Jump to content

I'd like to thank Bob Crites (my warranty department)


Deang

Recommended Posts

Every once in a while I run into a problem I can't figure out, because I don't have all of the test equipment that I'm supposed to have. Never felt a need for them since I'm mostly just a builder. My thinking was that as long as I could test the networks to make sure they're passing a signal, and sound good when I hook them up to my LaScalas, then I'm good to go.

This only works as long as I don't branch out into new designs that I can't test, and sure enough, the first time I tried to do it, I got bit real bad.

About a year ago I came up with what I thought was a pretty nifty idea. What I liked about the idea was that it was simple, and "simple" means I get to maximize the quality of the parts, namely the capacitors.

An ever increasing popular mod is to use the big honkin BMS driver with Dave's or Greg's mid horn, and one of the aftermarket tweeters. I really like the sound of the Type AA built with the Jupiters or Jensens, and the idea that came to me was to scale the primary cap for the new impedance, and let her rip. The only real problem I had was finding someone with the setup to help me nail down the attenuation, which would effect the capacitor value. I also ran a poll, and the consensus was that 12dB down sounded the best. However, I still needed the opinion from someone running the setup, because I knew that all of the opinions were from those running ALK-like networks, which incorporate the swamping resistor -- so I couldn't be 100% sure until I got someone's ears on the project.

Results led me to believe that I had to at least provide two settings, -12dB and -15dB, which was easy enough to do. On the last set I built, I removed that option, deciding just to trust the math and go with -12dB down. I told the customer that it was an easy fix if he needed more attenuation.

After receiving the networks I got a strange response. I was told that while it was evident that a lot of good things were happening, something didn't sound right.

"Hi Dean. Well I got them hooked up. Bass is overpowering the mid and top end. By the time I turn it up enough to actually hear the vocals, the bass is pounding and shaking the room, literally ... I really like what the caps are sounding like, they just are way, way low."

After some troubleshooting, I told the customer, Jay, that these would have to be sent to my warrant department.

"Warranty Department?"

"Yeah, you need to send them to Bob Crites. He don't always hear real good but he's the best troubleshooter there is. Where do you live?

"Arkansas"

"Well, shipping won't be to bad then."

"I can drive to Bob's, he's about 40 minutes from me."

"Would you mind?"

"Not at all, will he care if I just show up?"

"I'll let him know you're coming."

"Will he be okay with this?"

"Oh yeah, he loves a good puzzle, and gets that nice warm feeling because I screwed something up - and then I get the test equipment lecture. It's painful, but I only have to endure it every couple of years or so."

"Okay, I'll head over."

What transpired after that was a series of emails that amounted to this:

"The traces look terrible. Something is wrong, but I can't figure out what's going on."

"I couldn't have messed it up that bad, it's only got six parts. I'd say bad autoformer or a bad cap, but it's both networks."

"Yeah, it's strange. I'll figure it out."

"I hope so, because this customer has been waiting two years, and he's been issuing death threats."

Okay, what ensues next is the craziest of rides.

"Well, By now we have replaced every part, one part at a time and it still does not work. So, now we are down to things like bad solder joints, something conductive where it should be insulating, something insulating where it should be conductive, misaligned dilithium crystal, intermittent flux capacitor operation or ... I am going to reheat every solder joint."

One day later ...

"To show you how crazy this is getting, I have just taken all the stand-offs loose and insulated them from the board in case the finish or wood was conductive. That was not it.

Heated up all joints, no change. Used jumpers to connect to each end of each wire to rule out a broken wire inside the insulation.

To summarize:

Have replaced every part, one part at a time.

Breadboarded another one that tests good.

Heated up all solder joints

Jumped every wire

Reminds me of “Y” School in the Air Force. Troubleshooting Crypto Equipment, the instructors would put in a bug and time us finding it. They would do things like put in a marked 5 ohm resistor that really measures 50 ohms or take a wire smaller than a hair and solder it between two points on the back of a circuit board. A wire so small as to be invisible at a quick glance.

I must be missing something, but darn if I can think of what it could be. Got to finish up the shipping for the day right now."

