Using LTspice to Simulate Klipsch Crossovers

[mboxler]

I started a similar thread on the Klipsch forum a few years ago that didn't get much interest. Even though that will probably happen here as well, I thought I'd still offer to assist anyone curious about how Klipsch crossovers work. I'm by no means an expert, so I know I'll learn some tips as well.

I chose LTspice because it's the only free software I could find that allowed for autoformer simulation. My driver simulations (inductors with DC resistance) are always a work in progress.

To get things started, here's a file I have of the Type AA crossover. It's named Type AA.txt because this site doesn't allow for a .asc extension. Once downloaded, rename it to Type AA.asc and open it in LTspice. At least I hope this work around works.

After running the simulation, then clicking on K33, In1, K55, and K77, you should get the following plots.



Mike

 

[Gigantic]

thanks for sharing this. I certainly need more rabbit holes to get lost in...

 

[ClaudeJ1]

thanks for sharing this. I certainly need more rabbit holes to get lost in...

I'd rather stay even with the grass. Just improved me Titan II bass horn by measuring the actual output of JBL 2220, horn, and a Hysteresis-Free 1.2 mH choke in series instead of the recommended 2.5 mH iron core by Dr. Edgar which was a "one size fits all" approach to any of the Four different 15" woofers he used over the years. RIP Bruce Edgar!

Simulations can be the "spice of life," I know, but Acoustic Measurements, Room EQ, and extended listening sessions over several days will keep Audiophile DIY gus from having the face the dirt in Rabbit Holes! HAH

 

[Gigantic]

Keep in mind, I've only been at this for 8 months; there is a lot of catching up and learning to be done! Fortunately, my partner tolerates my obssesive curiosities and only chuckles when new things appear in our flat, whether its a bigger horn or a new guitar farkle or gee-gaw. She seems to be enjoying the hi-fi nearly as much as I am and has definite opinions on the various things I've done, mostly supportive.

 

[ClaudeJ1]

Keep in mind, I've only been at this for 8 months; there is a lot of catching up and learning to be done! Fortunately, my partner tolerates my obssesive curiosities and only chuckles when new things appear in our flat, whether its a bigger horn or a new guitar farkle or gee-gaw. She seems to be enjoying the hi-fi nearly as much as I am and has definite opinions on the various things I've done, mostly supportive.

Of course. I'm glad you have those obsessive curiosities! They serve as a baseline for communication, of which, learning can only happen with contrasting experiences. We learn nothing from repeating what we already know, other than creating history during out journey to sonic nirvana, which we will never reach.

 

[mboxler]

Simulations can be the "spice of life," I know, but Acoustic Measurements, Room EQ, and extended listening sessions over several days will keep Audiophile DIY guys from having the face the dirt in Rabbit Holes! HAH

If REW shows a 6db spike, how do you determine the component values needed to flatten it? I can easily add/change components in LTspice to get real close to the values.

 

[mboxler]

Keep in mind, I've only been at this for 8 months; there is a lot of catching up and learning to be done! Fortunately, my partner tolerates my obssesive curiosities and only chuckles when new things appear in our flat, whether its a bigger horn or a new guitar farkle or gee-gaw. She seems to be enjoying the hi-fi nearly as much as I am and has definite opinions on the various things I've done, mostly supportive.

There's no way I would have learned how the components worked without LTspice. Once I figured out out to simulate the autoformer, I was hooked.

I'd be happy to help in any way. I also plan on adding a Universal model in the first post. Can model others as well.

Edit...Looks like I can't add it to the first post, so here's a Universal using a 3636.



Mike

 

[ericc129]

Of course. I'm glad you have those obsessive curiosities! They serve as a baseline for communication, of which, learning can only happen with contrasting experiences. We learn nothing from repeating what we already know, other than creating history during out journey to sonic nirvana, which we will never reach.

Yikes.....now we just need a simulation with the PRV 2000Ph on tap 2 or 3 and tweeter on tap 4

 

[ClaudeJ1]

There's no way I would have learned how the components worked without LTspice. Once I figured out out to simulate the autoformer, I was hooked.

I'd be happy to help in any way. I also plan on adding a Universal model in the first post. Can model others as well.

Mike

No way? How about many other ways in addition to? My comment was not against the LTspice way, it was commenting on the "rabbit hole" by a non-LT spice DIY guy. Curves, of which I'm a proponent require interpretation and judgment, then ACTION to improve what they discovered. There are more than just LT spice curves involved. I'll bet PWK never used LT spice either, but if it were available "in his day," he probably would have.

