• Please note that Audiokarma will be offline briefly for updates early on Monday morning, September 28th.

GAMMA 30 experiement

Hmm crossing this system at 4KHz is challenging, in some ways, it is hard to make it in a way that looks like an improvement over my previous filters.

I guess the way to make it is with a 6dB/oct filter on the woofer, and a 18dB/oct filter on the tweeter, otherwise the tweeter will go too low, it REALLY likes crossing at 5KHz with a 12dB/oct filter.

Yea I have to see what those phase plugs can do.
 
This filter compensates for woofer directivity without lowpass.

b64969cd_GAMMAmkIIIv1.03.png
 
Yes it seems 3 is a lucky number as the mkIII v1.03 crossover is the best so far.

It's directivity compensation is not as aggressive as the Audioscan filter or mkII v1.04, and it has no lowpass filter.

The highpass filter pushes the ribbons a little, not as much as the Audioscan filter though.

Edit: eh no I think this filter pushes the ribbons just as much as the Audioscan one, if not more, it doesn't sound as rough though.
 
Last edited:
And yet another filter:
a9fc7c6e_GAMMAmkIIIv1.04.png


This one sacrifices linearity at the optimal listening position, for nicer behavior across a wide range of angles.
To is done by using a gentle roll off slope on the tweeter, it should avoid humps or bumps as the crossover point slides around depending on listening angle.

Sounds fine, and isn't as hard on the ribbon as v1.03, more testing needed.

Edit: This filter sounds really good, it has been a very fruitful day, the lesson is, work WITH the problem, not against it.

Edit2: I here by name filter mkIII v1.04 "Liquidator Special Edition".

Edit3: loud volume tested, sounds fine.
I really like how this crossover works with speaker directivity without causing any gross errors, just a change in temperature.
 
Last edited:
The mkIII v1.04 filter was warm sounding compared to the DF's, v1.05 is tuned to sound brighter.

f2f487a6_GAMMAmkIIIv1.05.png


Edit: lol I'm getting it all mixed up, v1.03 and 1.05 are the same!
 
Last edited:
Introducing mkIV v1.00, this filter tries to do what the mkIII filter does, without pushing the ribbons so hard.

f5c0689c_GAMMAmkIVv1.00.png


Edit: I keep messing the version numbers up now, this filter was a dead end anyway.
 
Last edited:
This filter tries to attain the flattest frequency response at listening position.

Without the coil (mkIII v1.03) the speakers sound mid centric, and lean.

With the coil (mkIV v1.00) the speakers sound mostly flat, maybe the upper mids are a bit recessed (tone sweeps) but with music listening they sound fine so far.

I used the mkIII v1.03 for quite a while, I'll do the same for this one, maybe adjust the highpass filter if I find a reason to, or try a 0.6mH coil on the woofer.

This lowpass filter is far less detrimental to performance, then the 12dB/oct one (mkII).

More testing required :yes:

6a8605a9_GAMMAmkIVv1.00.png
 
Last edited:
Here is the frequency response plot (mkIV), it should be considered an approximation.
It correlates well with what I hear though, but keep in mind that directivity data is not included, nor can I simulate it at this point.
I think I get more bass then the graph would suggest.

Still it makes it much easier to see what's going on.

02499189_approx.png
 
Loud listening test has just been finished, I'm wrecked, in the best possible sense, these speakers really draw you in.

Anyway, they seem to handle loud volumes better with the mkIV filter.
Without a lowpass filter (mkIII), I get a "shimmering" effect when moving my head around while playing loud, on some recordings.

It's probably caused by high frequency phase noise, the GAMMA woofers keep making sound all the way to 10KHz.
But the mkIV filter solves that problem.

The directivity issue has been greatly diminished, I don't consider it a problem at this time.
The mkIV has a lot going for it, as long at the tweeters are a ear height and pointing towards me (as always with horn loaded ribbons), the speakers sound consistent across a fairly wide angle.

And lastly, a flat frequency response, it's easy to think of that as being a nice thing, but not too important as long as it sounds good.
Well, chances are, a flat speaker will sound better, especially at loud volumes, it seems frequency response inconsistencies, get massively over exposed when you approach the "comfortable" SPL limit.

I have decided to change my approach.
While making the mkIII filter, the goal was to see if I could make my speakers sound good, with the simplest possible crossover.
I consider that project complete, it IS possible, but I'd like to have a subwoofer if I where to use that filter, or maybe woofer mods could make it sound warmer.

The main advantage of that approach is that the woofer performs beautifully.
The downsides are a mid centric sound, and obvious directivity issues.

So the mkIII filter has one major advantage, but two downsides, and only of of those problems (mid-centricity) seems to be something I can solve (without crossover).

The mkIV filter has three advantages, all pretty big ones too, and only one downside (slightly reduced woofer speed).

This is my new approach, if I want to avoid getting stuck trying to satisfy a design principle that may have great potential benefits, but some rather big compromises, I have to use a more balanced decision process.