Okay, at this point I'm starting to worry. Bob usually has things figured out for me by the end of the day. My biggest fear was a bad batch of capacitors, because dealing with distributors and manufactures is a nightmare, even if you have solid evidence from a state of the art spectrum analyzer. I also had the problem that it was a new product, and there several in the field using them. This of course left me with the question, why isn't anyone else complaining? I did have one odd experience with a single customer who told me they sounded great, and then later told me they were unhappy with them. Because of the contradictory feedback, I wrote that one off as a fluke.

At this point, it would be more interesting to hear the rest from Bob's point of view. If he isn't up to the typing, I can finish, but I would like to give him a chance, because he tells a great story.

Jay will have to decide if he wants to jump in or not.

What started this was my own personal experience with my LaScalas. It really didn't take long for me to figure out that I really liked what I heard from the Type A and AA using the Jupiter and Jensen capacitors. I'd been down this road before, and even though I already knew what I liked, I decided to go through it all one more time, building up some Universals and a set of SuperX's. During this timeframe, I found Alex Crespi from Uptone Audio, who had parted ways with Robert Hovland, and had resurrected the MusiCap. Alex hooked me up with his single distributor, and in no time I was ready to listen to a set of AAs loaded with the new version of the MusiCap, one of the finest film and foils ever made. The sound was outstanding, but after a couple of hours, I decided that I had to knock 3dB off of the midrange. I'm fully aware that many like the stock setting, but It just doesn't sound balanced to me. However, and I've been very vocal about this -- if I build with Jupiters or Jensens, the stock settings sound just fine.

When given the choice or ability to attenuate, almost everyone, without exception, will choose to take a little bit off of the midrange. Now, there are two ways to do this; the first is to choose your attenuation level, and then rescale the primary capacitor value so the crossover point stays the same. As you attenuate, the reflected impedance that the amplifier sees from autoformer goes up, so the capacitor value goes down. Now, this is very cool, because it allows us to use some very good capacitors at the primary position and otherwise (because smaller good caps cost much less than larger good caps). There are two "downsides", 1) the attenuation level is fixed, and 2) impedance changes with frequency, and though this is completely normal, zero feedback amplifier designs will find their frequency response tracking the impedance, which will result in non-flat response. The HF is tipped up somewhat because of the higher impedance, but in my experience, SET listeners actually prefer this. Regarding the fixed attenuation; if the math is right, and you employ the use of parts conducive to a positive horn listening experience -- you won't find yourself second guessing your autoformer settings.

The second way is the way most of us are used to seeing; a resistor is used to "swamp" the network back to a relatively stable 8 ohms. This allows for "on the fly attenuation", since no matter what setting you choose, the impedance "seen" by the amp stays roughly the same. There are other considerations, but I don't want to discuss any of them except one: The 8 ohm impedance. The lowered impedance requires a very large primary cap. With the last iteration of the Universal, this capacitor value went to 56uF. So, the trade off is this: to get adjustable attenuation, you have to stay with a metallized capacitor at the primary position, and for me, that's is the complete antithesis of what I'm trying to accomplish with my high end builds.

My choice in build is to eschew adjustable attenuation and a large metallized primary capacitor in exchange for properly fixed attenuation and something like a Jupiter or Jensen for the primary capacitor, which is responsible for passing/carrying the majority of the full range signal to the rest of the network/drivers.

So, I built a pair of networks consisting of seven parts, and with exception of two parts, built exactly like a Type AA. The additional part is a .40mH coil running from tap -12dB to squawker positive. The other change is the Jupiter primary capacitor, scaled down because of the higher impedance.

If anyone wants to takes a shot at trying to figure out what I messed up, knock yourself out.

A big "thank you" to Bob for taking the time out of his busy week to not just figure out and rebuild my networks, but for saving an important part of my product line, thanks Bob!

Edited by Crankysoldermeister
  • Like 2
Link to comment
Share on other sites

I assume Bob looked at the response.  You never mentioned how he tested them, just that he went nuts "mechanically" with the parts.

 

What did the response (attenuation and crossovers) look like?

 

Did he listen to them and confirm they had an audible issue?

 

He said he bread boarded one with his own parts.......and it was good?????????

 

Did the customer wire them up properly?  (Sorry).  :(

Link to comment
Share on other sites

I'm afraid I don't have any more than an elementary understanding of the technical problems you are describing.  But in my view, if the problem is not what you think it is, it might be what you don't think it is.

