Active Tube Crossover:
*** Final PC board is now available.
*** - Steve Bench designed a well done 2 way crossover design based on the 6DJ8 dual triode.
He was kind enough to allow me to produce a batch of PC boards based on his design to simplify construction. As his website is no longer up I've mirrored the appropriate files here:
Note that all schematics on these pages are (c) Copyright,
1996-1998, Steve Bench. All rights reserved. You may use or
build these FOR YOUR OWN use freely. You may NOT build and/or sell
for commercial profit, any of these without a royalty agreement
in place with the author.
You may place these schematics on your web page if you credit the
Note that high voltages are present in these circuits. The author
IS NOT RESPONSIBLE for ANY injuries, damages, etc. associated
with construction, or use of these circuits.
The Tube (valve) crossover shown here is a 3 pole lowpass and
highpass used for a 2 way crossover design. The circuit needs
+150VDC and 6.3V filament power for its operation.
The circuit is a modification of a unity gain non-inverting
Sallen-Key circuit that provides 18 dB per octave rolloff
characteristics. The center frequency shown on the schematic,
but may be modified to any frequency with suitable scaling.
Some representative value changes are shown in this document.
A 3 way crossover may be implemented by cascading two similar
in---------- xover hp----------------------------high freq amp
-------------------lp----xover hp----------------mid freq amp
-------------------------------lp----------------low freq amp
Value changes for other crossover frequencies:
LF Side LF Side HF Side HF Side HF Side LF Side LF Side LF Side HF Side
Freq R4 R5,6 R9 R10 R11,12 C2 C3 C4 C6,7,8
800 90.9k 100k 88.7k 18.2k 1.3M 2200pF 0.01u 270pF 2200pf
80 90.9k 100k 88.7k 18.2k 1.3M 0.022u 0.1u 2700pF 0.022u
85 84.5k 93.1k 82.5k 16.9k 1.2M 0.022u 0.1u 2700pF 0.022u
100 732k 806k 154k 32.4k 2.2M 2200pF 0.01u 270pF 0.01u
125 576k 634k 124k 25.5k 1.8M 2200pF 0.01u 270pF 0.01u
170 422k 464k 154k 33.2k 2.2M 2200pF 0.01u 270pF 5600pF
200 365k 403k 158k 33.2k 2.2M 2200pF 0.01u 270pF 4700pF
400 182k 200k 174k 36.5k 2.7M 2200pF 0.01u 270pF 2200pF
850 84.5k 93.1k 82.5k 16.9k 1.2M 2200pF 0.01u 270pF 2200pF
1000 73.2k 80.6k 154k 32.4k 2.2M 2200pF 0.01u 270pF 220pF
2200 73.2k 80.6k 154k 32.4k 2.2M 1000pF 4700pF 120pF 100pF
3300 73.2k 80.6k 82.5k 16.9k 1.2M 660pf 3000pf 82pF 270pF
4800 73.2k 80.6k 75.0k 16.2k 1.1M 470pF 2200pF 62pF 91pF
Calculating other crossover points:
Other crossover values are best calculated by Simple scaling!
The ratio or reactance to resistance is essentially constant
for all frequencies. You can raise the resistance or raise the capacitance
to alter the frequencies. Works for low or hi pass sections.
Find the ratio for each RC section at the given crossover frequency
(R=R, Xc=1/(2*pi*f*c). If you change the frequency, you need to maintain
the same ratio of r/xc for each pole pair. Easiest is to use a "common"
capacitor value, then scale the R accordingly. When the value of R
gets far outside "normal values", change the C and try again.
For instance to lower the frequency form 1000Hz to 900Hz (a 10% change)
simply raising the resistor values by 10% works. You could likewise
raise the capacitor values by 10%, but that's harder to do from a
now-try-to-actually-find-them view :-)))
For the low capacitor values, you also have to take into account the
circuit stray capacitance (a few dozen pF, but if you have a 100pF, that
is a significant change). For low resistor values, you have to take in
the "source" resistance into effect... the gm at the chosen op point
is about 10mS, making the cathode effectively look like about 100 ohms.
Thus, values about 5k and above are "safe" as the effect is minimal. For
driving the crossover from a "source", a 1k source implies values above
50k are "safe" as the effect is minimal. = Steve Bench
For the Altec 604C, a correspondence with Steve:
In a message dated 01-08-02 15:29:00 EDT, you write:
> I'm exploring options for an active crossover to be used with Altec 604C'
> s. Your design appears simple and straightforward. I wonder if you could
> suggest parts values for a 1,600 hz crossover point. Also any hints or
> suggestions would be greatly appreciated.
1600Hz values the simple way: use exactly half the capacitor values shown
on the schematic. Sooo
C2,6,7,8 = 1100pF (which can be 1000pF in parallel with 100pF
C3= 5000pF which can be 2- 0.01uF in series
C4= 135pF. Taking into account 20-30pF of stray capacitance in the
circuit due to construction, That would mean using a 120pF capacitor.
Best Regards, Steve
The prototype PC Board was populated with the tube socket "up" and the passive components "down" to allow easy mounting of the tube up thru a chassis.
I ran out of the bareboards, but am planning on making another batch up shortly. Email me if interested, and please us an appropriate subject or you'll get spam filtered
Thanks to Steve for making his design available!
Back to Main Page