I have to weigh the advantages and disadvantages first, and then see how I can mix my design principles in to that (two way system, with minimal crossover).

I have gotten so far at this point that it's not a matter of making the speaker work, or sound good.
No, all I'm doing at this point is voicing the speaker to fit my taste, the mkIII v1.03 filter sounds just fine, but it needs a subwoofer or maybe a BR box.
I prefer a simpler solution.

Ramble ramble, back to listening!
 
I think that, for the final crossover network, I'll do two optimizations.

I'll put a ClarityCap MR in the highpass filter and use a symmetrical lowpass filter.
I'll put 0.4mH coils in series on both sides of the woofer, the end result is the same as using a single 0.8mH coil.
I can't explain the advantage of using a symmetrical crossover at this point but it seems there is one.
I think I'll see if I can make a symmetrical highpass filter as well.

Edit: no symmetrical highpass, I'll have to use a radically different network for that, with much higher values.
 
Last edited:
Should I?

Not now of cause, but as one of the last tweaks, along with recabling the things.
I'd have to stick to ClarityCap ESA, I have two 6.8uF already, I'd just need the OFC coils and two more ESA's.

5b5116b7_SymmRC.png


I don't think I can stop my self from making that, I have to make sure the mkIV v1.00 performs well in the long run first, if it does, then...

Edit: the tweeter has neg. polarity, unlike the diagram above.

Edit2: aaah, it's probably better to use a more simple highpass filter and go symmetrical for the woofers only.
2x 6.8uF on each tweeter seems a lot compared to 1x 3.3uF doesn't it?

Edit3: but it's sooo... symmetrical!!! :sigh:

Edit4: I should make this one.
05b459f7_OptRC.png
 
Last edited:
Lol look at this, I came across this on the Raven ribbon website:

kimon21.jpg


Those can't be anything other then GAMMA's.
 
The directivity of the woofer has changed, it's hard to say how exactly.
The overall sound seems to have changed too, but I'm not sure how.

To me it seems that phaseplugs are nice, but not a requirement if you decide to remove the dust cap.
I'm not going to remove the phaseplugs though, I haven't even attached in the best way yet, they are held in place with a bit of foam tape, so there are some cavities under them.

To make a few guesses, I think the woofers roll off more gently when heard at an angle.
Without phase plugs, the woofers roll off fairly steeply after 3KHz, when heard at an angle.
They don't really start getting directional at all until above 3KHz.
With the phase plugs it seems that the rolloff after 3KHz is more gentle, maybe.
I think maybe the frequency response above 2KHz is smoother, making the upper mids, sound recessed (not by much at all) but more noticeable then it was the other day.

But I don't really know, could just be that my ears are totally ready for tone sweeps today, so variations seem more pronounced.
Also, my amp has barely completed it's warmup.

Oh well, more listening.
 
Last edited:
Hmm, another idea of whats going on is that maybe disprsion has changed in the 1-3KHz region as well.

Normally the GAMMA woofer start getting directional above 1KHz, but not seriously so until 3KHz.
Without dustcap this is improved upon, and the woofers are barely directional up to 3KHz, a bit but nothing to worry much about.
I think the phase plugs might make it so that the woofers don't start getting directional at all, until 2-3KHz.

The result is slightly elevated amplitude in the 1-3KHz region, compared to before, and are more gentle rolloff.

Just did another tone sweep it sounded flat flat flat, maybe the 2.5-3.5Khz region is a bit recessed, I think that is the last remnant of dispersion problems.
When listening to music, it sounds fine at all angles, maybe a bit "essy" and mid centric, when looking straight in to the woofers, but that is not the intended listening angle.

I'm starting to notice an improvement in clarity, especially in the upper midrange.
I even notice it on tone sweeps, before phaseplugs, tone sweeps in the 2-4KHz region (inconsistent), could sound a bit "screamy", while it sounded smooth at other frequencies.

Now it just sounds smooth.

I made quite a few realizations as I wrote this, donno it that shows.
 
Ah I have been listening to speech for a while now, and I think what has changed is that, without phaseplugs, they sound more aggressive, with phaseplugs they sound more neutral.
 
Indeed the mkIV v1.00 filter is very good, but I wanted to try something else.

So I put the VLD on their side, for better vertical dispersion and more overlap between the drivers.

So far I think it has improved dispersion in the, guess what, vertical plane.

I have also put a 0,3mH coil in the highpass filter, seems to improve phase integration, without changing the frequency response much.
It also strains the ribbons a little less.

Ok as I write this I can hear that the ribbons need to be vertical, sound terribly edgy at an angle, they really have to be upright.
 
Hmm putting the drivers back to their "normal" positions didn't change anything, must be the coil causing phase problems.

Hmm now I have confirmed that I can't simulate the correct phase response at this point.

Gotta' revert, good, good!!
 
You must log in or register to reply here.
Back
Top Bottom