Getting back into the swing of things.

Kornbread

School is out!  Time to catch up on things around the house and get back to s.t.u.f.f. 

This is a two way project using a FaitalPro 5fe120 and an xt25tg30. 

XSim data is in-box measurements @39" @0*, on tweeter axis, but inside which means the gate times were really small.  The mosquitoes are still too bad for outside measurements.   

 

And this is the result measured without moving anything.  Red is with polarity reversed.  Green is xsim model.  I feel there may be a bad connection with one of the cheap alligator clips but have redone just about everything with no difference.      

 

Measured impedance.  Green is XSim, blue is measured.  Close until 1k. but is that small of a difference responsible for the measured .frd response?        

 

dcibel

The most likely culprit is incorrect acoustic offset data entered interested xsim, or incorrect phase data in the measurements used in xsim. The simulation shows everything in-phase at the crossover frequency, however phase is not aligned in reality.

I'd verify the offsets in xsim first.

dcibel

As a verification, once the offset in xsim is corrected, you should see the simulation match up with the measurements you've already taken above.

jhollander

XSim put a little (m) next to the S driver label when you modify the sensitivity or delay.

Kornbread

@dcibel ; Yep, you are correct.  I never noticed.    

Kornbread

More better. 

The offset wound up being 1.25".  Thought that was a bit too much but ... pink is XSim, blue and red are measured.  Wonder what affect all those cheap jumpers have on the equation?

Using some reclaimed magnet wire to roll the inductors.  It's really too small for the larger coils, but it's free.  Can't find any motors at work with anything larger than ~18 gauge which may work fine on smaller coils; ordering 10lbs. of 15gauge magnet wire from eBay ($93) for the larger stuff.  The larger gauge should prop up the bottom end.  Damaged the threads on the only stainless steel bolt that is long enough to hold the homemade roller together, so probably not rolling any more inductors today anyway. 

 

jr@mac

Offsets always surprise me, good evidence that they can be difficult to determine without measuring. 

Kornbread

These have been sitting around waiting for a while.  The drivers are mentioned in the first post.  Had to tinker with the crossover a little after figuring the offset (thanks guys) and because the new coils with the larger diameter wire have a lot less resistance. 

Dunked the coils in poly as Ben suggested so they will be ready to go when we return from hauling boy back to college tomorrow.

Playing around with xsim looks like the inductors can be left alone and still have plenty of adjustment for voicing.  They have some bad spots on the clear that needs attention but if these sound good, they are small enough to lug around to a meet ...              

        

Kornbread

The crossover has evolved a lot from the last post ... build, listen, change, repeat. 

Running through the spotify fav song list and things are starting to gel, only after prolonged listening did those pesky s'sss become noticeable.  Where are the s'ssss at in the freq. range, ~5-7k?     

Wolf

5-8k.

Kornbread

Sweet Jane by the Cowboy Junkies confirmed it.  These are nowhere as bad as the carver amazings, but still there.  On the amazings, that song is nearly unlistenable due to their accentuation of Timmin's s'ss.     

Thanks Ben.  That coincides with what was measured.  Had a ~2.5db hump that started from ~5.5 and ran up to ~9k.  Dropped that to just over 1db but I guess that is a fairly broad area of frequency in a sensitive area of our hearing so the s'ss were still a bit too much.  Easy enough to fix without affecting anything else.           

Don't know if this is the right way to go about it, but I think the basic crossover is sound.  There are 3 resistors that can be adjusted with little affect on the phase which should allow for adjusting the overall tweeter level, that pesky peak ~7k, and the depth of the trough from ~1-3.5k. <this one needs kept in a certain range or it does affect the phase.  Think I'll solder and pot the other crossover parts leaving those 3 resistors loose and just listen for a few weeks while swapping resistors around.                          

rjj45

Kornbread said:

The crossover has evolved a lot from the last post ... build, listen, change, repeat. 

Running through the spotify fav song list and things are starting to gel, only after prolonged listening did those pesky s'sss become noticeable.  Where are the s'ssss at in the freq. range, ~5-7k?     

As always when trying to swat pesky anomalies, it's very useful to listen to the problem tracks with one driver at at time. Not demonstrating any great mastery, just that it pays to try a bunch of different things sometime. 
How do you like that Faitalpro? I've looked at it several times...
Those drivers _should_  play nicely together.

rjj45

Oh, and build your own jumpers with heavy duty alligator clips, and solder 18 or 20 gauge stranded wire to the clips and double up the clips on your test board if there is any doubt of the connection. I learned the hard way on my last build. Poor connections can cause bad measurements. 
Typically sibilance is caused by some peaking 5-7k. If the woofer is way down in that range, it must be the tweeter.
Sibilance really annoys the heck out of me, and so I have several "torture" tracks of female vocals that are prone to show any problems.  Good luck - you'll get it.

Kornbread

rjj45 said:

How do you like that Faitalpro? I've looked at it several times...
Those drivers _should_  play nicely together.

