_[sdiy]_ICL8038 worst ic off all time!

[Scott Gravenhorst]

Hey Karl, Thanks!

=?iso-8859-1?q?Karl=20Dalen?= <karldalen at yahoo.se> wrote:
> --- Scott Gravenhorst <music.maker at gte.net> skrev: > Magnus Danielson
><cfmd at bredband.net> wrote:
>> >From: Scott Gravenhorst <music.maker at gte.net>
>
>> >As I recall it, in the 4046 you have basically two inputs and one of
>> them is
>> >really a current input, so that's your entry into the world. I did
>> exchange
>> >emails with Björn Juhlin back then, and I don't recall exactly but I
>> think I
>> >helped him with that particular part (I might have just caused
>> confusion, since
>> >I was a little hasty to analyse things and made a few errors where he
>> was
>> >right as I recall it).
>
>I think he did that on his own, but there was a later list discussion
>on the matter as i can recall it. However referring to archive and some
>off list correnspondance, the usage was of pin 11 and the later discussion
>where about using pin 12,(Magnus suggestion) however what he didn't say
>(i almost said, "as usual") was "why" he used pin 11 and not pin 12.
>If reading the text really close of the design it speaks about the
>"non linearity" of the internal current Pfet mirror.

Wow, well, apparently the PDF file he sent me wasn't his final version.  What I have uses pin 9 as
a CV input, pin 11 is grounded through 10K and 12 is open.  I found this to be musically useless as
it doesn't track a linear CV very well at all.

>And off list he says:_
>:as much as a shot in the dark one idea are to "use" the current sink fet
>:of pin 11 to compensate for the slight exponentially behavior of the
>:internal current sinks by either statically set an offset voltage on pin
>:9 or dynamically by a continuos voltage on pin 9 controlled by the very
>:same V/octave voltage sent to the input of the exponential converter
>:scaled in one way or the other!? A crude way to use the Nfet of pin 11 to
>:compensate for the Pfet in the mirror!!...No satisfaction guarrentee..!!
>:And a shot in the dark!......

This is interesting, pin 9 goes to the gate of an N MOSFET and 11 is the source of the same
transistor.  Does my PDF represent his idea and that the 10K resistor is the best value to curb the
nonlinearity?  Am I to use pin 9 as my CV input?  I can send the PDF if you need it.

My own idea was to current sink through pin 12 leaving 9 and 11 open, controlling the oscillator's
charge current using the external current sink (and internal current mirror).  But clearly, if
there is a nonlinearity in the current mirror, then this will not work and I did see some awful
tracking, enough to give up on it.

>However, quite interesting idea, could be worth for Scott to try out?

I would like to, but I wasn't privvy to the private emails.

>So, i think he did that pretty much on his own, he just didn't say
>everything or "why" certain things was done. 

This is a problem for me because I'm not an EE and don't really know about the "slight exponential"
nature he refers to.

>Also looking closer to the design he also suggest de franco compensation
>of some sort,And i know he did some pretty complex "sort of de franco"
>compensation by splitting up this slightly expo response of the mirror
>into several smaller expo stages by using resistor dividers on both side
>of the charge cap and some biasing! Quite bizarre as it sound to me but
>he really tried to squeeze out the bananas of that IC to se what could
>be done.

This sounds like a "see if I can make it work at all" challenge, but quickly becomes complex, more
complex than is appropriate given that there are some incredibly simple saw VCO designs that are
quite inexpensive (4069UB is one).  As I had said, I have another idea to use a 4007 and replicate
the MOSFET bridge used in the 4046, then supply a VC current source to power the top of the bridge.
 I would use a 4027 to drive the gates of the bridge transistors.  Ultimately, I think the 4007
would work better with less nonlinearity compensation stuff.

>Great work anyway. And as he says a "neat cheap syncable LFO!!""

Ah, but an LFO doesn't necesarily need to be "linear".

>
>KD
>
>> Hmm, I'd still like to get this working.  Yes, you're right, one pin
>> (12?) is a 
>> current sink pin.  I've designed a standard linear v-i converter that
>> sinks to 
>> ground, but again, I think I was trying to run the current up too high. 
>> The 
>> idea was to allow the v-i converter to control the current by setting
>> the 4046 
>> VCO CV input to the positive rail (thereby opening it up full) and let
>> the 
>> external sink do the work.  I'll have to try this again, but be more
>> mindfull 
>> of the control current.
>> 
>> Something else I've wanted to try is to build the oscillator core from a
>> 4007.  
>> The one thing that stopped me was that the reversing cap charge scheme
>> would 
>> cause the cap voltage to go below Vss and that concerned me.  Then I
>> re-read 
>> the data sheet stuff on the VCO and found that, by design, the cap is
>> actually 
>> discharged through the protection diodes.  I'm thinking that I can do
>> this with 
>> a 4007 as long as the diodes can withstand the discharge current (that
>> means 
>> the cap should be as small as possible).  Using a 4007 this way, I could
>> avoid 
>> the possible nonlinearity of the MOS current mirror that supplies cap
>> charge 
>> current in the 4007.  It's a curiosity, I know I can make a saw VCO
>> other ways, 
>> I just think the design is interesting.
>> 
>> >I haven't heard from Björn for several years now.
>> 
>> Yes, he did disappear.  Too bad.
>> 
>> >Cheers,
>> >Magnus
>> >
>> 
>> =========================================================  
>> - Where merit is not rewarded, excellence fades.
>> - Hydrogen is pointless without solar.
>> - What good are laws that only lawyers understand?
>> - The media's credibility should always be questioned.
>> - The only good terrorist is a dead terrorist.
>> - Governments do nothing well, save collect taxes.
>> 
>> -- 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/
>> -- Autodidactic Master of Arcane and Hidden Knowledge.
>>  
>
>Höstrusk och grå moln - köp en resa till solen på Yahoo! Resor på 
>adressen http://se.docs.yahoo.com/travel/index.html 
>

=========================================================  
- Where merit is not rewarded, excellence fades.
- Hydrogen is pointless without solar.
- What good are laws that only lawyers understand?
- The media's credibility should always be questioned.
- The only good terrorist is a dead terrorist.
- Governments do nothing well, save collect taxes.

-- 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/
-- Autodidactic Master of Arcane and Hidden Knowledge.