Fostex T90A Supertweeter Filter

Dedicated to those large boxes at one end of the room
User avatar
Nick
Site Admin
Posts: 15706
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#16 Re: Fostex T90A Supertweeter Filter

Post by Nick »

Like everything else when measuring, you first need to verify if your measurement equipment make is doing what you expect. I would check what the noise floor in the room is. Then verify that you can detect single band frequencies (or even better filtered noise bands).

If you can hear output from the super tweeter at 2kHz. Check its actually from the supertweeter and not your output transformers. Does it stop if you cover up the supertweeter, or disconnect it?
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
Scottmoose
Needs to get out more
Posts: 1802
Joined: Sat Sep 08, 2007 11:03 am
Contact:

#17 Re: Fostex T90A Supertweeter Filter

Post by Scottmoose »

simon wrote:I was really surprised to hear any output at 2k, let alone as much as there was. This just isn't adding up to me.

Anyway, I want to try 1uF just to see if I can produce a roll off. Assuming I can do that then I can play a little more.
A single series capacitor is simply a 1st order electrical high pass. The tweeter is down -3dB at the nominal crossover point (in this case corner frequency rather than an XO point per se) and output only decreases by -6dB/octave relative to the unfiltered response assuming a stable impedance. Some general points / thoughts, for what they're worth:

-Driver excursion will continue to increase below the XO frequency with 1st order high pass filters. You need at least a 2nd order electrical high pass to prevent that happening. Power-handling is usually poor, which isn't directly too much of an issue in this case (relatively high efficiency drivers & low power amps) but distortion levels increase with it.

-The above takes [even] greater significance if you get a rising impedance around the tweeter's resonant frequency or elsewhere as it's going to start playing louder (hence them being relatively demanding of the tweeter in terms of its power handling and distortion behaviour. Again, circumstance dependent, & less of an issue here, but still worth noting).

-The Fostex plots of the era the T90a was launched in are all 3rd octave smoothed (minimum), and Fostex have been known to be a trifle 'optimistic' in their sensitivity figures. Their impedance data is also on a logarithmic scale so it can be mashed into the same graph, and is therefore visually also somewhat flattering. Don't get me wrong -I like the T90a, but I'd be wary of taking any bald figures at face value.

-A carefully implemented low Q 2nd order high pass network can give you a pseudo-1st order rolloff for around the first octave of the stopband, transitioning to a steeper slope below that -you also get more electrical protection & limit excursion below the XO frequency. Don't expect the overall response to be the most linear on the planet when supporting unfiltered widebanders though! Whether that's such a big deal or not -YMMV.
'"That'll do," comes the cry of the perfectionist down the ages.' (James May The Reassembler)
Website www.wodendesign.com
Community sites www.frugal-horn.com & www.frugal-phile.com
simon
Thermionic Monk Status
Posts: 5600
Joined: Thu May 24, 2007 11:22 am
Location: People's Republic of South Yorkshire

#18 Re: Fostex T90A Supertweeter Filter

Post by simon »

Who knew a simple filter for a supertwatter would be so complicated :-D.

Thanks for the offer of measurement, sounds good to me Chris and Simon (sorry for the dreadful pun!). I don't have a Saturday free till the middle on November, I'll see whether there's a Sunday available.

Sunny Donny so not far from you Chris if you're still in Sheff? Bit more of a trek from Saltaire, if you're still there Simon? Weekends okay for you two? Even a second opinion/reality check would be useful.

Probably won't get chance to play with the filters till the weekend now. Thanks for the advice Nick and Scott. Need to digest some of that.
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#19 Re: Fostex T90A Supertweeter Filter

Post by IslandPink »

Scottmoose wrote: -A carefully implemented low Q 2nd order high pass network can give you a pseudo-1st order rolloff for around the first octave of the stopband, transitioning to a steeper slope below that -you also get more electrical protection & limit excursion below the XO frequency. Don't expect the overall response to be the most linear on the planet when supporting unfiltered widebanders though! Whether that's such a big deal or not -YMMV.
This sounds good, it was where ( approximately ) I ended up with the 'Bazzilla' variants with 208Ez and Aurum G3 on OB.
"Once you find out ... the Circumstances ; then you can go out"
simon
Thermionic Monk Status
Posts: 5600
Joined: Thu May 24, 2007 11:22 am
Location: People's Republic of South Yorkshire

