Doepfer

If in fact the Frequency Shifter uses a quad. osc, then they've 
done it the right way and I might buy one!  I was curious about 
that.

That part of the circuit is required in correct a problem inherent in 
the 1496's design - it only issues a 40db cut of the opposing 
sideband. A corrpsponding signal 90 degree out of phase with 
the originalis required (via the quad. osc) iin order to phase 
cancel it out altogether, thus giving you true phase shift.

Tim, 

if you can get to the scanner, let me see the schemo, if you can.  
Dieter needs not worry - for $170 for the completed module, 
there would be little reason to attempt to replicate it from scratch 
- that would end up costing someone a lot more $$!.


> Tim
> 
> [The views expressed above are entirely those of the writer and 
do not represent the views, policy or understanding of any other 
person or official body. Personal email. This email is personal. It 
is not authorised or sent on behalf of the sender's employer. 
This email is the personal responsibility of the sender.]
> 
> 
> 
> The Information contained in this E-Mail and any subsequent 
correspondence
> is private and is intended solely for the intended recipient(s).
> For those other than the recipient any disclosure, copying, 
distribution,
> or any action taken or omitted to be taken in reliance on such 
information
> is prohibited and may be unlawful.

[... and will Yahoo! let me play today - if you're reading this then I have finally twigged what has been 'locking me out' for over a fortnight ('ray, 'ray!) - what a minefield!]

I have 'reverse engineered' the A-126 - if my drawings are complete enough and look 'scannable', and *if* I can get easy access to a scanner, I could probably make them available to others if you like. Its one of the modules I haven't completely sussed yet - from memory it is quite a beast: something like 6 all-pass filters for each of 'downshift' and 'upshift' and a voltage-controlled quadrature oscillator, all feeding somehow into 2 balanced modulators which actually (presumably) give the shift effect. The lower limit of shift (50-60 Hz from memory) is presumably a function of the voltage control range of the oscillator, which thus could probably be lowered, but then how well the multiplication in the balanced modulators works out, who knows...

Tim

[The views expressed above are entirely those of the writer and do not represent the views, policy or understanding of any other person or official body. Personal email. This email is personal. It is not authorised or sent on behalf of the sender's employer. This email is the personal responsibility of the sender.]



The Information contained in this E-Mail and any subsequent correspondence
is private and is intended solely for the intended recipient(s).
For those other than the recipient any disclosure, copying, distribution,
or any action taken or omitted to be taken in reliance on such information
is prohibited and may be unlawful.

Peter

> if you can get to the scanner, let me see the schemo, if you can.  
> Dieter needs not worry - for $170 for the completed module, 
> there would be little reason to attempt to replicate it from scratch 
> - that would end up costing someone a lot more $$!.

	Had a quick look at my 'scribblings' on the 126 last night -
unfortunately I have only drawn out a small part of it neatly, but if
I get time over the weekend I'll draw some more out tidily (and then
see if I can sweet-talk our secretary to let me use her scanner in
works time...). The quad osc uses the 2 VCAs in a CEM PA382 chip
(exponential response). The circuitry summing the shift CV which is
then applied to the 382 includes an 'offset' preset, and this should
thus be setting the lowest frequency of the oscillator. From my rough
calcs, assuming the gain of the VCAs is 1, the oscillator would run at
approx  15kHz: estimating the gain at the actual voltage applied for
max and min shift gave me frequencies of 6.5kHz and 30Hz, which tied
in well with the measured values on my 126 of 5.8kHz and 57Hz. Thus
reducing the lowest shift possible may be simply a matter of twiddling
this preset a bit (although the osc is AC coupled to the 1496 chips of
course...). This also suggests that it would be nice to see the
service manual for the 126 anyway, to establish just what Doepfer's
set-up procedure with respect to this preset is (and of course this
will include the schematic anyway...) - it's a real shame that its
priced just a bit above what I'm prepared to pay, just for the sake of
having it, but I could be even more tempted if Doepfer made it
available on a CD, 'cos then at least one could use babelfish  to get
around the language problem!

Tim

[The views expressed above are entirely those of the writer and do not
represent the views, policy or understanding of any other person or
official body.]

Got a note from Dieter last night, who saw these listings and he 
went into the theory behind the Shifter and how it cancels out the 
second sideband to isolate the sum and difference.  Much as I've 
always heard its done, by [hase cancellation using the 90 
degree output of a quadrit. oscillator.  It's a very complex circuit.

