A post of this picture of Tysen elicited a large number of requests for more information. So in aid of saving me from answering individual questions multiple times, let;s see if we can cover it here.
Tysen (named after our godson in the picture) is a Fostex FF85K & a CSS SDX7 (in both cases my fully modified version).
I chose sealed for the SDX7 because it works well sealed and with a bit of EQ can reach into the 25-30 Hz range in room. Tysen is 17 litres, but anywhere 14-21 litres net (including effect of stuffing).
The FF85KeN is in an aperiodically damped 10:1 taper TL. I wanted to tune the TL as low as i could withing the limited space i had available, Damping starts out really light getting increasingly dense towards the terminus (i call this arrangement a midTL).
Because of the efficiency of the SDX7 is lower than needed for a passive XO, and because it is so much easier, the system is active (biamped).
Chris & i are very pleased with how these turned out, and everyone who has heard them has been favourably impressed. First public exposure was on the Sunday of last years' VI diyFEST.
dave
Tysen (named after our godson in the picture) is a Fostex FF85K & a CSS SDX7 (in both cases my fully modified version).
I chose sealed for the SDX7 because it works well sealed and with a bit of EQ can reach into the 25-30 Hz range in room. Tysen is 17 litres, but anywhere 14-21 litres net (including effect of stuffing).
The FF85KeN is in an aperiodically damped 10:1 taper TL. I wanted to tune the TL as low as i could withing the limited space i had available, Damping starts out really light getting increasingly dense towards the terminus (i call this arrangement a midTL).
Because of the efficiency of the SDX7 is lower than needed for a passive XO, and because it is so much easier, the system is active (biamped).
Chris & i are very pleased with how these turned out, and everyone who has heard them has been favourably impressed. First public exposure was on the Sunday of last years' VI diyFEST.
dave
How we got there..
Tysen was not formed out of whole cloth.
After successfully hatching the mFonken, we still wanted something smaller. This led to the uFonken with the FF85. When we started this we knew that they would often get used with a woofer, and concurrently ordered a Keiga 2.1 amp and built a 14 litre sealed SDX7 woofer to go with the uFonken and hold the Keiga amp.
After listening to this for awhile* we decided that a pair of woofers that would act as stands.
*(not quite as planned -- the satelitte amps in the Keiga were prettry poor sounding, so we fudged that -- turns out that there is a compression cicrcuit in the HF amp chain that when removed improve things immensly)
This is what we started with.
I borrowed DCX2496, set upo the old G3 Lombard PowerBook with measuring mic/Fuzzmeasure and played with all sorts of XO arrangements.
With the DCX the only one happy with the sound was the measuring mic, but i learned a lot about XOing the combo -- for instance, as pictured the woofer is upside down -- i got quite severe combing until i fliipped it do the woofer was close to the mid-tweeter.
Attached is one of the FR curves. The limits at the bottom may not be the speaker but the Lombard's analog input and the roll-off at the top is from too long a mic cable. Althou a bit saddle shaped it is +/- 2dB from <30 Hz out to beyond 15k. There is a small amount of lift set at 25 Hz.
In the end i choose a 1st order XO on the woofer & 2nd order Bessel on the midtweeter at about 350 Hz. The low order XO on the woofer helps fill in the baffle-step loss. In this case the natural rolloff of the SDX7 above 1k doesn't hurt either.
A close approximation of this can be implemented as a PLLXO (Passive Line Level XO http://t-linespeakers.org/tech/filters/passiveHLxo.html).
dave
Tysen was not formed out of whole cloth.
After successfully hatching the mFonken, we still wanted something smaller. This led to the uFonken with the FF85. When we started this we knew that they would often get used with a woofer, and concurrently ordered a Keiga 2.1 amp and built a 14 litre sealed SDX7 woofer to go with the uFonken and hold the Keiga amp.
After listening to this for awhile* we decided that a pair of woofers that would act as stands.
*(not quite as planned -- the satelitte amps in the Keiga were prettry poor sounding, so we fudged that -- turns out that there is a compression cicrcuit in the HF amp chain that when removed improve things immensly)
This is what we started with.
I borrowed DCX2496, set upo the old G3 Lombard PowerBook with measuring mic/Fuzzmeasure and played with all sorts of XO arrangements.