#20 Re: Fostex T90A Supertweeter Filter

Post by simon »

Okay, very quick play before tea. Sound is definitely coming from the supertweeter - it's quite odd listening to music on supertweeter alone. Also checked simply that the signal being generated is as expected. Even used an alternative spectrum analyser - still free (cough). This one has a log scale, but it still looks pretty similar to me.
Image

The increase in background noise in the lower frequencies below 2k Hz I think is probably due to the phone generating background noise (whadaya want for free!)
chris661
Shed dweller
Posts: 2556
Joined: Mon Feb 22, 2010 7:29 am
Location: Sheffield

#21 Re: Fostex T90A Supertweeter Filter

Post by chris661 »

Got one of those too. Did quite well, but I never quite trusted it, having seen some of the curves online. The Beyerdynamic one comes with a unique graph, so I've got something calibrated.

Do you have a way of connecting the ECM8000 to your computer? - REW is certainly worth having if so.

Chris
SimonC
Old Hand
Posts: 545
Joined: Mon Sep 24, 2007 11:34 am
Location: West Yorkshire

#22 Re: Fostex T90A Supertweeter Filter

Post by SimonC »

simon wrote:Who knew a simple filter for a supertwatter would be so complicated :-D.
Thanks for the offer of measurement, sounds good to me Chris and Simon (sorry for the dreadful pun!). I don't have a Saturday free till the middle on November, I'll see whether there's a Sunday available.
Sunny Donny so not far from you Chris if you're still in Sheff? Bit more of a trek from Saltaire, if you're still there Simon? Weekends okay for you two? Even a second opinion/reality check would be useful.
Probably won't get chance to play with the filters till the weekend now. Thanks for the advice Nick and Scott. Need to digest some of that.
Yep, still up in Saltaire. Its not a big deal heading down your way, especially as my extended family live in Nottingham so I could combine you all into one weekend.
SimonC
Old Hand
Posts: 545
Joined: Mon Sep 24, 2007 11:34 am
Location: West Yorkshire

#23 Re: Fostex T90A Supertweeter Filter

Post by SimonC »

chris661 wrote:Got one of those too. Did quite well, but I never quite trusted it, having seen some of the curves online. The Beyerdynamic one comes with a unique graph, so I've got something calibrated.
Do you have a way of connecting the ECM8000 to your computer? - REW is certainly worth having if so.
Chris
Hi Chris,
Yes, I've got an external Behinger USB mic amp/phantom supply. I can't remember the exact model off the top of my head, but its the same one that Ed used for his measurement setup.
Si
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#24 Re: Fostex T90A Supertweeter Filter

Post by IslandPink »

Still doesn't look right to me. That response curve is awful between 2k and 20k - up & down 20db, with loads of spikes & noise.
Does the T90A actually sound nice ( in a tinkly sort of way ) when played on its own ? It should sound clean and with some tone on things like cymbals, bells and metal guitar strings. If it sounds good and not aggressive then I can't see the curve above representing real life.

I will try & dig out my measurements of the FE208Ez on OB too, for additional checking. Was on my speaker thread somewhere 2/3 years ago.
"Once you find out ... the Circumstances ; then you can go out"
User avatar
Nick
Site Admin
Posts: 15706
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#25 Re: Fostex T90A Supertweeter Filter

Post by Nick »

I don't understand whats going on at 8kHz, where has that notch come from. I would be tempted to try a single 12kHz tone, to make sure you get a single peak.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#26 Re: Fostex T90A Supertweeter Filter

Post by IslandPink »

Some stuff about crossover for Aurum G3 ribbon and Fostex FE208Ez here :
http://www.audio-talk.co.uk/phpBB3/view ... la#p121175
( Page 90 of my speaker thread, if that link doesn't work )
"Once you find out ... the Circumstances ; then you can go out"
simon
Thermionic Monk Status
Posts: 5600
Joined: Thu May 24, 2007 11:22 am
Location: People's Republic of South Yorkshire

#27 Re: Fostex T90A Supertweeter Filter

Post by simon »

Made a few more measurements this evening. I'm not sure what they all mean, the more I do the more confusing they seem. Alone the T90A sounds fine, for a supertweeter, and in combination with the 208 sounds pretty good I have to say. I'm probably chasing unnecessarily, but it is interesting.

