Env vco!

[Magnus Danielson]

>>>>> "B" == BJ  <zzynt at swipnet.se> writes:

Hi there!

 >> BJ, you forgot to put a value on R11. I suggest a 100k.

 B> Sorry i forgot, the value depends on the range you want to use.
 B> 100k is fine. 

Right, but as things happends 100k is also a very good value for the
linearity. The linearity problem is partly due to the input FET, so
you may acheive better result by hooking in directly to the current
input, then you are pulling current directly from the current-mirror.
The linearity that the 4046 has is acceptable for most of the PLL
applications that it is intended for.
 
 >> Also, why don't you let the cursuit take more of all that space?
 >> I try to scale up my schematics when they don't take that much space.

 B> Because im lacy. :)

OK, I'll by that ;)

In my Postscript schematics I have a scale factor, so it is a small
thing fixing it compared to all the work getting the schematic in.
But then, my way of doing schematics is a bit braindamaging in the
first place, so I better have some freedome as a benefit ;)

 >> 
 B> And the MC202 and SH101 envelopgenerator and a simple VCO are found at:
 B> http://home8.swipnet.se/~w-88660/schpcb.htm
 >> 
 >> For those that wishes to use an exponential cursuit with the 4046 VCO
 >> may hook it in on the current input of pin 12. This input wishes to
 >> see a current being drawn towards ground, so hooking it to the current
 >> output of a standard dual-transistor expo-cursuit (like the ASM-1 for
 >> instance) should go well.

 B> Yes you may do so, i have , but the linearity is no good, One may use
 B> the HC7046 who has even better linearity, i think it was approx 0.15%.

Hmm... was there a need for high-end compensation maybe?
 
 >> Also, you may still use the VCO input for a real linear input (pin 9).

 B> Yes again, this is good, i dont have to explain my own designs 
 B> Magnus does it wery well indeed, great i can keep doing my 
 B> lacy ,sleepy things here! :)

Well, since you didn't someone had to ;^)

 >> The old CMOS versions of this VCO would scream up to 0.8-2.4 MHz
 >> (depending on powersupply), the newer onces get even higher. So, to
 >> get a pretty good responce one should consider pretty quick op-amps.
 >> The reset time is quite small, but if one would find it a trouble,
 >> then you could allways add the Franco compensation resistor in series
 >> with capacitor.
 
 >> The upside about using the 4046 as a core for a VCO is that you get
 >> the reset-cursuit for free as well as phase detectors to make PLL
 >> loops, in case you would fancy that game. BJ reused one of those to
 >> make a MS-20 style ringmod.

 B> Yes again, wery nice thanks for the "going thru" my designs, ooh i think 
 B> i have a cup of teee, and a smorgas while you other listening to Magnus
 B> .:) 

Funny, I was just about to get a cup of tea and a smorgås too...
 
 >> Oh, be carefull about loading the capacitor, use high-impedance
 >> op-amps (about 1 Tohm). 

 B> Yes that right, or one may use FET's as buffers.
 B> The frequency will shift down a bit if the cap is loaded to much.
 B> When i did this design, it was for a pecussion synt voice,
 B> I was just measuring around, and foundout about the charging
 B> phases for the VCO and saw then the possibility to have a SAW wave.

Also, a little warning: When the 4046 flips, it will earth the
capacitor end that is positive, thus making the negative end going
below 0V (to -2.5 V in BJ's 5V design), so either is op-amps on a
differential supply (which the -5V feed to the summing op-amp needs
anyway) or (when you have the op-amp in +5V/0V single feed) you must
have op-amps that does not go crazy for the input to go below the
negativ rail.

 B> Pitty that the EN pin dont reset the charging phase of the cap!

Certainly, but if you add a CMOS (4066 or something) switch over the
cap, you may get a reset anyway. Another method would be a pair of
FETs on each side reseting to earth.

 B> Regards
 B> Haddock!

Cheers,
Tintin