LC 600ohm xover

Design and use of Chokes for PS, anode and filament use.
Post Reply
radio_elasais
Posts: 11
Joined: Sun Jan 13, 2008 11:04 am

LC 600ohm xover

Post by radio_elasais »

Ciao!

First post here but I have been looking at Dave's forum since ages.

I prototypied a 600Ohm LCL xover I am driving with my 801A preamp and I currently use it for my biamped system. Having a line xover allows me to run my horns and onkens free and it is a huge improvement over the "traditional" caps/inductors nets we attach in front of the drivers. Pultec made it in the golden years, why shouldn't I do it now?

Selectable freqs 600, 750, 900 and 1100Hz, 18db/oct and I am waiting for some NOS langevin 600Ohm stepped attenuator to replace the linear pots I am using to balance the low/high.

The basic schematic is pretty simple (the value you see in the attached file are NOT updated ... that was an old spice sim). An input 600:600 tranny, a R to offer the pre a flatter load, and the LCL - CLC network. The last 600R is a tapped resistor or pot. And an output 600:600 tranny.

At the scope I found the -3dB points are pretty close to the theoretical. The low freq filter cuts a at a freq a bit higher than theoretical while the hi freqs a bit lower. That's fine. The pendence is not exactly 18dB but that is not really an issue. Please refer to the xls graph for experimental values. The components values and the theoretical cutting freqs follow

High Pass 1100Hz
C 0.32uF
L 33mH
Freq 1095Hz
Impedance 454Ohm

Low Pass 1100Hz
C 0.626uF
L 68mH
Freq 1090Hz
Impedance 466Ohm

High Pass 900Hz
C 0.47uF
L 33mH
Freq 903Hz
Impedance 375Ohm

Low Pass 900Hz
C 0.94uF
L 68mH
Freq 890Hz
Impedance 380Ohm

High Pass 750Hz
C 0.47uF
L 47mH
Freq 757Hz
Impedance 447Ohm

Low Pass 750Hz
C 1.096uF
L 82mH
Freq 751Hz
Impedance 386Ohm

High Pass 600Hz
C 0.47uF
L 82mH
Freq 573Hz
Impedance 590Ohm

Low Pass 600Hz
C 0.94uF
L 150mH
Freq theo 564Hz
Impedance 466Ohm

As you see the impedance of the filter is lower than 600Ohm. Actually I found that the lower is the better using the ear as measurint instruments. The drawback is that the preamp will suffer because of the lower load. 400-600Ohm seems to be a good compromise here. Going up with the impedance was not giving good results, the driver will be happier but the overall results was not as good as a 400Ohm filter. Don't ask me why.

Does anyone care to share his experience on the subject?

Ok, a compulsory pic of the prototype is attached. I am using tapped inductors from cinemag and teflon caps from CCCP (line trannies are from cinemag too). It sounds right but I am considering an upgrade to nickel cored inductors.

Values of the inductors are 33, 47, 68, 82, 150mH with a signal in the 5Vrms range.

Dave, is there any chance you can mix some copper and nickel for this project? I need to see if other values may be of interest (thinking to go down to 300Hz or so).

Gianluca
Attachments
This is the ratnest... The 801A power supply exploded after adding other components ... i need to work on it.<br /><br />The xover are finished using exotic wood and expert craftsmanship (LOL)
This is the ratnest... The 801A power supply exploded after adding other components ... i need to work on it.

The xover are finished using exotic wood and expert craftsmanship (LOL)
13012008.jpg (259.21 KiB) Viewed 8426 times
This is the basic schematic. It is NOT updated as values have been adjusted on the bench.
This is the basic schematic. It is NOT updated as values have been adjusted on the bench.
801A_XOVER.jpg (524.14 KiB) Viewed 8426 times
Measurements. Theoretical vs actual.
Measurements. Theoretical vs actual.
filter_test.jpg (158.01 KiB) Viewed 8426 times
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Hey Gianluca,

Sorry for the late reply and I figured this would be the best place for the topic.

I find it interesting that you found the lower impedance networks sounded better. Thomas Mayer also insists this to be the case so you are in very good company. I find what is required to drive the Low Z networks causes more harm than simply using a higher Z network (very limited sample on my part)

The tapped inductor approach is definitely the way to go, are the cinemags on nickel cores? I'd hate to give you something similar to what you already have. Chances are my versions would have a fairly substantial gap (which would also be adjustable) but again if your cinemags are already there (or close) I'm not sure how much I can improve things.

I'll gladly spin some copper for you and the stash of nickel and brass awaits your command.

If you still have the LTspice file, please upload it since it will save me a few minutes work.

dave
radio_elasais
Posts: 11
Joined: Sun Jan 13, 2008 11:04 am

Post by radio_elasais »

Ciao Dave

Thanks for the reply.

I am glad to hear others agree on the low Z. Cinemags are wound on small ferrite torroids and sound good. Pultec had ferrite inductors at the end of the day and their sound is out of question!

I was thinking to a gapped nickel core actually. I'll try to recover the old spice file but it will take a bit longer ... I am away from my office for a business trip for the whole week.

More later this week.

Ciao
Gianluca
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

hey,

no need to recover the spice file if it isn't handy. I just find it nice practice to post the .asc with any screen captures to others can play along.

I for one would be very interested in hearing a report of the difference between the two core materials in question so when you finish your trip we can work out the details.

dave
radio_elasais
Posts: 11
Joined: Sun Jan 13, 2008 11:04 am

Post by radio_elasais »

Ciao,

it seems that the old asc file has been erased. Btw I made some calculations by hands (no spice at home) and

150mH/0.79uF (2*0.47)
330mH/1.41uF (3*0.47uF)

would give me a 330Hz hi pass and low pass filter. Impedance slightly higher than 600Ohm. It sounds good to me.

The inductors should be tapped at 33, 47, 68, 82, 150 and 330mH. Expected Q and/or DCR would be a useful piece of informations. What core size are you thinking of?

Hey! I'll drop a mail off line for other info.

Ciao
Gianluca
Post Reply