Pass Diy High Low Pass Please

[ThorstenL]

Konnichiwa,

Originally posted by macka
Has anyone done any ground work or built up their own xover using Jfets or whatever.

I thought a generic 2-3 way mother PCB with the power supply and input/output buffers and plug in daughter boards for the filtering offering 1-4th order / bessel, butterworth/ LR would make the project almost univeral.

Well, this is going back about two decades, but I once needed a good, easy to make electronic X-Over, 4-Way, 4th Order (L/R) that allowed easy change of x-over frequencies and needed to be made from very generic parts. It also had some tricks applied in circuitry. This whole thing was for PA Applications.

On main "advantage" was the lack of a requirement for the presence of a PCB. It used Plugs and sockets similar to the modern "DB9" Format (PC -Serial Port) for the frequency determining components (yes, a 4-way X-Over had a total of 8 Plugs per channel with a HP and LP per channel and four level controls).

The basic active circuit had 4 transistors & 2 resistors, composed as Darlington Emitter Follower and "ring of two" current source. This makes a basic 4-Pole circuit with input, output and +/-V (I used +/-24V as I had suitable complete 24V PSU's at hand). The resulting buffer is pretty good sonically and measured and can literally be build by supergluing 4 Transistors with a BCE pinout plus adding 2 Resistors. I used to wrap the transistors in addition info copper tape.

The filter around this buffer was standard Sallen/Key 2nd order, 2 daisychained sections to give 4th order slopes. The Buffers where soldered directly to the Sockets for the filter plug-ins and then their copperwrap and PSU connections here soldered together per socket, making a compact, frontpanel mounted single filter module. The Filter RC components where fitted to the DB9 style and then the plug case was attched to protect the components.

The main problem is that these circuits have around 1.2-1.4V DC offset. This normally prevents DC coupling, however if the buffers are made up alternatly from NPN and PNP transistors it is easy to DC couple throughout with the offset being eliminated.

In the highpass where Capacitors are present anyway and block DC, here the final sections bias resistor is returned to a spare buffers output whose input is grounded (via a suitable value resistor to minimise offset after 2 sections).

You of course need added buffers after the Level controls (before going to the filter) and in my case I had further buffers ahead of the level controls and of course a line input transformermer for balanced inputs. Equally, the outputs where also balanced using transformers.

IIRC I also had inserted between the post level buffers an insert point for equalisation like the HF EQ for compression driver/horn combos and LF lift for LF horn/reflex LF sections again as plug in modules.

This approach saved developing a PCB and making it (in those days I had to handpaint and etch PCB's, no CAD programs, easy print & transfer via cloth iron onto PCB material. The entire X-Over was bolted to the front of a 19" Case with square cutouts for the RC module sockets that had the active circuit on the back. XLR sockets and transformers where on the back of the Case, PSU's external in a further 19" case. I build quite a few of these and never had one coming back for service over quite a few years, despite road use.

I think this sort of thing would still suit a DIY'er who just wants a good sounding active X-Over that is easy to make and is highly flexible. Okay, looks are pedestrian unless you fit the plug in RC modules inside the main case.

Sayonara

[moe29]

Grey,

If you do post a AXO article, and it's half as good as your previous
posts - the wait will be well worth it. I for one appreciate your
contributions - they always ADD something to the community, and
always bring up excellent discussions and above all else, another
box that makes music.



m.

[Fred Dieckmann]

Don't go to pieces I have talked to a couple of people working on some simple SK follower circuits that should work pretty well......

picture removed on the request of the copyright holder.

[Peter Daniel]

Fred,

Are you waiting for Grey's crossover too (like you need it?)? Otherwise, your lack of patience is hard to explain

[Variac]

Yeah, It looks like Fred can't come up with his own!

[AKSA]

Don't bet on it, Fred has many designs already built!

Here's one to consider, likely way off beam, but interesting, and simple: http://www.ne.jp/asahi/evo/amp/el86/report.htm

This amp is a hybrid, the one I am considering is the one with damping factor of 5.7 and distortion of 0.045%.

Has anyone built this amplifier? I'm interested in giving it a go. Looks very interesting; uses a Rush cascode, too, novel.

Cheers,

Hugh

[uli]

I wonder whats the problem to build an active Xover using
NP´s ideas. If you take a look in the manual of the original
Passlabs Xover you can find everything needed to build an
Xover. NP described building discrete opamps in Audioxpress
or was it AA at that time (about 99).
Ok now I can use his opamp article ( article )
and fit those ops into the circuit shown in the manual.

Uli

[Nelson Pass]

Electronic crossovers are not hard, it's just that the range of
choices is daunting.

Nevertheless I will give you the first officially frozen spec:

One side of the PC board is 8.05 inches long.

[Fred Dieckmann]

You mean that was a simplified circuit? No wonder I couldn't get it to work! Most of the Sallen Key filter circuits for active crossovers can be built with followers and the Q manipulated by component ratios. The no feedback guys won't wring their hands that way.....

[GRollins]

As of yesterday, I finished the first draft of a story I've been writing, so now I've got a little breathing space. (Got a story in the current issue of Analog [Jan/Feb '04] if anyone reads SF.) I'll try to get started on the writeup for the crossover dingus--perhaps as soon as this afternoon.
As I said earlier, and as Nelson just said, it ain't all that hard to fiddle together a crossover, it's the ga-jillions of permutations that make it so intimidating. And choices mean explanations, and explanations mean lotsa writing to try to say it clearly. What, you think Nelson wanted to sit down and write a small novel for owners of the XVR1? Unlikely.
Any fool can toss out the equations and say contemptuously,"Here, go build yourself a crossover." But what topology? What frequency? What Q? For that matter, just what the hell is Q, anyway? To someone who's just starting out, the question of Q could be an entire thread on its own. And compound slopes? Jeez, most people don't even know they exist, so how to explain?
Whew! I'm tired already, and I haven't even started the writeup. Somebody fetch me a glass of wine, this is gonna take a while, and I'll need something to soothe my tattered nerves.
(I just figured out why Nelson wants me to go ahead with this crossover...I'll end up being the shock troops storming the beach and weathering the heaviest barrage of questions! Ah, well. Once more into the breach...)
And all this for a niche product that will only appeal to, what, 5% of the audio community? It's no wonder that there are so few commercial crossovers. I urged cj to do one all the way back in about '84. Whether my words had any effect I cannont say, but they did eventually release one. It's no longer available. Why? Probably because it wasn't commercially viable. ML used to have one, but I don't think they do at this time. Etc. etc. etc. It's a labor of love and/or necessity, not something you do with the expectation of making any money. Unless you hope to sell more amps in the process.
Anyway, I'll try to get started on some simplified (i.e. smaller, so they'll come in under the silly pixel limits) schematics, then get words stitched around them. If someone else wants to lead the charge, be my guest. But you get to take point answering questions. I'll sit back and watch.
Huh...maybe I'll just go camping this weekend and gather my strength...

Grey