 

[ClaudeJ1]

Yikes.....now we just need a simulation with the PRV 2000Ph on tap 2 or 3 and tweeter on tap 4

OK, then you should do it.

 

[mboxler]

Yikes.....now we just need a simulation with the PRV 2000Ph on tap 2 or 3 and tweeter on tap 4

If you post a schematic, I can try to get close .

 

[ericc129]

OK, then you should do it.

I know......I was just joking......as it related to the other thread wherein everyone is starting to drop the prv driver a tap and/or find a correct shunt value vs bandbass

 

[mboxler]

After all these years, I may have found the missing piece of this puzzle. 10 years ago, someone on diyAudio wrote a program that, when fed a .frd file (REW), and a .zma file (DATS), will create a file to simulate a real driver in LTSpice. No more using a resistor as the load I created such a file using a K77M. Here's the model, no crossover components



And here's the resulting SPL plot



I should be able to quickly modify a crossover and see the real results within minutes. This is a game changer since LTSpice is the only software that allows me to use an autoformer.

 

[Panelhead]

You are a beast.
Can you simulate the AL4 crossover with a B&C DE-120, Altas PD-5, and Eminence Kappa 15C in a La Scala?
I bumped up the mid a couple dB to match better at crossover points. Sounds glorious with a big honking sub. I need MORE.

 

[Gigantic]

You are a beast.
Can you simulate the AL4 crossover with a B&C DE-120, Altas PD-5, and Eminence Kappa 15C in a La Scala?
I bumped up the mid a couple dB to match better at crossover points. Sounds glorious with a big honking sub. I need MORE.

you need a 4500X circuit to tame the rising rate of the B&C DE120. That will take care of it.

 

[mboxler]

You are a beast.
Can you simulate the AL4 crossover with a B&C DE-120, Altas PD-5, and Eminence Kappa 15C in a La Scala?
I bumped up the mid a couple dB to match better at crossover points. Sounds glorious with a big honking sub. I need MORE.

Right now, I can only simulate FR for individual drivers, and only the drivers I have. I do have the Altas PD-5, and I have an AL-4. A good test will be to measure the driver in a
K-400 (no crossover), measure the driver through the AL-4, and then simulate it in LTSpice. I have a CT-120 as well and can test the tweeter section. I don't have an Eminence Kappa 15C.

In the case of a crossover with no autoformer, XSim might be a better tool??? Unfortunately, I can't simulate an autoformer in XSim.

Hopefully my enthusiasm is justified. I've fooled myself way too many times

 

[mboxler]

I thought I would run a simple test using a 4uf cap in series with a CT-120 (DE10). This would represent the A-4500 tweeter circuit. The DE10 has a massive 50 ohm impedance peak around 1800hz, which I have shown to be an issue with the DE10. Sure enough, the simulation validates this.

White is the DE10's FR with no capacitor. Green is through the series 4uf capacitor. The FR with the cap is actually greater in that area. So far so good!

 

[mboxler]

You are a beast.
Can you simulate the AL4 crossover with a B&C DE-120, Altas PD-5, and Eminence Kappa 15C in a La Scala?
I bumped up the mid a couple dB to match better at crossover points. Sounds glorious with a big honking sub. I need MORE.

Did you keep the tweeter filter stock? I ran two simulations, the first is with the K77M, which I believe is what Roy used. Impressive FR.



This is with the CT-120.



I'll test using real crossover and drivers. This should let me know just how well the simulations are working, as this is a seven-element filter.

 

[Panelhead]

Did you keep the tweeter filter stock? I ran two simulations, the first is with the K77M, which I believe is what Roy used. Impressive FR.

When rebuilding the crossovers left tweeter section stock (no 10ohm resistors) and removed the fuse. My version of AL4 crossover had a jumper wire where the 10 ohm resistors were in earlier ones. And tweeter and midrange are negative polarity.
Saw where in earlier testing, you found a 4ufd cap across the DE-120 tamed the rising output. Was thinking of trying that when opening them up next time.
They actually are sounding so well balanced right now I do not know what to expect. Think switching to KAPPA 15C and bumping up the K55 a couple dB in crossover smoothed the midbass. Maybe helped the handoff from woofer to mid.

 

[Panelhead]

Correction, the resistors that the later AL4 did not have are the 5 ohm, 10 watt. In series with K77. So they are six element.