So, this evening I switched to generating tones and sweeps from Audacity on the laptop. This first plot is the result of a series of single tones from the 208 and T90A, just to check I was reading what I should be reading. There were a few harmonics at some frequencies (only second and third that I could see), but the peak profile is interesting as it's up and down a little. The volumes were obviously different to my hearing too. So I guess either Audacity or the laptop aren't consistent, or the T90As aren't.

Image


Next image is also 208 & T90A with a sweep 20Hz to 20kHz

Image


And now the same sweep on just the T90As. The "drop out" between 6k and 8k is there again. Not sure why it's not there when the 208s are connected too.

Image


T90A alone with composite individual tones. The "drop out" between 6k and 8k isn't the same magnitude compared to what's happening above 8k. A little odd.

Image


So I tried running one of the T90As out of phase. I don't know whether this looks better or worse.

Image


Both supertweeters in phase, but with the Lpad removed. This should look something like the third image (with the "drop out" between 6k and 8k) but doesn't.

Image


Here I've replaced the 2.5uF cap with a 1uF. Still no Lpad. Maybe it rolls off just a little earlier? Maybe. And the frequency response is a little different again too.

Image


And here's the last one with the 208s added. So 208s & T90As with 1uF and no Lpad. This looks like the relative levels of the 208s and T90As are about right, but in practise listening to music there's too much T90A.

Image


It's interesting to me (at least) that the speakers sound a whole lot better than the plots. And I guess the dodgy nature of the plots probably comes down to the dodgy nature of the mic and analyser hardware.
User avatar
Ali Tait
Eternally single
Posts: 4373
Joined: Fri Jun 08, 2007 8:10 pm
Location: Galashiels

#28 Re: Fostex T90A Supertweeter Filter

Post by Ali Tait »

Something like this could be helpful Simon-

https://www.minidsp.com/products/acoust ... ent/umik-1

Should give good results with REW.
simon
Thermionic Monk Status
Posts: 5600
Joined: Thu May 24, 2007 11:22 am
Location: People's Republic of South Yorkshire

#29 Re: Fostex T90A Supertweeter Filter

Post by simon »

The bits for the second order crossover arrived a few weeks back. After a brief run-in I think it did sound better - a little more clarity perhaps. But that's what I'm expecting so maybe it's what I want to hear...

Anyway, Ed dropped me an email a couple of weeks ago offering me a go with his Clio system. Seemed too good an opportunity to pass up!

I met him this evening in a lay by off the motorway to pick it up, I felt like I ought to be passing him a brown paper envelope stuffed with notes :lol:. Might be a few days before I can play with it as I'm under, er, um, domestic pressures with higher priorities. Just hope I can get my head around it.

Ed sends his regards and says hello to all BTW. He lost his login a while back but does look in occasionally evidently. Twas good to see Ed again, all too briefly, another valued member to have fallen by the way (file along with Will and Max)
chris661
Shed dweller
Posts: 2556
Joined: Mon Feb 22, 2010 7:29 am
Location: Sheffield

#30 Re: Fostex T90A Supertweeter Filter

Post by chris661 »

simon wrote:Thanks for the offer of measurement, sounds good to me Chris and Simon (sorry for the dreadful pun!). I don't have a Saturday free till the middle on November, I'll see whether there's a Sunday available.

Sunny Donny so not far from you Chris if you're still in Sheff? Bit more of a trek from Saltaire, if you're still there Simon? Weekends okay for you two? Even a second opinion/reality check would be useful.
Hi Simon,
Sorry I missed this. Keep me informed, I'd be more than happy to come over and do some measurements. I'm still around Sheffield, yes.

Chris
Post Reply