My little Beast (Updated - Panel Time)

[SSLtech]

HI all,

I've been watching this approvingly, -just haven't posted in the thread yet.

Lukas, I got your emails, and I'd already been watching the thread, but I don't feel like I have much to contribute just yet... Also I've been a little tied up with installing a few new rooms, and round here the Christmas break is the only time we get to put them in... You've seen the place, so you know how it is!  ;D

But looks great though.... those black PCBs are GORGEOUS, and that trace layout really does put the word "art" back into "PCB ARTwork".

Looks GREAT though... I'm watching the thread; -very inspiring!

Keith

 

[livingnote]

Aww, Big Hug! I was kinda thinking you were probably somewhere in your ball of wires
but it's great you posted because now I can finally tell the little voice to shut his trap  ;D

Things are just starting to get way out of hand here...

Anyway, where were we?

Right  ;D

 

[Kingston]

Before it all gets out of hand with too many ideas, I would just like to ask, why a transformer anywhere in a SSL variant, output or input?

besides galvanic isolation, what would be the point? you already have opamps coloring everything, why more? save your money. you could instead improve the gsll output with true differential design, instead having the pseudo balancing of the current design. see rane.

 

[livingnote]

Well, mainly for creative choice. If you have this iron that does sweet stuff you can
have more fun when you're at work - you never know what you'll encounter and a
little bit of trafo sweetening for premaster would sound like a good choice for
certain kinds of recording...

Can you show me where the alternative output is, though? Don't think I'll really
get it on board with the space considerations, but who knows?

 

[emrr]

I'd sure be tempted to change the input and output to THAT 1240 and 1646.  Good work!

 

[radiance]

... why a transformer anywhere in a SSL variant, output or input?

So you could leave out the in and output decoupling caps...meaning less coloring.
An input transformer has been done and supposedly it sounds cleaner, not like a ssl actually.
Output transformer did not made much of a difference and was replaced with a line driver.
Here's the thread > http://www.groupdiy.com/index.php?topic=8174&start=0

Like emrr said..try the THAT variant..or maybe input iron plus a 1646 output stage.

...Or go all the way and use DOA's although I don't know if it's worth it having a VCA (obviously) in the audio path as well...

 

[livingnote]

THAT sounds good too...aww it all sounds good. What do you do as a designer?

I think I'll just make a couple different footprints, I see I won't be getting out
of this project without trying some different IO's. Just for curiosity, too...

...but hang on. Do we even need to debalance if the signal is going through
the VCAs phase-flipped anyway? Sounds like washing the dog and then going
to the tar pit...

 

[Kingston]

... why a transformer anywhere in a SSL variant, output or input?

So you could leave out the in and output decoupling caps...meaning less coloring.

I disagree with this. In a properly balanced input and output configuration, you will have negligible impact from any I/O caps. On the other hand, there isn't a negligible (ie. transparent) sounding transformer in the world. Lundahls get close, but they still have a distinct transformer sound.

Besides, the I/O caps aren't exactly mandatory here. Try it. They are only for extra protection.

Can you show me where the alternative output is, though? Don't think I'll really
get it on board with the space considerations, but who knows?

http://www.rane.com/note124.html

opamp count increases by one, but it's favourable to the above-mentioned THAT options, since it still leaves plenty of choice for opamps.

Similarly (but not related to any of the the above), the input should be a "superbal".


PS. The I/O stage upgrades are not my idea originally. tv (the prodigy pro member) suggested them in one of those recurring opamp replacement threads.

 

[tv]

There were some ideas floating around in this thread...

http://www.groupdiy.com/index.php?topic=33191

And perhaps somebody could mod the existing input stage like this (look at the sch in this post, not the one before which is faulty):

http://www.groupdiy.com/index.php?topic=36967.msg455520#msg455520

(with 22k and 11k resistors you effectively get a quite balanced input impedances and time constants).

 

[Kingston]

tv,

what's the basis for this with regards to the cheap mod: "I'm certain that the last fig. I posted would sound tighter/more focused than original."

I understand the trade-off between less common-mode rejection, and the resulting "mismatch" of the impedance between of the two inputs. But why is the mismatch a positive thing? Why would it theoretically result in a better sound?

 

[tv]

I don't think there's any "tradeoff" present in the correct sch. I posted. It's functionally equal to the original input stage, but it has the same (22k) imp. both in negative and positive leg (was 44k in the positive before).



