Tysen Variation FE167E with CSS SDX7

[ljfont]

Originally Posted by ttsto

One thing bothers me, you combine a high sensitivity speaker like FE167E with a low sensitivity woofer...9db difference...how will this be compensated? by reducing the sensitivity of the high sensitive driver I presume...

This is intended for a bi-amped system with separate volume control for each speaker. Effective volume, balance and EQ with 4 dials. So efficiency does'nt really matter here too much. Just turn up the bass and/or turn down the midrange hi's.

I won't mess with the FE167E. I think it will be happy getting current from a NP Firstwatt F3. See http://www.firstwatt.com/downloads/cs-amps-speakers.pdf for power transconductance amps with full range drivers. A BSC circuit might not be needed either. A future update will be getting my FE167E eNaBled.

I built a DIY gainclone based on the LM3886 Chip which I will use for the SDX7. Gainclones have a somewhat high damping factor which should be good.

I'm currently working on a buffered PLLXO. Something similar to the NP B4.
See sketch in this thread Single Ended Active Crossovers for Open Baffle Speakers

Not sure if his design will be published publicly, but it is the perfect device for biamping with full range drivers. It's not yet for sale either.

I made a DIY B1 buffer, and it's sitting in front of the gainclone powering the FE167E's and it makes a HUGE difference in the sound when you have proper impedance matching, so a B4 type circuit is the way to go for me. Perhaps one running dual supply that is regulated to get rid of the bypass caps in the signal path.

The biggest risk of this design is placement of the speakers FE167E and SDX7 within this cabinet, I tried to physically spread out with the driver, port and sub. But it's really unknown to me how it will sound in the room. From simple listening tests with the current speaker build and just firing up the SDX7 from another amp with no crossovers it sounded very promising.

The hard part now is getting a buffered XO system to work with both the FR and Sub in this cabinet. Adjustable electronic XO and adjustable Q point with removable shelves gives me a bit of a tweak factor and many hours of listening this system to get it right. I estimate that it will take several months.

[ljfont]

After hearing these EL70's at CSS. I almost thought about replacing the FE167E with Dual EL70's wired series like Bob's at CSS.

Possible future addition is to have these interchangeable with the FE167E's in the same cabinet. (Sometime next year perhaps, wife permitting, yes I have a budget. )

One of the differences is that the stuffing will have to probably change dramatically for each speaker swap. There is a bit of ringing.

Need to think about a supra baffle/speaker caddie for each FR system. What are the choices of materials. Hardwood, softwood, MDF, non-wood?

I guess something more ridgid?

[chrisb]

Originally Posted by ljfont

Need to think about a supra baffle/speaker caddie for each FR system. What are the choices of materials. Hardwood, softwood, MDF, non-wood?

I guess something more ridgid?

yup, them EL70's are quite interesting little driver

if you're planning on swapping baffles, I think the choice of method to attach to main enclosure could ultimately be as important as the material for the baffles themselves.


I'd be inclined towards either hardwood cleats screwed to the perimeter of the baffle opening and threaded insert fittings, or 3/4" x 3/4" x1/4" aluminum angle, tapped for machine screws.


For the baffles, as they'd not be permanently glued to the enclosures, but rather sealed with weatherstripping, etc., you'd be free to play with solid woods, plywood, 1/4" or thicker aluminum plate - the sky's the limit. I've always wanted to find an excuse to try aluminum plate.

[ljfont]

Originally Posted by chrisb

yup, them EL70's are quite interesting little driver

if you're planning on swapping baffles, I think the choice of method to attach to main enclosure could ultimately be as important as the material for the baffles themselves.


I'd be inclined towards either hardwood cleats screwed to the perimeter of the baffle opening and threaded insert fittings, or 3/4" x 3/4" x1/4" aluminum angle, tapped for machine screws.


For the baffles, as they'd not be permanently glued to the enclosures, but rather sealed with weatherstripping, etc., you'd be free to play with solid woods, plywood, 1/4" or thicker aluminum plate - the sky's the limit. I've always wanted to find an excuse to try aluminum plate.

Thanks Chris.


I'd like to try different hardwoods. Give's my a chance to play around with the router a bit more.

I used T-Nuts to attach the FE167E, don't know If I can re-use those in the same location. We'll see.

[ljfont]

Circuit Description:
- Dual supply with JFET source followers
- Sallen Key circuit
- Linkwitz Riley 2nd order

The simulation comes out okay, the frequency range for the XO should suffice the output Impedance of this circuit. Calculated range is about 40Hz to 400Hz, without too much loading issue. If so, a "white follower" can be employed.


References:
http://www.borbelyaudio.com/adobe/ae699bor.pdf
Crossovers
ESP - The Linkwitz Transform Circuit
http://en.wikipedia.org/wiki/Sallen%...93Key_topology
The Art of Electronics, 2nd Edition, Horowitz and Hill, 1989

[ljfont]

Reducing the rails to +/- 9V, required reducing the source resistance a few ohms to keep the output DC offsets low.

Simulated Data:
Bias Id = 6.35mA

Source 1kHz sine
X0(Hz) THD(%) DC offset (uV)
40 0.000 14
120 0.003 25
200 0.009 86
400 0.032 324

Source 10kHz sine
X0(Hz) THD(%) DC offset (uV)
40 0.000 11
120 0.003 45
200 0.010 92
400 0.039 400


Slightly higher DC offsets in the higher frequency sim. But very small, uV.
THD increasing with fo.

I will try with different bias for Id (Need to see the curves on the datasheets) and "white follower" topology to see if the THD lowers. So loading is a little bit of a factor at the higher XO frequencies.

More info later.

[ljfont]

Lowered THD @ 400Hz. Calculated 0.001% with 1V peak input sine. Reduced the filter Capacitance to lower loading and distortion.

White follower not needed since it can drive this load okay
Cascode topology doesn't need to be called in since it's a lower frequency XO range.

Specs:
Solen Film and Foil Polypropelene Caps
Matched JFETs
Metal Film Resistors

Well, that's it for a while. Can't build anything till next month. Out of town.

[ljfont]

So, I've been living with these for awhile now, and they really sound great. However, there are probably some tweaks that can make them sound better.

Q: Which is preferred?
a) A BSC circuit
b) High Q HP filter for the FR

Q: Which is preferred?
a) A passive circuit at the speaker, which can effect the speakers response (quoting Linkwitz)
b) A passive buffered circuit which has caps and/or coils in the signal path?

Q: What relationship exists between stiffer damping at the terminus and cone excursion?

Thanks