    I'm hoping to make it to an event this year so maybe you can hear for yourself.  Work is a big question as we employed 1100 in 2012 and now employ 240 with those remaining working 6+ days a week.  I'm keeping an eye open for new opportunities.  Hate to start all over and loose 4 weeks of vacation, or go to night shifts, but this place reminds me of a turd circling the toilet bowl.               

Nicholas_23

Kornbread said:

rjj45 said:

How do you like that Faitalpro? I've looked at it several times...
Those drivers _should_  play nicely together.

    I'm hoping to make it to an event this year so maybe you can hear for yourself.  Work is a big question as we employed 1100 in 2012 and now employ 240 with those remaining working 6+ days a week.  I'm keeping an eye open for new opportunities.  Hate to start all over and loose 4 weeks of vacation, or go to night shifts, but this place reminds me of a turd circling the toilet bowl.               

Oof , tough call on the job . I would try and be proactive vs reactive if something happens. Been there and it's no fun . If you're single it can make things difficult for sure. 

rjj45

Kornbread said:

rjj45 said:

How do you like that Faitalpro? I've looked at it several times...
Those drivers _should_  play nicely together.

    I'm hoping to make it to an event this year so maybe you can hear for yourself.  Work is a big question as we employed 1100 in 2012 and now employ 240 with those remaining working 6+ days a week.  I'm keeping an eye open for new opportunities.  Hate to start all over and loose 4 weeks of vacation, or go to night shifts, but this place reminds me of a turd circling the toilet bowl.               

Eesh, been there, had that done to me. Nick is right - be proactive, sorry to say, don't trust your boss. They will look you in the eye and smile right up to the time HR comes to your desk. Good luck and God bless, bro.

Kornbread

On the fatial pro; for the price it should come with a cast basket.  Like the sd215, the frame is stamped and not that strong.  Personally, I think it is a smooth sounding driver that may not be the most revealing.  It digs way deeper than its tiny size might suggest.

On the job front.  I have been approached by both the plant engineering manager (twice) and the plant quality manager for positions that are "going to open up in the next couple of weeks" only to have this never happen.  The hr manager has stated her interest in my role in a technical position at the plant.  Funny how they say this when another manufacturing plant in the neighboring town is hiring, but never follow through once the other plants have filled their positions.  In the past 2 weeks, they have terminated an engineer, maintenance supervisor, and weld tech.  Engineers are hard to come by and if they are cutting engineers things are getting serious.  Anyhow, it is a union plant and I'm just a shop floor dummy with 30+ years of seniority.  The plant will have to close its doors to get me, but I spent the last 6+ years going to school to better myself, not remain in the same dead-end job.  Sob story, I  know.  Got some important decisions to make ...              

rjj45

re: sob story?
Not at all, bro. Stuff happens, 'specially w jobs/careers and corporations.
I would always tell my peers, study hard, get your IT certifications - few listened to me, 
and they paid the price the next time the HR grim reaper came around.
Going back to school is hard, but can really open new paths. (My personal experience)
The flip side of all the layoffs is that some people stay around until they lock the doors,
and that can be a very long time. 

tajanes

so, my 2 cents, looks very good, possibly consider a very shallow, but broad, notch filter centered around 7kHz... 
and, gotta ask, is that a golden doodle?

Kornbread

It's been revised a 'few' times since then. As I posted in another thread, I'm concerned my hearing may be an issue with detecting that rising tail above 12k..  May have to do something with it, but to my ears the crossover is pretty close to getting soldered and potted.  You guys are probably looking at this and thinking 'I could do this crossover with 3 components', well ... not me.  Maybe someday, but not today. 

Notched the tweeter resonance and cut some breakup? from the woofer around 6k. that seemed to muck up the overall sound.         

R3 controls the trough from ~1.3-4k.  R2 controls the overall tweeter level.  R1 controls the cut at ~7k.  For the most part these can be adjusted, within a certain range, without messing up the phase too much and they make a huge difference in the sound.  Been playing around with them measuring/listening/adjusting on the go.  Surprisingly, a change in the sim closely correlates to the actual measured response.  Yea, I'm doing something right!  

Got the ebay soldering/hot air station in yesterday (I know, you guys told me to get the better stuff but I need the hot air gun too).  Thanks to the mail woman for leaving it sitting in the rain unprotected.  Going to un-hotglue some of my previous mistakes and start soldering/potting this one leaving R3, R2, and R1held in by jumpers.  Those resistors 'should' provide enough flexibility to make adjustments as listening time increases.

Starting to like this little 5fe120, it digs pretty deep and holds its composure very well at spl levels way above where it should be playing ... and that's being on an edm kick as of late.  Dubbacle>Gengar is a decent low-end test track and at these volumes, it has to be hitting its xmax, but if it does, it does so gracefully.  No harsh rat-a-tat machine gun fire stuff like when the rs225 bottoms.  Smooth also comes to mind, maybe not the most revealing mids, but smooth. 

