on the LM1894 & Re: [sdiy] BBDs -= wrap-up =-

[anthony]

> This sounds very interesting to me, I'm totally unfamiliar with the 
> LM1894.
> Seems to me that I ran across them somewhere - I'm thinking it was the
> dreaded Futurlec.

I think National still makes them in SMD form. You have to have a license to 
use them in new gear.
No matter - I have a near endless supply of them.
It really works a lot similar to the NE570 series (but yet no so...) except 
it's single-ended. It uses a Gilbert cell just like the NE570, but instead 
of being basically a VCA, it's a VCF. The detection path cuts off signals 
below 800Hz then tracks the high frequency content. When no signal is 
present or when it's mostly bass the filter is kicked in. As the high 
frquency content increases, it backs the filter off. When you are hearing 
the music you don't notice the noise. In the end at your ears what happens 
is the same exact thing except with DNR (LM1894) it's just a one-step 
process - no silly expander at the end. This helped me keep my parts count 
down since I only needed one for stereo.

It also has an output for a meter so you can see just how much it's kicking 
in. You could use it for a limited envelope follower if you really wanted 
to. Which could be a cool way to change the delay time with frequency 
content.

It's sad this thing never caught on that much. Because I think in the realm 
of analog effects it excells over compandors in many ways. Although I still 
can't imagine why someone would design gear to use it BEFORE the BBD's.

Possibly the only disadvantage would be the need to use a limiter. But I 
think reducing the noise makes up for the loss in dynamics control. And 
Juergen Haible mentions that NE570's really aren't fast enough for use in 
noise reduction. If I had to use a compander I think I'd use something like 
he did in the Storm Tide. I do have the choice pair of 3080's...

>>using 3 BBD's instead of 2, but my goal is to use just ONE grid board
>>(276-168B) from Radio-Shack!
>
> Whew!  That's a lot to squeeze onto a board.  I bet you'll feel like a
> qualified micro-surgeon after its done.

Just call me Edward Tweezerhands.

Seriously - I have 3 different tweezers and 3 different angle pliers that I 
use constantly. And a Dremel tool for any recalcitrant parts.

Bending is the key to my 3-D construction technique. My philosophy with grid 
board is that if I can get all of the chips on there then I can get 
everything else on there too. Daughterboards are not forbidden either - 
though I don't forsee needing one. I even have room for the split power 
supply. That's one of my departures from the original Dim C. I'm using a +/- 
7.5V (+ 6.5V) supply from 2 9V batteries instead of one 9V. I'll get better 
performance out of the MN3007's that way. Mine won't be as small as the 
original C, but it will still be small enough for a stompbox. Really I have 
the advantage that I don't have to mass-produce this so I can do whatever I 
want. I have all the parts and then some. And I'm using better components 
for most of it. None of them are inferior at any rate.

What I do is put together sections and then put those on the board. Since 
most of the components I use have short leads this means touch soldering 
lots of leads. It's very relaxing though - like Zen gardening. I try to 
think of the best way to stack them up and I also consider aesthetics - I 
always throw in some clear-yellow, orange and boxy yellow polycarbonate caps 
(I wish I had a lot more of those - Philips brand). And these days the 
newest electrolytic caps are really small (and the ones in computer supplies 
are excellent). The biggest (physically) cap in my Dim is the 4.7uF in the 
LFO. I used a polycarb instead of a non-polarized electrolytic (although it 
would be cool to try larger values here - I have some 25uF non-polars). It's 
folded over on top of everything - looking cool with the thermal insulation 
on the leads (it's what I had around...). If noise is a problem I have some 
brass mesh that I'll use for a ground plane (wouldn't that be a car?).



>>bunch of inline M5218's. I wish I had a lot more of those.
>
> I was only able to find those in DIP package.

I got mine from lots of old stereo gear. I really wish I could find a lot 
more inline dual op-amps. Everything else is fine
DIP I guess. But dual op-amps are way better - especially when you're using 
grid board. I wish I could find some real actual veroboard around here. It 
looks more versatile than the Radio Shack stuff.

>>I have 6 dual
>>op-amps on there! (Along with everything else.) I'm using 2SC2240GR's for

Actually it's 7 I forgot about the CV inputs! :-)

>>the LPF's, but I have some 2SC732BL's and a bucket-load of 2SC1815GR's...
>
> I used the Japanese transistors mainly because I wanted to keep things in
> the audio path as close to the original as possible.  I'd planned to (but
> never did) sub in other transistors to see what worked out or not.

A lot of Japanese transistors are excellent. I like the 2SC732 and the 
2SC2240 and the 2SA725. The 2SC1815 and 2SC1740 are pretty good too. A lot 
of stompboxes use the C1815GR's. I think this is the one I have the most of 
by far. Well except maybe for the C1740S. I have a ton of those too. I think 
the transistor I have with the highest probable gain (haven't tested it) is 
a 2SC1775F.


>>The 3207 is in my Radio-Shack reverb which I've hacked and thrashed a lot.

>>the op-amp on the right side. What? I know you guys all have one of these
>>somewhere in a closet.
>
> Mine disappeared years ago =-(.

Schade.
Zut alors!

>>soon). This a lot like what's done on the Storm Tide but I wanted a 
>>bigtime
>>"Theta Processor"
>
> One of these days I'd like to try something like that out.