With the DCX the only one happy with the sound was the measuring mic, but i learned a lot about XOing the combo -- for instance, as pictured the woofer is upside down -- i got quite severe combing until i fliipped it do the woofer was close to the mid-tweeter.
Attached is one of the FR curves. The limits at the bottom may not be the speaker but the Lombard's analog input and the roll-off at the top is from too long a mic cable. Althou a bit saddle shaped it is +/- 2dB from <30 Hz out to beyond 15k. There is a small amount of lift set at 25 Hz.
In the end i choose a 1st order XO on the woofer & 2nd order Bessel on the midtweeter at about 350 Hz. The low order XO on the woofer helps fill in the baffle-step loss. In this case the natural rolloff of the SDX7 above 1k doesn't hurt either.
A close approximation of this can be implemented as a PLLXO (Passive Line Level XO http://t-linespeakers.org/tech/filters/passiveHLxo.html).
dave
Attachments
The PLLXO
The parts box yielded a PLLXO at ~ 333 Hz. 0,1, 0.01 uF caps and 10 & 100k Rs (IIRC). I built it into an old Macintosh 800k floppy disk case.
This would actually be best built into the amps.
I thot i'd done a map of the XO, but i've not been able to find it... i may only have sketched it on paper.
dave
The parts box yielded a PLLXO at ~ 333 Hz. 0,1, 0.01 uF caps and 10 & 100k Rs (IIRC). I built it into an old Macintosh 800k floppy disk case.
This would actually be best built into the amps.
I thot i'd done a map of the XO, but i've not been able to find it... i may only have sketched it on paper.
dave
Attachments
A buffered PLLXO
As set up the PLLXO requires an amp with a 100k input. In aid of more versatility we did a fully biffered version. We put an OpAmp buffer at the in and the out of each filter. We choose opamps so that we could just use 2 9V batteries as a supply. Daniel did a real tidy job, and performabce was pretty good. Not quite as transparent as the unbuffered one on a system that can hack the PLLXO impedance requirements, but better on systems that can't.
To use a PLLXO the HF amp really should have 50k+ impedance, and the preamp needs to be able to drive a 10k load, The 2nd order hipass is the most restrictive bit, the 1st order lowpass makes that work almost anywhere.
Depending on the amps & preamps used you can get probably away with fewer buffers.
I always figured that the Firstwatt B1 would be a good platform for this XO, now i see that Nelson has had the same idea. The B2 Biamp (2nd paragraph coming soon http://www.firstwatt.com/watts_new.htm)
dave
As set up the PLLXO requires an amp with a 100k input. In aid of more versatility we did a fully biffered version. We put an OpAmp buffer at the in and the out of each filter. We choose opamps so that we could just use 2 9V batteries as a supply. Daniel did a real tidy job, and performabce was pretty good. Not quite as transparent as the unbuffered one on a system that can hack the PLLXO impedance requirements, but better on systems that can't.
To use a PLLXO the HF amp really should have 50k+ impedance, and the preamp needs to be able to drive a 10k load, The 2nd order hipass is the most restrictive bit, the 1st order lowpass makes that work almost anywhere.
Depending on the amps & preamps used you can get probably away with fewer buffers.
I always figured that the Firstwatt B1 would be a good platform for this XO, now i see that Nelson has had the same idea. The B2 Biamp (2nd paragraph coming soon http://www.firstwatt.com/watts_new.htm)
dave
Attachments
Trapezoidal Tysen
Chris made the suggestion that instead of propping the front up to get the right angle of fire on the FF85, we build a trapezoid similar to Scott Dunn's bipolar WR126.
This ended up working really well and turned out, IMO, fantastic from an aesthetic POV.
Note that instead of having an additional piece inside the box to support a woofer rebate, we surface mounted the woofer and later added an additional layer of 12mm to facilitate the rebate. Turns out this was a good thing, with the energy generated by the SDX7 the stethescope turned up more cabinet ringing than we like with just the 12mm walls. The extra layer also allows a larger bevel on the front.
dave
same caveat on these plans
Chris made the suggestion that instead of propping the front up to get the right angle of fire on the FF85, we build a trapezoid similar to Scott Dunn's bipolar WR126.
This ended up working really well and turned out, IMO, fantastic from an aesthetic POV.
