Buffer for Pass' LXmini OX/EQ

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Note to moderator: I hope I have posted this in the relevant section - otherwise I apologize, and please move it.

I was interested in trying out Nelson Pass' analog XO/EQ for the LXmini (Analog XO/EQ) . However I don't have enough of the 2SK170/2SJ74 JFets, which is problematic for two reasons: they are quite hard to come by, and if you find them they are very expensive. So I tried some Fairchild J113/J176 - they are not really complementary, but they might work anyways.



In the test set-up I used matched FETs with about 12 mA Idss.

With +/-12 Volts supply it'll supply around 7 Vrms without visible clipping. Soft clipping occurs around 8 Vrms.
THD and noise is below 0.025% at about 3Vrms (I cannot supply a higher voltage distortion free).
Frequency response: Flat from DC to at least 1Mhz. 100kHz square wave shows a tiny amount of ringing, but this could probably be removed if the bandwidth is lowered.
Rise time estimated at 0.5 µS.
Output impedance is around 35 ohm. This rises to about 70 ohms if a trimmer potentiometer (100 ohms) in installed in the sources to balance out small differences between the Fets.

These specifications are, of course, far from as impressive as the Toshiba pair - but we're talking about line level signals and how audible is 0.025% THD after all?

Is there anything I have over-looked?
 

Attachments

  • FET_Buffer.jpg
    FET_Buffer.jpg
    5.3 KB · Views: 1,551
The one and only
Joined 2001
Paid Member
These specifications are, of course, far from as impressive as the Toshiba pair - but we're talking about line level signals and how audible is 0.025% THD after all? Is there anything I have over-looked?

No, that's about right. By comparison the Toshibas or LSK's will get you into
below .01%.

You can tweak these parts a bit, but I wouldn't get too excited about it - that
.025% will be 2nd harmonic, not all that problematic, and for the LX mini
you are really listening at the <1 volt levels with a 30 dB gain amplifier.
 
The one and only
Joined 2001
Paid Member
SL is happy with the bi-amp version, and will be playing with the 3-way at
his leisure, but I expect no more than a tweak, since it is essentially the 2 way
with a bass channel added.

I will have copies at BAF, although likely the LX display will be using Frank
Brenner's power box with analog crossover, which I am told is an op amp
version of similar topology.

I expect to be able to make kits consisting of boards, or boards + Jfets
available by the beginning of the year through the DIYaudio store. A more
complete kit is not out of the question.

:)
 
The one and only
Joined 2001
Paid Member
2SK170, 2SK370 and LSK170 at about 10 mA are my favorites, but you can
use J113's if you are prepared to AC couple the follower output and degenerate
the current source for lower current.

A J113 with maybe 100 ohm source resistor will give you about 20 mA, which
is about the most I would run, dissipation becoming an issue at higher current.
 
The one and only
Joined 2001
Paid Member
BTW, probably final artwork is done, and I think SL will be happy with the
latest iteration of the 3 way version after I relented and put gain in the
circuit for the woofer.

We should have both crossovers as partial kits in the store real soon now,
which would include boards and the parts that go on the boards. They are
being set up to fit in Galaxy 2u 230x230 chassis and use a 24V external
supply.

As soon as SL signs off on the latest, we will be set to go.

:)
 

mkc

Member
Joined 2002
Paid Member
I relented and put gain in the circuit for the woofer.

Hi Nelson,

May I ask how simple a gain-stage you ended up choosing?

The reason why I ask is that a couple of years ago I made an active cross-over, for a friend of mine, using OPA2134. It works great.

But, no fun in that, so we have been toying with the idea of trying to redo the high-pass section, which comes in from 110 Hz and up. The ideas has been from doing some of it using valves to my thought of trying to do it with jfets using B1 style buffers. But, I might have the need for some gain. Hence my question.

Being on the subject of B1 style buffers compared to the "new" B1 push-pull buffers. The original B1 buffer has very good performance for such a simple buffer. Any subjective reasons for choosing the PP version over the SE version in a high pass filter where capacitors in the signal path is unavoidable?

Btw. I do have some SJ74 and he is a very good friend :), but I would like to save them if the benefit is only a little better measured performance.

Mogens
 
The one and only
Joined 2001
Paid Member
The current status of the LXmini crossover is that I am waiting for some
potentiometers to arrive while packing up the kits to go to the DIYaudio
store. 100 should ship by end of next week.

Kit consists of pcb, packet of 16 selected/matched Jfets + resistors specific
to them, and the rest of the resistors, capacitors and pots needed.

This is powered by a 24V wall wart and fits in the store's Galaxy 2u
230 x 230 chassis. Management is planning a rear panel with holes for
8 rca and power connector.

Documentation will be posted shortly.

I just finished the Jfet gain stage required by the LXmini+2 and Studio,
so this will also be appearing before very long.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.