I may have to go with an active x-over, but will give them time to break in first. Interesting info, thanks for posting it. Quote Link to comment Share on other sites More sharing options...
John Warren Posted February 18, 2017 Share Posted February 18, 2017 Here's a cardboard model of the filter board taken from the Gerber file. The gain pots from each channel plug into the board at 90° 3-pin headers located on the bottom of the board (J1, J2). Rail voltage will be pulled from the filter caps to on-board Zener regulated supplies which will provide +/- rail voltage to the AD713 packages. In this configuration just about any quad package op-amp with low THD will work and I'll outfit the filter board with 14-DIP sockets to determine if that is the case. Small signal connections from the filter board (J3, J4) to the amp board will be made using the existing board jumpers (see the jumpers on the amp board). Installation should be very simple, the filter board will have a pair of L-brackets that attach to the Aluminum frame using existing holes in the frame. Each channel of the filter is then attached to each channel on the amplifier board using a couple of small signal cables outfitted with plugs and done. So the SPICE model of the entire amplifier with the filter shows it works so it must be ok!!! Quote Link to comment Share on other sites More sharing options...
The Dude Posted February 18, 2017 Share Posted February 18, 2017 John, I love how far you take things. Couple of questions. From what I gather this is somewhat of a dsp or crossover built directly into the amplifier? Is this only going to suit only Klipschorns? Quote Link to comment Share on other sites More sharing options...
John Warren Posted February 18, 2017 Share Posted February 18, 2017 2 hours ago, The Dude said: John, I love how far you take things. Couple of questions. Excellent! Quote From what I gather this is somewhat of a dsp or crossover built directly into the amplifier? It's an analog filter that is ENORMOUSLY simplified by using op-amps, in this case as amplifiers. If it was done using discretes the board would be the size of the amplifier (maybe not that large but you get the point). It's an amplifier with the mid and HF active filters built in. So channel 1 is the midrange and channel 2 is the HF. Quote Is this only going to suit only Klipschorns? Nope. There are many variants possible by simply changing out the resistors and capacitors. The topology is a 24dB Sallen-Key bandpass and 24dB Sallen-Key highpass for both filters. There are websites, SPICE models etc. that provide the R-C combinations for just about any responses wanted within the confines of the topology. The slopes can be changed too but I haven't thought that out yet at the board level. Conceptually it's easy. 1 Quote Link to comment Share on other sites More sharing options...
John Warren Posted March 4, 2017 Share Posted March 4, 2017 Getting there. The board fits in the little D45 without too much pain. Using the same connectors as the ones used by Crown makes it a bit easier. I've got a Crown XLi500 coming in next week to determine is this type of mod can be performed on it. 1 Quote Link to comment Share on other sites More sharing options...
jorjen Posted March 4, 2017 Share Posted March 4, 2017 Very nice John! Keep it coming. Very interested in the development of this project. Quote Link to comment Share on other sites More sharing options...
Deang Posted March 4, 2017 Author Share Posted March 4, 2017 I wish Crown still made and sold this amp. 1 Quote Link to comment Share on other sites More sharing options...
John Warren Posted March 4, 2017 Share Posted March 4, 2017 3 hours ago, jorjen said: Keep it coming. Very interested in the development of this project. First couple photos are the filter installed in the amplifier. I've also attached a couple of response plots. The is the first pass on the analyzer with a sinusoidal sweep. Both CH1 and CH2 have an 8 Ohm power resistor as a load. On the midrange plot, I knocked the mid down 2 "clicks" on the potentiometer, the tweeter plot was 1 click/sweep. The analyzer has only 8 colors for a screen shot. The range settings between the two drivers are uch greater than what I'm showing. You can shut the HF off and blow up the mid if your so inclined. Note that the attenuation appears to be logarithmic. So nothing blew up, the amp is stable and the distortion numbers are still excellent given the AD713 has no significant distortion. Note that I did install a socket for each channel which allows for so flexibility. The board has it's own rail power supplies to which can be set for optimizing whatever op-amp is used. The slate gray cables are the Crown cables that feed the gain pot. They do a half-assed job of feeding them thru the amp, clearly an after thought. So I've given the amp a real colonoscopy. Next is to drive and measure the top section of a Klipschorn. Then I can go back a "tweak" the filter. 2 Quote Link to comment Share on other sites More sharing options...
John Warren Posted March 5, 2017 Share Posted March 5, 2017 The black and white leads run back to the +/-V rails and provide power to the board. I didn't have the right size solder lugs on hand to make the rail connections to the power supply so I added bleed-down resistors and soldered the +/- leads to the hot sides of the resistors. The D45 powers supply is +/- 31V. Bigger filter capacitors in this amplifier reduce ps ripple alot. The op-amps are dual supply and that's handled by Zener regulated supplies on the filter board (+/-10V). I changed out the factory small signal leads going to the front panel gain pots. The factory leads were too tight (they were too tight even for the stock configuration). I made them about 1.5" longer which made installing the filter a ton easier. Quote Link to comment Share on other sites More sharing options...