Note that instead of having an additional piece inside the box to support a woofer rebate, we surface mounted the woofer and later added an additional layer of 12mm to facilitate the rebate. Turns out this was a good thing, with the energy generated by the SDX7 the stethescope turned up more cabinet ringing than we like with just the 12mm walls. The extra layer also allows a larger bevel on the front.
dave
same caveat on these plans
Tysen as an example
Tysen should be considered as an instance of a whole family of designs.
With the large number of small FRs now available, and a wealth of midbass drivers (most often seen in the ubiquitous midbass+dome tweeter 2-way) there is a large selection of components suitable for systems with a small FR coupled with a helper bass unit.
Tysen has inspired development of an FF85/SilverFlute W14 MTM. We cjhose this combo as it will allow a passive XO. We also have drivers to do silimar with the Alpair 5 & 6.
I bet a lot of you have bits laying around just looking for a project. I can dig out 4 different TB 3" (one set to be used with a pair of vinatge Coral 8" woofers/side), some Jordan J6T, and a set of FE83A (and that is nor considering larger mid-tweeters -- come to think of it Fonken + FonkenWoof falls into this). Some OEM Vifa P17 (for PSB), a set of Jordan JX150, some Foster 12s that are good well up above 1k... & i know i can find more...
I still have more to post, and i'm sure there are questions to be answered, but i've got to go put some gloss on some FE167eN before i go to bed.
dave
Tysen should be considered as an instance of a whole family of designs.
With the large number of small FRs now available, and a wealth of midbass drivers (most often seen in the ubiquitous midbass+dome tweeter 2-way) there is a large selection of components suitable for systems with a small FR coupled with a helper bass unit.
Tysen has inspired development of an FF85/SilverFlute W14 MTM. We cjhose this combo as it will allow a passive XO. We also have drivers to do silimar with the Alpair 5 & 6.
I bet a lot of you have bits laying around just looking for a project. I can dig out 4 different TB 3" (one set to be used with a pair of vinatge Coral 8" woofers/side), some Jordan J6T, and a set of FE83A (and that is nor considering larger mid-tweeters -- come to think of it Fonken + FonkenWoof falls into this). Some OEM Vifa P17 (for PSB), a set of Jordan JX150, some Foster 12s that are good well up above 1k... & i know i can find more...
I still have more to post, and i'm sure there are questions to be answered, but i've got to go put some gloss on some FE167eN before i go to bed.
dave
A couple of quick builder's notes-
With careful sizing of the rebated woofer cut-outs, you can get a very tidy friction fit of metal grilles. In our case from surplus recycled Monsoon computer speaker sub woofer boxes, but probably any number of car stereo installers would have spares that could fit.
The slanted version will definitely need some type of extended base plinth - stone or corian would be great
With careful sizing of the rebated woofer cut-outs, you can get a very tidy friction fit of metal grilles. In our case from surplus recycled Monsoon computer speaker sub woofer boxes, but probably any number of car stereo installers would have spares that could fit.
The slanted version will definitely need some type of extended base plinth - stone or corian would be great
midTL
The goal of the midTL is to attempt to absorb as much of the backwave as possible and to flatten the impedance curve. As can be seen by the attached impedance curve, i have mostly achieved that. The red curve is the FF85KeN in free-air, the blue curve in the Tysen midTL. If you imagine the blue curve as a typical double humped vented box curve, you can see that the minima is at about the Fs of 130 Hz, and that the lower hump is almost completely gone, and we have only a small resisdual of the lower hump.
This also shows why a passive XO can be difficult, but certainly not as difficult as the free-air curve.
If i ever get those Tysen back i'll measure the woofer impedance (JesseG borrowed them, took them up to Alert Bay with the "threat" that i may never see them again
)
dave
The goal of the midTL is to attempt to absorb as much of the backwave as possible and to flatten the impedance curve. As can be seen by the attached impedance curve, i have mostly achieved that. The red curve is the FF85KeN in free-air, the blue curve in the Tysen midTL. If you imagine the blue curve as a typical double humped vented box curve, you can see that the minima is at about the Fs of 130 Hz, and that the lower hump is almost completely gone, and we have only a small resisdual of the lower hump.
This also shows why a passive XO can be difficult, but certainly not as difficult as the free-air curve.
If i ever get those Tysen back i'll measure the woofer impedance (JesseG borrowed them, took them up to Alert Bay with the "threat" that i may never see them again
)dave
Last edited:
Hi
I think I would like to try something like those ie a small full ranger as a mid/tweeter and a helper woofer.
