I haven't gotten round to making one, but its basically the same as post #451 but with two stages of filtering to make it 24dB, the bass section output is taken from the JFETs sources.
Each section is a small block so shouldn't be too hard to work out, the CX and RX are the filter components, the rest are JFET input and sourse resistors.
There is a preset at the output of each filter (HP & LP) for level matching, link out as required.
The big caps and resistors are PSU filtering, the diodes are what I call 'idiot diodes', they only conduct if you connect the supply backwards.
DC offset shouldn't be too much of a problem for the HF section, but there could be a small offset on the LF output, an offset voltage could be injected to null the output (à la B1), but I decided to trimm the apropriate JFET source resistor to match the IDSS of it's counterpart.
I haven't been able to find any pro/con info on the new tapping point (must be the serch perameters I'm using), if you get round to it, I would be interested in your findings Andrew.
I just stumbled over this thread -- brilliant stuff! How could I miss thisone for so long? I admit I haven't read the whole thread yet, so it might be in there somewhere, but... is it possible to use this for a Linkwitz-Transform, too? The arrangement of the R and C parts in the filter section look very much the same as in the typical opamp based Sallen-Key filters, and I suspect this would be the same for a Linkwitz-Transform. Correct?
Coming up with a flexible PCB that allows building different filters would be really cool, maybe even with a cascade of several filter blocks in series for more complex arrangements (could be jumpered if not used).
Coming up with a flexible PCB that allows building different filters would be really cool, maybe even with a cascade of several filter blocks in series for more complex arrangements (could be jumpered if not used).
The B1 filters are an implementation of a Unity Gain S&K.
That needs a Buffer with a very high input impedance and a very low output impedance. An opamp wired as a Buffer and the two jFETs wired as a B1 do the same job.
Go and read
Active Filters
You will see that most of the circuits do not use a Buffer. The opamp uses the feedback circuit to achieve the transform.
mbrennwa, posts #341: http://www.diyaudio.com/forums/pass-labs/156094-b1-active-crossover-7.html#post4255228, #370: http://www.diyaudio.com/forums/pass-labs/156094-b1-active-crossover-8.html#post4265322, #391: http://www.diyaudio.com/forums/pass-labs/156094-b1-active-crossover-8.html#post4270362 and #438: http://www.diyaudio.com/forums/pass-labs/156094-b1-active-crossover-9.html#post4503609 pretty much sum up what I have come to know about the topic. Post #434: http://www.diyaudio.com/forums/pass-labs/156094-b1-active-crossover-9.html#post4443741 shows a sketch of a proposed layout. Notice there's a wiring mistake at the second buffer, the one after the BSC.
Edit: Post #274: http://www.diyaudio.com/forums/pass-labs/156094-b1-active-crossover-6.html#post3437016 show pics of my completed build. Still up and running.
Edit: Post #274: http://www.diyaudio.com/forums/pass-labs/156094-b1-active-crossover-6.html#post3437016 show pics of my completed build. Still up and running.
Last edited:
Guys, this circuit is very easy to etch on a copper plate plus it sounds good so better use the best passive elements you can. My next one will use polystyrene caps. Remember you can also add a volume pot in front of the first buffer and take one extra pair of outputs from this buffer to have a "preamp" function, apart the XO outs... 
What are you using to get rid of the offset?
I thought about variable voltage regulators for the PS and also 1MOhm pot to ground for the outputs with persisting offsets.
Thank you very much.
M.
What are you using to get rid of the offset?
I thought about variable voltage regulators for the PS and also 1MOhm pot to ground for the outputs with persisting offsets.
Thank you very much.
M.
The HP section has series caps, so DC offset is less of a problem there. Good matching is required for the LP section if you want low cumulative offset. I managed single digit mV offset with nicely matched JFETs. See post #274 http://www.diyaudio.com/forums/pass-labs/156094-b1-active-crossover-6.html#post3437016
Last edited:
The high pass S&K block the offset at each stage.
So, the accumulation of offsets from cascaded stages is worst for the low pass S&K.
This could be blocked at the last stage or at various points along the way.
B1 with fairly closely selected Idss within each stage generate quite small offset. The accumulated offset may be so small that it is insignificant. All that is needed is the normal input and output AC coupling that one applies to every single ended supply amplifier.
So, the accumulation of offsets from cascaded stages is worst for the low pass S&K.
This could be blocked at the last stage or at various points along the way.
B1 with fairly closely selected Idss within each stage generate quite small offset. The accumulated offset may be so small that it is insignificant. All that is needed is the normal input and output AC coupling that one applies to every single ended supply amplifier.
If you decide to go with a DC coupled topology, then I strongly suggest you look at adding other detection/protection systems.
Depends on the impedance before and after the filter. If you have a
high source impedance compared to the impedance of the buffer, then you
need a buffer before the filter. If you have a load impedance which is low
compared to the filter, then you need the buffer after the filter.
You might need both.
high source impedance compared to the impedance of the buffer, then you
need a buffer before the filter. If you have a load impedance which is low
compared to the filter, then you need the buffer after the filter.
You might need both.
Where you referring to the LF alternate tap?
https://pdfs.semanticscholar.org/4022/c82039d953256183210f72fe1f615791a5df.pdf
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.