John Warren Posted March 6, 2017 Share Posted March 6, 2017 First swag at an acoustic response, Klipschorn top section. The purple is with the mid filter out, and the blue is active. With the values selected for the filter, the break-up of the K55 is completely attenuated, the driver is essentially silent above 5kHz. It would take a handful of coils and caps to do that passively. I gave the setup a listen, the midrange was "pleasant" enough (but I'm biased!), The ability to adjust both mid and HF drivers while listening to the Klipschorn from 20' appeared to be more useful that I'd first thought. Regarding different crossover frequencies, there's a lot of flexibility in the design, limited entirely by the drivers. Quote Link to comment Share on other sites More sharing options...
John Warren Posted March 6, 2017 Share Posted March 6, 2017 4 minutes ago, John Warren said: Regarding different crossover frequencies, there's a lot of flexibility in the design, limited entirely by the drivers. SPICE simulations changing lo and hi pass sections. Quote Link to comment Share on other sites More sharing options...
John Warren Posted March 10, 2017 Share Posted March 10, 2017 On 2/18/2017 at 9:34 AM, John Warren said: In this configuration just about any quad package op-amp with low THD will work and I'll outfit the filter board with 14-DIP sockets to determine if that is the case. Appears to be the case. 20kHz %THD with AD713 (blue) and TLE2074 (purple) packages ramped to full power output on the tweeter channel. There's potential for audible noise on the high pass so resistors values were limited to a few k Ohms, metal film with <1% tolerance. Quote Link to comment Share on other sites More sharing options...
John Warren Posted March 12, 2017 Share Posted March 12, 2017 Ok, more acoustic responses on the Klipschorn top section (i.e. more fun with active, integrated filtering). The mic is about 4 feet from the top section and position so it's axially at the same height as the top of the K401 mouth. The compression driver is the K55M. The response is smoothed somewhat at 1/24 octave. The nets are 4th order (24dB/oct) band-pass for the mid and high-pass for the horn tweeter. I tweaked the midrange filter get a bit more out of it between 5 and 6kHz. So first graph shows red plot the tweaked filter and the blue is the first prototype that is a "model" Sallen-Key build. The midrange output is controlled by the CH1 potentiometer. The second graph is a plot (green) with the K77 turned up, mic in the same location as above. Response with mic in this location is within about 8dB from 500Hz up to the limit of the tweeter. The third graph is the result of first shutting down the midrange and plotting the tweeter response (blue) then turning the mid back up and dialing the tweeter down. It shows how the two individual responses make up the system response. The forth graph is what the midrange does with level adjustments at the front panel gain pot. The tweeter does the same sort of thing. 1 Quote Link to comment Share on other sites More sharing options...
mikebse2a3 Posted March 14, 2017 Share Posted March 14, 2017 Very Impressive John...!!! miketn Quote Link to comment Share on other sites More sharing options...
babadono Posted March 15, 2017 Share Posted March 15, 2017 @John Warren have you ever tried LME49720 Opamps? There duals only, there used to be a quad LME49740 I think but no longer avail. So you would have to get some kind of 2 duals to quad adapter. I redid the guts of two DBX-4BX dynamic range expanders using them and was impressed by their performance. Anyway very cool project, I wish had the time to delve as deep as you. Quote Link to comment Share on other sites More sharing options...
John Warren Posted March 20, 2017 Share Posted March 20, 2017 On Wednesday, March 15, 2017 at 2:39 PM, babadono said: @John Warren have you ever tried LME49720 Opamps? There duals only, there used to be a quad LME49740 I think but no longer avail. So you would have to get some kind of 2 duals to quad adapter. I redid the guts of two DBX-4BX dynamic range expanders using them and was impressed by their performance. Anyway very cool project, I wish had the time to delve as deep as you. No, haven't considered it but thanks for the suggestion. Truth is the filter configuration is not too sensitive to the type of op-amp used (gain bandwidth does not limit performance). I also have four XLi1500 Crown amplifiers here for mods. I've installed the filter in one and the responses and distortion numbers are good! These are a bit easier to mod than the D45 because there's more room to work in. One idea is to use the amplifier with active, internally integrated steep slope filters to drive the Klipschorn or Jubilee bass sections (or any bass horn). The %THD of the XLi amplifiers are quite low below 1kHz. I've started a thread on my forum to describe the teardown, measurements, etc. http://www.northreadingeng.com/Forums/viewtopic.php?f=14&t=22 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.