SA602 ring modulator?

[christian]

Has anyone had any experience with this chip? SA/NE602?

I came across this one, but the datasheet was kinda vague. Its basically like 1496 gilbert-cell along with other circuitry to provide onboard HF oscillator etc..
It should be low power, max Vcc 8 volts unlike 1496 and other modulators that work best with lot larger voltage supply.

Then I read about that chip on latest ARRL handbook which gave better block diagram and the oscillator pins could probably be used as linear oscillator inputs.

Though about making this "deluxe frequency analyzer" with this chip and the xr2206 chip for VCO (needs more power than 9v though so I'll need either dual battery, yuck, or 12vdc input, another yuck), along with LFO.

Then again, DOD gonkulator uses 1496 with 9v succesfully..

ch.

[R.G.]

That one always interested me, too. I playes with the NE602 a bit. I never could get it to be sufficiently quiet when not playing through it, even tweaking the thing.

It's primarily a RF part, and I think the low frequency operation is good enough for RF where  you always have follow up filters to remove spurious stuff, but won't quite make it at audio where  you need -80db or better carrier suppression in quiet passages.

I always thought that I could apply a noise gate to the output and get it to work OK, but ran out of time back when I was working on it and shelved it.

[christian]

It's primarily a RF part, and I think the low frequency operation is good enough for RF where  you always have follow up filters to remove spurious stuff, but won't quite make it at audio where  you need -80db or better carrier suppression in quiet passages.

This was my primary concern, that its supposed to be used on RF, and with lower voltages and currents the gilbert cell wont also act that well.
I never had 1496 to try, but I'd be interested in how well that works with 9v like in that gonkulator. I'd assume pretty badly?

ch.

[R.G.]

Signetics has been making the 602 so long and have their process tweaked in so well that they no longer do functional testing on the chips, only DC tests. I found that out from one of their semiconductor process engineers at a seminar a few years back. They really don't know how it works at RF anymore except that people buy them and are happy with what they get. Oh, I'm sure they still periodically sample, but there is only DC testing on the production line for them.

The 1496 is a good chip for this. The problem with the 1496 is just that you run out of voltage for the transistors inside when you only have 9V, and so the signal range before you get signal clipping is limited. This may be OK if what you're trying to do is make really strange sounds.

If you up-vert 9V to over 15-17V with a charge pump, you have enough voltage to run one pretty well.

My experience with ring modulators in general is that they all need a noise gate to get rid of the modulation whine unless they have been freshly nulled out. Temperature, drift, phase of the moon, etc. all let it drift off null. Probably the right answer is to use a gate with them and not rely on the modulator to null the modulation out.

[christian]

My experience with ring modulators in general is that they all need a noise gate to get rid of the modulation whine unless they have been freshly nulled out. Temperature, drift, phase of the moon, etc. all let it drift off null. Probably the right answer is to use a gate with them and not rely on the modulator to null the modulation out.

Add that to the fact that ring modulation is pretty special effect that you aren't using all the time (unless you're in Devo) if at all, you may not want to bother.

This sounded like I'm replying to myself.. I'm not bi-personal or anything

ch.

[Paul Perry (Frostwave)]

My commercial Frostwave Blue Ringer ring modulator stays nulled, at least for five years or so. And it is reasonably quiet.
BUT..... I'm using regulated + and - 15v rails and an AD633 as the multiplier. Plus I advise turning up the input to maximise the signal to noise ratio. And layout is hyper-critical, if you want to avoid any carrier bleedthrough. Forget about loose wiring for stuff like this... it's all PCB mount pots.
BTW here is something that drove me mad while prototyping: any hum on the input, makes it seem (at least to my ear) that there is bleedthrough during the silences. Because, say the osc is at 1000Hz, then 50Hz hum on the input gives you 950 and 1050 out, and as you rock the local osc control, it sounds like bleedthrough of the 1000Hz!!!
How good can a passive ring mod be?? with careful matching of diodes, and again care in layout, you can get the carrier 60dB down, accoding to someone who should know (and has more patience than I do!). I guess a passive ring mod isn't going to drift much, so there is a point in is favor.

[R.G.]

I can believe that. Using stable power supplies, a carefully optimized multiplier chip and very careful layout are going to get you quite a ways toward proper operation. It's a lot of painstaking work, and my hat's off to  you if you got it stable enough. That's saying something.

[R.G.]

I meant that it's saying something notable that you got it to work, not that my hat's off. I lose my hat a lot. 8-)

[christian]

BUT..... I'm using regulated + and - 15v rails and an AD633 as the multiplier.

Isnt this chip discontinued now? EH used AD533 in old FA, what are they using now?

BTW here is something that drove me mad while prototyping: any hum on the input, makes it seem (at least to my ear) that there is bleedthrough during the silences. Because, say the osc is at 1000Hz, then 50Hz hum on the input gives you 950 and 1050 out, and as you rock the local osc control, it sounds like bleedthrough of the 1000Hz!!!

Yep, this is specially for guitar players, better use mic scene that cancels the hum. Strats hum like hell with just their bridge mic on

Paul, I read a few weeks ago about your +/-15v supply and you said you use 14v transformer? Can you do that? Shouldnt the incoming voltage be higher than outcoming in regulator? Or does it have something to do with ripple voltage?

ch.

[Paul Perry (Frostwave)]

Digikey & Jameco still have the AD633 in DIP (about 8 or 9 bucks).

AS for the power supply, 14V AC mains rectifies to a peak of 14x1.4 = 19.6V, plenty of headroom for the 15v regulators (require 2.5 v or so margin).

[Jaicen_solo]

Hey Paul, sorry to hijack this thread but I was wondering if I could pick your brain a little regarding your ringer.
I'm working on a ring mod design using AD633's, simply becuase I have a few. I had planned to use the 1495's since they're readily available but hey.
Anyway, what I was wondering is wether you've ever done any work on different carrier waveforms, for example sawtooth rather than sine. I imagine the harmonics will cause considerable bleedthrough, but how does it sound? So far i've got my modulator modulating but i've been using external signals rather than onboard oscillators. Any tips for cleaning up the bleedthrough?