ETI VCO (4600, 5600 etc) schematic and q about CMOS switch

Ricko

Here is the schematic for the ETI 4600 VOC. It is a linear, triangle-core, through-zero VCO with tri, sine, saw, ramp and pulse outputs. The version shown here is actually the one from the Maplin 5600, which has sync input and output as well.

I am rebuilding my ETI 4600 and I have found that replacing all the 301 op-amps which are not used for positive amplification improves the sound a good amount. (Because the decoupling capacitors have the wrong values, as I wrote in a previous post.) For example, replace IC 1 and IC 2 with TL081 and IC 7 with TL071.

Anyway, I am now stuck sourcing a 4416AE quad CMOS analog switch for the fourth VCO. It seems that quad switches have a fairly standard pinout for the 4*16 range, but different specs and generations of technology. I cannot find a datasheet for the 4416 (I think Maxim made them, perhaps SCL and Fairchild?)

Does anyone have any ideas about what modern 4*16 can be substituted?

Thanks, and I hope the VCO schematic is interesting. (The boards are still available, by the way, from RCS Australia, for about $20 a board IIRC.)

Cheers
Rick

eti5600-vco.JPG

Description:

VCO from Maplin 5600: same as ETI 4600 + sync

Filesize:

190.59 KB

Viewed:

903 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

yusynth · Joined: Nov 24, 2005 Posts: 1314 Location: France

Ricko · Posted: Mon Jun 16, 2008 6:23 am Post subject:

That is what I thought too.

The input to the circuit is usually +/- 5V, and therefore the output of IC2 is usually positive.

As you say, if the input is negative, it is easy. But since the input is usually positive, that is a little more interesting. The documentation speaks of the diodes being there for "excessive voltages" (i.e. for large positive inputs.)

(The circuit definitely works, by the way!)

eti5600-vco-info.JPG

Description:

Filesize:

197.23 KB

Viewed:

246 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Ricko · Posted: Mon Jun 16, 2008 6:36 am Post subject:

Damn, I think your ae right it isn't though-zero, but not because of the diodes. I don't see where the logic of the +ve and -ve selection switch IC3/1 and IC3/2 flips to match any flipped outputs of IC1 and IC2 for a negative input: if it doesn't flip the oscillator will just lock up.

I'll go home and test it!

yusynth · Joined: Nov 24, 2005 Posts: 1314 Location: France

Ricko · Posted: Mon Jun 16, 2008 7:03 am Post subject:

Something like that, yes.

What about this logic: in order to convert this to a through-zero, we would have to ensure that the inputs to the switch are always negative at the top and positive at the bottom. In other words, full-wave rectification... does that make sense?

This could be done by a diode bridge after R11 and R12. That sounds scary: the linear CV is very sensitive to non-linearities.

Another approach would be to have a non-inverting op-amp with a gain of 2 in parallel with IC1, but with a diode in series with its input, blocking the usual case of positive voltages. So the sum of the outputs is the absolute value of the CV input. That would give no non-linearities on the normal operation (e.g. for normal non-FM use, but provide some through-zero for FM use, would be the theory.)

If this is not complete rubbish, it would work on any linear VCO. (Mathematically, the negative CV wave might be out-of phase, but I don't now that it matters.) In fact, you could build it in after the linear-to-expo converter of a conventional VCO design, I guess. It seems an easier method than the switching that you see in some designs.