(Just discard any comments referring to the faulty sch. I posted before, they arent valid for this sch anyway.)

 

[radiance]

... why a transformer anywhere in a SSL variant, output or input?

So you could leave out the in and output decoupling caps...meaning less coloring.

I disagree with this.

Well, I should not have posted it as a fact...it's more like: the person who tried this out thought it sounded cleaner. Input transformer used was a Jensen JT-11P-1

The reason he tried it was also

input transformers because:  better CMRR, lower noise, natural RFI shielding (as Bill Whitlock noted).

 

[regularjohn]

I was really pulling for Sebastian Vettel this year.  Bummer...
Sorry...off topic

 

[livingnote]

Boy chewing the fat on F1, I'm kinda outta my depth

But here's my idea for where to take this thing long-haul:

In L+ Buffer     ->     VCA  \                                         /  Out L+
                               THAT Reciever  ->  CnB  ->  THAT Driver  
In L- Buffer     ->     VCA  /                                         \  Out L-

In R+ Buffer     ->     VCA  \                                         /  Out R+
                               THAT Reciever  ->  CnB  ->  THAT Driver
In R- Buffer     ->     VCA  /                                         \  Out R-

This way I surmise we can have a serious whack at CM where we
need it, getting VCAs to cancel each other's distortion and doing standard
laundry service. Then make the CnB highly transparent so it flies
below radar and output to scrumptious 1646.

Choice of input buffers might be an interesting thing...

 

[tv]

If you are targeting transparency etc. wouldn't it be better to simply buy a Duende?

 

[livingnote]

Does the Duende sound that good? I've only read about it...never heard one.

 

[tv]

Its digital. I'm not owning one. It saves you a roundtrip trough converters each time you want to compress something. Personally, I don't see a point in overdesigning a circuit that is a "dead" technology. If I was fooling with it (but I'm not), I'd seek to throw out of the schem every little bit that isn't needed for the device's "mojo", while optimizing and pimping the stuff that contributes to the "magic".

But I'm biased - I'm always seeking the shortest signal path and relegating duties to digital whenever I feel something isn't worth my time and effort. However, I spend ridiculous amounts of time fooling with instrument preamps. For me, this is where it's at.

 

[strangeandbouncy]

Dudes,


  I have a Duende. it sits in a box. It is a pain in the arse. LAtency issues on busses etc etc. Sonically, It is quite good. but that's as far as it goes.



   the mixbus compressor, like all the digital recreations of the SSL mix compressor, is remeniscent of the real deal, but it does NOT do the same thing for me! there is no "interaction" with the signal, and it just fails to ignite any fuse for me. Very handy if you have no outboard, but that is as far as it goes for me! I have yet to hear any recreation that glues and sucks and smoothes like any of the SSL's, Al Smart's, G-SSL's(turbo) with any of the vca's I have got(2001, 202c, 2151, 2181). They all sound (slightly) different, but the duende pales for me.


   There it is. Sorry, SSL, I mean you no disloyalty, and everyday I WISH that it was possible to just use a plugin in the box - recall's, switching chewns. etc . . .


   . . . and I would say fizz not transparency . . . .


          ANdyP



           

 

[livingnote]

Yeah that's what really got me going too - my friend Marco has a whole pack of plugins studied together
out of x load of "the best" from what's on offer, and we kicked the **** out of all of them with the ol' GSSL
and that wasn't even in Turbo mode back then.

Now I was just kinda getting a tad worried about sending the signal through so many 5532's in direct succession
as to how much is too much - I heard that the THAT recievers and drivers have a character of their very own
and after all the datasheet reading about how fantastic CMR is and the lot, that this might be a nice cocktail
when it comes down to it - color, yeah, but without loss in detail, which was a bit of the disease of my older rev,
so in the end I just couldn't leave it alone. I guess it's just this thing about making something analog that
doesn't have to hide from the Neumann and the Halos when it comes to signal, and ultimately I guess more
to do with getting myself to the point where I can design really cool stuff rather than copying same old same old.

 

[audiovisceral]

Can't wait til you're selling these. The old G-SSL project was too bloated with hpf, turbo, crush & blend, and crc. Nice to see it all come together. If you can add a switchable output trafo like you mentioned that would rock!

Beautiful stuff!