So....Dieter does read these listings (HELLO TO DIETER...YOU 
ARE A VERY NICE GUY AND A BRILLIANT DESIGNER).  So, 
watch your P's and Q's in here, guys...who never know who's 
lurking!!!

best,

Peter





--- In Doepfer_a100@y..., "stinchcombe_t" <tstinchcombe@q...> 
wrote:
> Peter
> 
> > if you can get to the scanner, let me see the schemo, if you 
can.  
> > Dieter needs not worry - for $170 for the completed module, 
> > there would be little reason to attempt to replicate it from 
scratch 
> > - that would end up costing someone a lot more $$!.
> 
> 	Had a quick look at my 'scribblings' on the 126 last night -
> unfortunately I have only drawn out a small part of it neatly, but 
if
> I get time over the weekend I'll draw some more out tidily (and 
then
> see if I can sweet-talk our secretary to let me use her scanner 
in
> works time...). The quad osc uses the 2 VCAs in a CEM PA382 
chip
> (exponential response). The circuitry summing the shift CV 
which is
> then applied to the 382 includes an 'offset' preset, and this 
should
> thus be setting the lowest frequency of the oscillator. From my 
rough
> calcs, assuming the gain of the VCAs is 1, the oscillator would 
run at
> approx  15kHz: estimating the gain at the actual voltage applied 
for
> max and min shift gave me frequencies of 6.5kHz and 30Hz, 
which tied
> in well with the measured values on my 126 of 5.8kHz and 
57Hz. Thus
> reducing the lowest shift possible may be simply a matter of 
twiddling
> this preset a bit (although the osc is AC coupled to the 1496 
chips of
> course...). This also suggests that it would be nice to see the
> service manual for the 126 anyway, to establish just what 
Doepfer's
> set-up procedure with respect to this preset is (and of course 
this
> will include the schematic anyway...) - it's a real shame that its
> priced just a bit above what I'm prepared to pay, just for the 
sake of
> having it, but I could be even more tempted if Doepfer made it
> available on a CD, 'cos then at least one could use babelfish  
to get
> around the language problem!
> 
> Tim
> 
> [The views expressed above are entirely those of the writer and 
do not
> represent the views, policy or understanding of any other 
person or
> official body.]

But did he say (or is it possible) to lower the shift range, or is it 
more trouble than it's worth?

--- In Doepfer_a100@y..., "petergrenader" <petergrenader@h...> wrote:
> Got a note from Dieter last night, who saw these listings and he 
> went into the theory behind the Shifter and how it cancels out the 
> second sideband to isolate the sum and difference.  Much as I've 
> always heard its done, by [hase cancellation using the 90 
> degree output of a quadrit. oscillator.  It's a very complex 
circuit.

he didn't go into it.  I would think he wouldn't be too keen in giving 
us the means to mods. 

This shouldn't be that much of a problem though.  I'm getting 
mine today, give me a wekk ro so to get aquainted and let's see 
what can be done.  I'm no Einstien, but I have had luck with a few 
simple mods of Doepfer, Analogue  Systems and Analogue 
Solutions gear.

Speaking of which, if anyone wants to add a pure pulse out (very 
very short on-time) to a Doepfer LFO, get in touch with me 
off-line.  There is a way of doing it using the square out that 
involves two passive components in a simple RC config that 
could by added either under the faceplate  with a new jack added 
(as I did it) or even be incorpotaed into a special patch chord that 
would be used for this conversion only.

best,

P


--- In Doepfer_a100@y..., "code_pig" <code_pig@y...> wrote:
> But did he say (or is it possible) to lower the shift range, or is it 
> more trouble than it's worth?
> 
> --- In Doepfer_a100@y..., "petergrenader" 
<petergrenader@h...> wrote:
> > Got a note from Dieter last night, who saw these listings and 
he 
> > went into the theory behind the Shifter and how it cancels out 
the 
> > second sideband to isolate the sum and difference.  Much as 
I've 

> > always heard its done, by [hase cancellation using the 90 
> > degree output of a quadrit. oscillator.  It's a very complex 
> circuit.

> Got a note from Dieter last night, who saw these listings and he 
> went into the theory behind the Shifter and how it cancels out the 
> second sideband to isolate the sum and difference.  Much as I've 
> always heard its done, by [hase cancellation using the 90 
> degree output of a quadrit. oscillator.  It's a very complex circuit.

Yes, even one look at the board tells you this, and having traced most
of it out, you are not left in much doubt! Since I haven't totally
fathomed it out myself yet, I'd quite like to see the explanation too,
so if you *are* listening Dieter and would care to forward your
response to me also, I'd be most grateful!

Tim

[The views expressed above are entirely those of the writer and do not
represent the views, policy or understanding of any other person or
official body.]

> But did he say (or is it possible) to lower the shift range, or is it 
> more trouble than it's worth?

Having spent some more time studying the circuit over the weekend I
now have a reasonable idea of the way it all works, and I suspect the
answer to your question is probably the latter - more trouble than
it's worth. Adjusting the oscillator to run at lower frequencies seems
feasible, but it brings a few other problems with it: there is a
considerable DC offset on the signal coming out of the oscillator, and
it depends on the control voltage (i.e. frequency) of the osc - at
it's lowest freq its about +2V, at the high freq end its about -2V.
This DC is blocked from the 1496 chips doing the main multiplications
by some capacitors - removing these to allow low frequencies through
would also allow the DC through, and this is likely to upset the whole
1496 set-up (which is pretty hard to figure out even *with* a 1496
datasheet..). Also the amplitude of the osc output appears to decrease
as frequency decreases, therefore there is likely to be a point where
it just stops oscillating reliably. Which is not to say that it is
impossible to lower the lowest shift frequency, just that it would be
non-trivial!