 

Could the problem be in the pre amp or amp?  This might be tested by using a different AVR, amp, etc.

 

My other question is with the customer LS.  Does the XO work in an other LS?

 

I would do some investigating of the customer's setup, and reduce the audio chain to the least common denominator, to eliminate variables.

Link to comment
Share on other sites

Bob eventually figured out - I left you all hanging to see if you could figure it out.

It's safe to assume that Bob is competent. I mentioned "traces", and there was plently of electronic testing. Trust me, showing you the traces isn't going to help you.

Anything you can think of was tried - I put everything into the post.

Link to comment
Share on other sites

I suppose a clue is in order.

When the bread boarded version was built, it was NOT identical, even though Bob thought it was. The bread boarded version was built by Todd, who accidentally built it right.

So, the genius who came up with the idea messed it up, while a guy who had better things to do that day managed to wire it right.

I don't think Bob was happy that it worked. :)

Link to comment
Share on other sites

Bob eventually figured out - I left you all hanging to see if you could figure it out.

 

You stink!  :P

 

If it took Bob some time to figure it out and he had the network IN HAND, what makes you think us schmucks can figure it out?

  • Like 1
Link to comment
Share on other sites

I doubt Dean would make this error, but I have.

 

Reversed the input wires on the auto former.

 

Dave

 

I haven't given up on this, but I think you are on the right track.  The question should be, what would make the woofers run so hot?

 

We know this:

 

When the bread boarded version was built, it was NOT identical, even though Bob thought it was. The bread boarded version was built by Todd, who accidentally built it right.

So, the genius who came up with the idea messed it up, while a guy who had better things to do that day managed to wire it right.

 

 

 

Todd, who was using the same components, managed to get it right.  So it's not a component failure.

 

Genius, who messed it up, still had the idea essentially correct, which must mean crossed wires somewhere.  Not crossed to attenuate the woofers, but crossed so twice as much power was sent to the woofers.

 

I'm saying the woofer and squawker were crossed on the schematic.  Edit:  Nah, that doesn't make any sense.  I bet the auto transformer wasn't grounded properly, or wasn't grounded at all, sending a hot signal to the woofs.

 

If I get this right, I expect to be assigned to the Warranty Department.

 

Contingent of course, upon me learning how to solder.  B)

Edited by wvu80
Link to comment
Share on other sites

Guest David H
The question should be, what would make the woofers run so hot?

 

Reversing the input leads would cause additional attenuation at the squawker.

 

Dave

Edited by GotHover
Link to comment
Share on other sites

Well, I actually had this thing figured out pretty quickly.  The proof of my solution was to be when I had Todd breadboard this entire circuit exactly as Dean had built it.  But, it worked when Todd did that, so threw me off the main track for days.  As Dean mentioned, Todd made a mistake in breadboarding the circuit that FIXED IT.  So, that mistake caused me to think I was wrong about the problem being what I originally thought it was.  That sent me off on all sorts of tangents and a flurry of testing for unlikely causes.

 

Anyway, Dean was building what was essentially an AA crossover with much more attenuation than a standard AA has.  He correctly scaled the input (to the autotransformer) capacitor from 13uF all the way down to 1.5uF.  Now Klipsch and the rest of us building the AA crossover always just connect the tweeter to the downstream side of the 13uF input capacitor.  Dean did that also with his scaled down input capacitor.  Normally the input cap (13uF) is so large that you can ignore it as far as the tweeter circuit is concerned.  But at some point, as you decrease that value, the cap would start acting as a part of the tweeter circuit. And at a value of 1.5uF, that makes a very strange tweeter circuit.  Actually the strange tweeter circuit becomes an almost dead short somehow and drags down the midrange circuit.  I am sure that someone could probably draw an impedance vector to show the effect, but been too long since I did that to want to try.

 

OK Solution:  Connect the tweeter circuit to the amplifier side of the autotransformer input cap.   That is what Todd did by mistake when I had him duplicate the circuit on a breadboard.  If he had not made that mistake, the circuit would not have worked and next step would have been for me to tell him then to connect the tweeter circuit to the input side of the cap and test again.  Circuit would have worked and we would have been done!

  • Like 2
Link to comment
Share on other sites

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.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...