                         

Kornbread

Chocolate labradoodle. 

Her eyes are always watching my every move.  This is one of those puppy pics where you can see just how important you are in your dog's life.  This tiny and innocent creature is looking up to a giant for everything she will ever need in life offering in repayment her faithful undying love.            

Tom_S

I've been working with the same tweeter again this past week. Are C6 & L6 for impedance compensation? You may be able to replace those with a simple parallel shunt resistor before the xt25. You can change the overall shape of your response by juggling that and your first resistor. It took me months, but I finally got something smooth(ish) where phase lined up well enough. Is it perfect - no, but it sounds pretty good.

* Wow - I just now removed a notch filter simply by lowering R4 from 10 to 5.6 ohms. So thank you for causing me to play around a little more this morning! We'll see if I can hear the difference. 

rjj45

My last build I worked on the crossover for months! Had at least 3 times where I deemed it "final" and ordered premium components, only to find that the next measurement was off, and my crossover test board jumpers had screwed me. Most of the versions tested and sounded good, but I kept going back to basics and reworking things. Got a good bit simper along the way. One day it "came alive", and I knew that was the keeper.
Don't know how some guys can do a sim on paper, build it w those values, and it's perfect. Takes me a while.

Tom_S

I'm also working on the Sim vs. Reality issue. There are just too many ways to screw it up!   

Kornbread

C6 and L6 are for the notch at tweeter resonance.  For some reason, playing around with this tweeter in the active (minidsp) 'not a khanspires', this tweeter sounds more refined with a steep crossover leading me to wonder if its resonance peak might be adding something to the picture.  Otherwise, for whatever reason, it can sound kind of clacky.   

That plateau from 2-3.5k bothered me.  Wish the 5fe120 rolled off as smoothly with only two components.  

  

Wolf

The resonance on the XT can lead to very nasal male vocals. If using a xover anywhere near 2k, you really should comp it.

...as to the sim vs reality, usually I am using real impedances as they are easier to get around and measure than the FR. Rarely do I do a sim without real impedance data. Once you measure the impedance and T/S, the box and modeled box impedances are likely not going to be faulty.
Beyond that- some people choose to not voice their xovers, or do not understand the process and/or it means very little to them. Some newer builders do not have a stockpile of parts (like say JR, JohnH, Chuck, or myself) and choose to not build a stockpile so they go with what they get. Measurements being for one point in space, and likely not taking off-axis or room (gated) into the picture can have a large effect on performance/results to where it should be adjusted. Some people just don't know that.

Just because it measures good does not mean it sounds good.

Kornbread

@rjj45 agreed.  The last sim posted was 7_3.1451.  It is the 7th complete crossover makeover.  The second digit indicates a major revision to that basic crossover, the 3rd digit a lesser modification, and so one.  Yes, I envy those guys that just sim stuff up overnight and it sounds great right out of the box.  I was so sure that a previous crossover for this speaker was going to sound good that it was soldered and potted only to have to figure out a way to un-glue hotglue to save the parts.  Thank goodness I can roll my own inductors as needed.                   

Wolf

  for rolling your own!
By potting do you mean gluing the parts down?

Kornbread

Wolf said:

  for rolling your own!
By potting do you mean gluing the parts down?

Yes, guess it's not really 'potting' but sounds better than 'hot-gluing parts into place'.  That stuff is a lot stronger than it looks.  Already destroyed a few components trying to remove it and going after it with a hot air gun next. 

With the rising response of 12k+ being an issue; the xt25 rolls off fast to the sides.  Is my thinking correct that because of this a little bit of toe might tame the rising response or that this fast horizontal roll-off might not load the room so the 12k+ may not be problematic?

Working on the cap/resistor stockpile.   Ok, never-mind, just seen ugly woofer's toolbox full-o-parts and a few plano boxes pales in comparison.                                

Wolf

The XT25 is not a tweeter with great off-axis response, and this is due to the type of tweeter it is; a ring-radiator. That is the reason for it's off-axis rolloff.

rjj45

Kornbread said:

@rjj45 agreed.  The last sim posted was 7_3.1451.  It is the 7th complete crossover makeover.  The second digit indicates a major revision to that basic crossover, the 3rd digit a lesser modification, and so one.  Yes, I envy those guys that just sim stuff up overnight and it sounds great right out of the box.  I was so sure that a previous crossover for this speaker was going to sound good that it was soldered and potted only to have to figure out a way to un-glue hotglue to save the parts.  Thank goodness I can roll my own inductors as needed.                   

I will usually do a crossover sim or two, and then possibly revisit a few times during "tuning", but my crossover flow is mostly "measure / modify / repeat". A big problem is keeping track of exactly what values produced which curve. Sometimes I print out a xSim crossover graphic and annotate it by hand, and sometime do the same for an Omnimic curve. But usually, I save the intermediate work files with as much info as I can, to wit: