Best/simplest non-MN3102 chorus/flanger clock circuit?

[ElectricDruid]

Hi all,

I've been looking at flanger and chorus clock circuits. I'm hoping to avoid the MN3102, so I've eliminated circuits that use those. That leaves various designs. Here's two:

1) 4046 PLL/VCO
There are several flangers that use the VCO from the 4046 as the high freq clock. Since it doesn't have biphase outputs, it is often followed by the 4013 flip-flop (which does).

2) 4047 astable + 4007
The 4047 astable has biphase outputs on the chip (it has a built-in flip flop), so that's handy. Unfortunately it doesn't have a CV input, so that needs faking. This mostly seems to be done by using one of the mosfets (is that what they are? I'm not big on transistor ident..) in the 4007 across the frequency-setting R value. This seems pretty simple, but I don't understand why you'd use a whole chip for just one transistor. Why not just use a mosfet instead?

Are there other examples you've seen? What else could I use for a high-frequency VCO with biphase outputs?

(I'm thinking of a chorus using one of my PIC LFO chips for multiple waveforms, but I need a VCO to feed it to).

Thanks,
Tom

[KarenColumbo]

*fetches popcorn*

[Scruffie]

You can set the 4046 up to have the two outputs, look at the Hollis Zombie Chorus & Ultra Flanger.

[ElectricDruid]

I've just realised the EHX Small Clone uses the 4047 as a clock too, but it doesn't bother with the 4007 to control it. Instead, it just feeds a voltage into pin 3. This is beautifully simple, so I might give it a try. It looks like this:



Tom

[Fender3D]

Hi Tom,

there's an interesting 4046 application paired with TDA1022 BBD (then dual clock signal):

here

[anotherjim]

I don't know exactly why 4007 is chosen over a discrete MOSFET or JFET, except that maybe it's Vt is more consistently amenable to the LFO sweep  or assumed to be by the designer. R Penfold published a few designs with 4007's used for just one Q in a VCR role, so maybe history is to blame.
If we want, we can make use of the rest of the 4007. Millennium bypass, tails gating or buffers - I know I would.

[ElectricDruid]

<4007 offtopic interlude>
So what's your top tip for what to do with the two transistors in the package that have one leg tied to the supplies? What does one do with those?
Also how do you get round the fact that pairs of them are tied together? What's that about? (Presumably someone thought it was a good idea for some intended use or they'd have spent the money on extra package pins?)

</offtopic>

T.

[anotherjim]

Left pair can only be either a single inverter or make a P or N common source amplifier. Cannot have source degeneration or source follower usage (unless we play with the chip power pin feeds).

Middle pair can be VCR, Inverter, Common Source, Source follower, Complementary DPST transmission gates.

Right pair same as middle but forces spdt operation of transmission gate due to common drain connection.

There's nothing to be done about the common gate to P & N. The drain or source that may be unused can be left floating.
Source or sink current less than discrete MOSFET.
No parasitic source-drain diode like a discrete Mosfet (so not always shown in the diagrams - parasitic D's go to Vdd or Vss instead)

...if only it was a 16 pin with a least 2 of those gate pairs separated. I have no idea why that wasn't pursued.

[Fender3D]

So what's your top tip for what to do with the two transistors in the package that have one leg tied to the supplies? What does one do with those?
...

You can get a dual gate (for signal and eventually feedback) + bypass system just adding a comparator.
Check TC Electronics XII schematic (I uploaded it on the other forum)

[ElectricDruid]

Can anyone tell me about the extra power supply circuit that EHX used for the clock and the BBD on the Small Clone?

Shown here, bottom right, the "VD" supply derived from VC by Q3 and associated parts:



What's the need for that? Is it just an attempt to keep clock noise from the rapidly switching logic out of the main power rail?

Thanks,
Tom

[bool]

It's a simple "capacitance multiplier". Most probably both to filter the BBD circuitry supply - and to isolate the audio supply from BBD/OSC.

[ElectricDruid]

Thanks Bool. I've never come across that little snippet before, but I knew someone here would know what it was.

The most interesting thing about it for me is that it drops the 9V supply by another 0.65V across the transistor, so we've got a best case of about 8.35V. That's getting significantly outside the spec'd supply voltage range for the MN3007. Although apparently it works - but don't rely on those datasheet noise figures at that voltage. My bet is they degenerate by ooh, I dunno, say 6 or 7dB to 73dB instead of 80dB (ahem! MN3207)

T.

[DrAlx]

Not a fan of daisy chaining the three supply lines (VB,VC,VD) like that.  Ticks in one section can affect the others
plus you can have other unwanted effects. For example if one of the supplies (like the VCO) draws more current when running at high rate
then it will cause a drop in the other supply voltages (because of the daisy chain) and that can affect other things (like BBD bias).
I prefer having the 3 supply lines in parallel to each other, each with their own RC filter direct to the main supply.

[ElectricDruid]

+1 agree, Doctor.

On the DigiDelay, I had two parallel voltage regulators complete with their own caps in an attempt to keep digital noise out of the analog parts.

Tom

[Aph]

I've just realised the EHX Small Clone uses the 4047 as a clock too, but it doesn't bother with the 4007 to control it. Instead, it just feeds a voltage into pin 3. This is beautifully simple, so I might give it a try.

There are three main circuits I've seen for a 4047 VCO that I've breadboarded: diode, jfet, and 4007. And that's in the order from worst to best (IMO). The 4007 approach is by far the smoothest sounding with the widest range. However, I was using a flanger circuit, so the diode one is probably fine for a chorus.
I assume the mosfets on the 4007 are just more consistent as opposed to discrete?