It's easy because they're all fixed. Get one of those boards from 
Radio-Shack that's just like the lay-out from one of their solderless 
breadboards (have one - couldn't live without it). Line up all of the old 
324's, 4136's whatever quad op-amp you have the most of - line 'em up flush. 
Then add capacitors and resistors. I've thought about making one with all 
trimmers - even trimmer caps, but I don't have enough of the caps for a huge 
one. I do have a lot of really cool (and probably very expensive when they 
were new) silver and ceramic trimmers. Not quite big enough. But I do think 
there is something to be gained by tweaking a fixed phase-shifter for use 
with BBD's in a Theta Processor.

>>I want to make something liek the STD. I want some MN3011's...
>>I just found a schematic today.
>
> That's my next big project - I'm laying the groundwork now.  I don't know 
> if
> I'd put 'STD' on a piece of gear, so I'll probably call it Variable Delay 
> 1,
> or VD-1 for short =-D.

HEE!

> I know you prefer to keep things to a small budget.  Having said that, 
> Small
> Bear has the MN3011, but it's not cheap - $15.95 apiece.  As expensive as

Indeed I do. Yeah I read about that. I think I can stretch my shoestrings a 
little for a couple of these. My whole approach is to keep my budget teeny 
tiny for the regular stuff then I can get cool expensive stuff when it comes 
along then feed my cat and then maybe eat.  B^]

I've been debating between starting with one of these or with a MN3005. I 
actually thought I would string together a bunch of MN3007's, but that only 
makes since up to about 8. After that it just seems like too much trouble.
That many MN3005's would be rad.

Except I think I can do that just as well with the digital delay I'm working 
on. The parts count on it is a lot, but I had enough in my junk to make like 
2 or 3 - a 74HCT4046A was perfect for the clock. I wish I new how to string 
up more RAM on it. I have so many 41256's. It should be simple.

I think a delta modulator type can sound just as warm as a BBD.

> Or, you could keep your eyes peeled for a broken Scholz Rockman - they 
> used
> it for the reverb.

That would probably be the part that was broke. If I had a broken one of 
those I'd fix it.


> more prized than the later MN3207 incarnations.  The MN3207 does pretty 
> well
> in the Dim C.  But there's a lot of emphasis, de-emphasis, companding,
> filtering and other histrionics to keep the clocks from hetereodyning, not
> to mention the equalization/crossmixing, all of which play into the 
> general
> overall sound of the thing.

Yeah my lay-out ended up being perfect for keeping the clocks apart. At 
first I focused on keeping the BBD's and stuff from the input, but where the 
SILs need to be dictated where everything else would be. Of course after I 
got it all soldered on I figured out that if I rotated it all 90 degrees and 
then shuffled it would be perfect but oh well. There's probably a problem 
with that too so I'll not dwell on it. I'm sure as hell not desoldering all 
of it. I usually don't solder all of the pins on the IC's before I have it 
all done, but they way I am building this it will be better that way - some 
of the parts have to be on the trace-side - just some resistors and 
capacitors here and there.

I studied the 'emphasis, de-emphasis, companding, filtering and other 
histrionics' a lot. My first instinct was to change it all because I thought 
it looked weird but then I remember what you and Juergen had written about 
it so I studied it more intently. I think I read somewhere that the D has a 
little different equalization. It'd be cool if I could find those details.

> Hmmm...I have an old Bounty Hunter metal detector that looks somewhat like 
> a
> Geiger counter chassis.  Hmmm....

Hmmm.... Indeed!

>>I have the Morely schematic. Mine's a little different. It uses an op-amp
>>instead of the darlington and the 4007 is hooked up a bit differently.
>
> It's a good clock setup, but as I understand it, won't hit the freq's a 
> 4046
> based clock will.

I think I might have actually been better off using 4046's in my Dim C, but 
I don't need high clock speeds.
There was something about the 4047 that I liked that I can't remember now. I 
wish it had the VCO inputs like a 4046.

>>I've thought about getting a CV flanger to track v/oct for some modeling.
>>I've heard some examples and actually love that psychotic flute sound. I

> I'd say you've got some good devices for that kind of stuff (the R5106's).
> It'd be nice to have some 128 stage devices for that kind of thing, too.

Yeah I have 4 MN3006's. I can get as many as I want. I just have to wait a 
little.
Except I put those in a flanger. I thought having them all in the same 
flanger would  be cool.

OK that makes 3 things I have half-finished...

>>Oh yeah! You're the Birth of a Synth guy!
>
> Yes, still gestating.  This water retention is killing me =-D

Heh.

> good flanger from something rather ho-hum.  An added benefit to higher 
> clock
> rates is also higher bandwidth, which is good for chorusing and flanging,
> but (IMO) not quite as vital for longer echo type delays, which sound a 
> bit
> more natural at lower bandwidth anyway.

Everyone should try clocking a BBD down to like 2kHz. It's not great but it 
is fun. The bandwidth is so low that formants get mangled. Weird. The 
Radio-Shack reverb is perfect for this kind of mucking about. I'd recommend 
it to anyone starting out making BBD stuff to get one of these on eBay.


>>ones from tape decks from the 70's: 2SC1815GR/BL's, 2SC732BL's... I
>>mentioned all this earlier I think...
>
> I like'em.  Especially the Japanese dual trannies.
I don't have any of those. All of the dual transistors I have are 
super-exotic selected ones from discarded cyclotron gear and such. Scavaging 
from hardcore scientific refuse has a sweetness that cannot be described. 
It's a hoot to find out how much the parts cost to replace. I wonder if I 
could get them insured? Nah. They aren't really worth anything to anyone but 
me. The only thing that sucks is that I can't find any datasheets for these 
dual trannies. But it's plain that they are choice. I guess I could just use 
them and see.

> I should close this book - I heard a distinct groan emanating from the
> Detroit region =-D.

/That's/ what I was hearing...

> Cheers,
> Scott

hasta l juego