But then it mostly would be a tweek fest with room placement and setting amp levels crossed highish at 333 Hz with the side firing woofer? I'm sure it could be sweet once dialed in tho.
Any tips on a current Parts Express/Madisound full ranger driver to go with a sealed Morel 8" or ported peerless 10" woofer. TIA
I think I would like to try something like those ie a small full ranger as a mid/tweeter and a helper woofer.
But then it mostly would be a tweek fest with room placement and setting amp levels crossed highish at 333 Hz with the side firing woofer? I'm sure it could be sweet once dialed in tho.
Any tips on a current Parts Express/Madisound full ranger driver to go with a sealed Morel 8" or ported peerless 10" woofer. TIA
infinia said:
Did you have a particular FR driver in mind? FWIW, the choice of FF85K for this project was serendipitous, but not happenstance. This is one of the more under-rated little gems in the Fostex line-up.
For those with another taste (i.e. non-paper) , the Mark Audio CHR70 or Alpair 5 would be great candidates for this application, and no doubt any number of Tang Band drivers as well.
As for the tweek-fest, isn't that why we're DIYers? - any idiot could just buy a B0$e lifestyle system and be done with it
Member
Joined 2009
Paid Member
Dave,
I'm glad you've started this thread, long overdue if you don't mind me saying.
The Tysen was the one that got me thinking CSS drivers and some low end assistance for my MoonOnkens.
At some point this year the 'itch' to build speakers again will likely pull me back. I am leaning towards the Tysen sytle approach. But I want physically smaller boxes. I'll be keeping an eye on this thread to see how these things develop
I'm glad you've started this thread, long overdue if you don't mind me saying.
The Tysen was the one that got me thinking CSS drivers and some low end assistance for my MoonOnkens.
At some point this year the 'itch' to build speakers again will likely pull me back. I am leaning towards the Tysen sytle approach. But I want physically smaller boxes. I'll be keeping an eye on this thread to see how these things develop
chrisb said:
No , just something that can do high SPL's ie at least 105dB/m steady and w/o sounding compressed when pushed with transients of 10 -15 dB more.
I guess what I'm really asking is...
Beside a different/sweeter SQ than the ubiquitous 7" midwoof and dome two way's could these things hang when pushed hard.
Based on the larger helper woofers listed in my above post.
Re: The PLLXO
I'd love to know the values for that non-buffered PLLXO, I'm thinking I may try this with my FE126E/W15 FAST. The current cross is 200Hz and even though it sounds unreal the crossover should be a tad higher. I'd love to build this curcuit into my interconnects to bi-amp with my LM3875 GC's, seems like it could easily fit into the braid of some Techflex, Hmmmmmm...... You have me thinking now Dave.... Very Dangerous it seems as of late for me to think at all LOL....
If you can think of the values I'd love to hear them Dave, Peace and keep well.
I'd love to know the values for that non-buffered PLLXO, I'm thinking I may try this with my FE126E/W15 FAST. The current cross is 200Hz and even though it sounds unreal the crossover should be a tad higher. I'd love to build this curcuit into my interconnects to bi-amp with my LM3875 GC's, seems like it could easily fit into the braid of some Techflex, Hmmmmmm...... You have me thinking now Dave.... Very Dangerous it seems as of late for me to think at all LOL....
If you can think of the values I'd love to hear them Dave, Peace and keep well.
Re: Re: The PLLXO
The values i used would probably work for the bass amp (10k, 0.1uF IIRC), but you are going to have to get creative to do 2nd order on the hi-pass as the input R on a gain clone is (usually) too low. what you will probably have to do is make the high pass gainclone into a 2nd order powered filter -- shrink the input C to move that pole up to where it needs to be, and then another C in the feedback loop for the other pole. Time to dig out the active filter cookbook.
dave
speakrsrfun said:
The values i used would probably work for the bass amp (10k, 0.1uF IIRC), but you are going to have to get creative to do 2nd order on the hi-pass as the input R on a gain clone is (usually) too low. what you will probably have to do is make the high pass gainclone into a 2nd order powered filter -- shrink the input C to move that pole up to where it needs to be, and then another C in the feedback loop for the other pole. Time to dig out the active filter cookbook.
dave