Tim

[The views expressed above are entirely those of the writer and do not
represent the views, policy or understanding of any other person or
official body.]

Tim,

Thanks very much for the detailed explanation, as well as the time 
you've taken on the whole thing.

Regards,
Kevin


--- In Doepfer_a100@y..., "stinchcombe_t" <tstinchcombe@q...> wrote:
> > But did he say (or is it possible) to lower the shift range, or 
is it 
> > more trouble than it's worth?
> 
> Having spent some more time studying the circuit over the weekend I
> now have a reasonable idea of the way it all works, and I suspect 
the
> answer to your question is probably the latter - more trouble than
> it's worth. Adjusting the oscillator to run at lower frequencies 
seems
> feasible, but it brings a few other problems with it: there is a
> considerable DC offset on the signal coming out of the oscillator, 
and
> it depends on the control voltage (i.e. frequency) of the osc - at
> it's lowest freq its about +2V, at the high freq end its about -2V.
> This DC is blocked from the 1496 chips doing the main 
multiplications
> by some capacitors - removing these to allow low frequencies through
> would also allow the DC through, and this is likely to upset the 
whole
> 1496 set-up (which is pretty hard to figure out even *with* a 1496
> datasheet..). Also the amplitude of the osc output appears to 
decrease
> as frequency decreases, therefore there is likely to be a point 
where
> it just stops oscillating reliably. Which is not to say that it is
> impossible to lower the lowest shift frequency, just that it would 
be
> non-trivial!
> 
> Tim
> 
> [The views expressed above are entirely those of the writer and do 
not
> represent the views, policy or understanding of any other person or
> official body.]

Don't know if it's possible (haven't seen a schemo) but thinking 
about the lowering the frequency range of the internal osc, I was 
thinking...why not just have an external input that bypassed the 
internal osc all together.

This probably isn't possible.  Not sure how the quadrature outs 
are generated - or if those four elements are redrived from a 
single oscillator i there somewhere, or if the only outs of the 
internal osc are in quad.  If it's the first option, then I would think 
that (somehow) and external input could be added.

Then you could run it down to any frequency your heart desired.

You could also run the SAME signal into both ins, which on a 
Buchla anyway gives off a very nice effect.

I am assuming this, but I betcha the audio out jack of this unit 
was added so they could conform the module to one of their 
standard faceplate configurations.  It surely ain't gonna be used 
much on my system.  Perfect place for sticking in an external 
input!

daft idea?

Dieter, if your listening (taps on microphone) could it??

-P

Peter,

I'd be very interested in the mod you suggested.  By the way, I've 
tried applying very low (as well as negative) voltage to the CV 
input, but have had no success.  Bypassing the A126's internal 
oscillator might very well be the answer.

Okay, I have to admit that basically I'd love to have that phase-
thick sound of the old EMS Frequency Shifter at a very low shift 
frequency (one of which I used own, and was VERY STUPIDLY was conned 
into selling about 8 years ago).

I'm not saying that I don't like the A126.  Quite the contrary.  The 
A126 is one of my favorite modules, and one I tend to use maybe more 
than I should.

Hey Dieter, and if you're listening, might there be room for either a 
second Frequency Shifter, or a modification to the existing A126?  
There seems to be a bit of interest here (not to mention the fact 
that other companies are coming out with FS's with substantially 
lower shift ranges).  Maybe I'm biased towards Doepfer, but I know 
I'd buy at least one, and I'm fairly certain others on this forum 
would as well.

Many thanks to Peter and Tim.  These two bolt-heads (and that is said 
with complete respect) have helped to clear the otherwise muddy 
waters of the A126.

By the way Peter, the CD of A126 examples was sent (as was yours 
Paul).

Regards,
Kevin

saw this and thought it might be of interest.  home-built freq 
shifter w lo freq capability..

http://home.t-online.de/home/jhaible/hj_fs.html

Jurgen's circuit is of interest, yes. However, I have a general rule:

If there's over 10 analog ICs involved, who in the f@*k am I 
kidding.

This has saved me from countless failed attempts to recreate 
circuits from my past in which I lust and why I stick to the easy 
stuff, like Buchla Low pass Gates and simple voltage 
processors.

Some day you should all try to have a lookey-see at Buchla's 
Frequency Shifter.  What a monster.  Four different supply 
voltages, fhree pages of pure Buchlian-type mumbo-jumbo 
schematic diagrams, notes scribbled all over the place.

But....it had no internal oscillator, just two inputs! and...it used the 
same IC as it's heart as the Deopfer unit - the 1496.

In any event, the Deopfer Freq Shifter has changed my life - 
highly recommended!





--- In Doepfer_a100@y..., "duffnuff" <duffnuff@h...> wrote:

> saw this and thought it might be of interest.  home-built freq 
> shifter w lo freq capability..
> 
> http://home.t-online.de/home/jhaible/hj_fs.html