[sdiy] MC1495 multiplier chip discontinued completely

[Czech Martin]

Well, like always in life it depends on what you want.
It is also a question of taste.

The "real ringmodulator" has a dirty sound due to many
shortcommings, S/N, nonlinearities, transformer bandwidth
etc. If you want that sound, perhaps you try one with a few
Ge or Si diodes and op amp balancing (to avoid costly
transformers) or transformer balancing.
It has something. I still do not know if the transformers
are really necessary to have "that" sound.

On the other end of the scale we would have the high S/N,
low distortion device, for example to use with piano sound
with some carrier. S/N is really important here, as well
as carrier suppression, also nonlinearities should be low,
in order to have a very clean, soft sound.
I can not describe it better.

I have both, one as transformer/diode circuit
and the later as software (with oversampling, so no 
problems due to Nyquist folding back). And a third,
which is discrete, has low noise, but more distortion, compared
to the software solution.

Some people think that audio modulation sounds "agressive",
"mad" , and that it is "against all intentions of music".
I think this is much due to the shortcommings of imperfect
solutions. The software modulator experiment showed to me
how -well- soft such modulation can sound if it is done
right.

The AD633 is expensive, but will do a good job for most people.
It is available. The downside is clearly noise, be it by technology,
transistor geometry or simply too few bias current.
S/N is somewhere between 70-80 dB, this is not much compared to
16 Bit recording, and we're heading toward 20 effective bits or so.

The 1496 is built for square wave carrier. I do not know 
if an external linearisation circuit can be applied nor
what it means in terms of S/N and distortion.
Some people on this list do not like these anti-distortion
circuits, at least for VCA and VCF applications.
The 1496 is available and it's use is wide spread in
all kind of vintage devices. This does not imply that
it is good, of course. If have to look for that, still.

The Fet sign changer is simple, but only useable for "effect" sounds.
I.e.: very agressive modulation, basically the same as a square wave
driven real multiplication device.
So it's use is very limited. Those "soft" sounds I have mentioned
are not available at all.

The 13600/13700 do have the problem that they are only two quadrant.
You can overcome this with using two vca devices, each for two quadrants
will make four quadrants. The problem is then, like in B type output stages,
the crossover distortion. This kind of distortion will not go away
if the signal gets smaller, so no good prospect.
Perhaps there is a possibility to have a tricky bias scheme, 
similar to A/B or A operation of output stages. However, a lot of mirrors
are involved, much more circuitry than the normal gain cell
has, I do not like that idea.

The problems of symmetry also applies to the last one, RC4200.
They are still around at Farnell, but I need to try them.

The discrete gain stage I made (BC549) is of course always possible.
http://www.geocities.com/SoHo/Museum/4459/circuits/ringmod.html
shows a (not too good) schematic.
The only problem with good beefy biasing is residual distortion
due to transistor tolerances and also carrier feedthrough due
to offset problems. With glue and expoxyd these thermal problems
could have been an order of magnitude lessened, also
dual transistors will help a lot.

One thing is also important: modulation should be possible
far out of the audio band, so 100kHz at least.
Using two such modulators gives you the oportunity to 
plug some modules together for frequency shifting
or other wierd stuff.

As I said, it depends on what you want.
I think for me the discrete design will most likely win at
the end. But perhaps there are surprises.


m.c.

-----Original Message-----
From: Scott Gravenhorst [mailto:music.maker at gte.net]
Sent: Donnerstag, 24. April 2003 17:12
To: synth-diy at dropmix.xs4all.nl
Subject: Re: [sdiy] MC1495 multiplier chip discontinued completely


And I *still* have not built an audio ring modulator.

I have a need for a very basic ring modulator, I have as choices: 1496
and AD633.  I also have LM13600 and LM13700.  Of these, what is the
best choice for this?  I don't plan to use it in a pitch shifter, just
for timbre modification to get metallic sounds.  Will one of these
work well enough for this purpose?

Also, on the other end of the scale are really poor multipliers, like
using a single FET which I guess is a one quadrant "multiplier".  Very
nonlinear, but does this have viable uses?  After all, if one can do
it to square waves with an XOR gate, an FET should be useful for
*something*.  

Watching this decay into "unobtainium" is depressing...

>From: "Czech Martin" <Martin.Czech at Micronas.com>
>> I just learned from Arrow/Spoerle that the MC1495 multiplier chip
>> is discontinued and they can not ship any more.
>> On Semiconductor has accordingly made an update to their
>> web site. 
>> 
>> So I guess the 1495 is *really* gone.
>> 
>> The good thing about this chip was that it contained more
>> or less matched transistors for the gain cell, so offset and offset
>> drift was under control. I.e. carrier suppression was good.
>> Compander circuits will not make a good job for this carrier
>> suppression, because it is not a wide band noise problem,
>> but very narrow, so masking is poor.
>> 
>> And it had an input linearisation circuitry for the 
>> carrier input, in order to have quite linear behaviour.
>> The (still available) 1496 is designed for square wave
>> carrier and has of course no input linearisation.
>> So you have to use low carrier amplitude (noise)
>> or accept excessive sidebands (distortion, intermodluation).
>> Also the 1495 allowed to taylor the gain cell bias current.
>> For good S/N you need a lot of current.
>> Most other integrated multipliers do not allow for bias
>> current change and are designed for low power consumption,
>> so S/N of 80 dB will be the upper limit, and most devices
>> will not even give that (AD633 for example).
>> For some applications this is still too much noise.
>> 
>> I want low noise and linear behaviour (smooth sound).
>> So next I try the RC4200 (?) and the 1496 with external DIY
>> linearisation stage. If this doesn't work I'll go back to
>> a complete discrete design, perhaps using a couple of
>> japanese dual NPN, glue, expoxyd ... whatever it takes
>> to cope with the offset problems...
>> 
>> 
>> m.c.
>> 
>> 
>> 
>

=========================================================  
- A smoking gun is one that's already been fired.
- That which gets rewarded, gets done.
- What good are laws that only lawyers understand?
- Government: The other religion.
- The media's credibility should always be questioned.

"When you see a rattlesnake poised to strike, you do not 
wait until he has struck before you crush him." 
  -- FDR

-- Scott Gravenhorst | LegoManiac / Lego Trains / RIS 1.5
-- Linux Rex         | RedWebMail by RedStarWare
-- FatMan: home1.gte.net/res0658s/FatMan/
-- NonFatMan: home1.gte.net/res0658s/electronics/