tri lfo with 90deg lag?(blush)

[Harry Bissell]

Disclaimer: heresay evidence...
I haven't worked with these chips much. My understanding is that the LM13600
has bias current comp for the darlington, which will extend the range but can
cause control voltage "step" or bleedthrough... The LM13700 doesn't have it. If
you look at the current drive for the OTA, you will see the extra tranny.  In
VCO apps you probably wouldn't notice the step as it would probably be masked
by the change in oscillator frequency, or just integrated out as a little extra
transient error. In a VCF or VCA you might hear the thump. The fet buffer seems
to be the favored solution for current starving in the buffer... :^) Harry

Grant Richter wrote:

> Using a LM13600 darlingtons directly without a FET buffer,
> the output will swing negative about six volts when
> the control current get below a few mocroamps.
> I believe this is due to current starving in the darlingtons.
>
> To get the maximum range, maybe use a FET (2N3819) before the
> darlington buffer. This lets you drop the output impedance
> of the darlington buffer lower (from 10K to 2.2K maybe?)
>
> BTW we cracked open a Big Briar twelve pole phaser.
> Beautifully designed, as is all Bobs later work.
> The transconductors are LM13700, which Bob also
> used on the Etherwave. He obviously prefers the buffers
> without bias current modulation. Although, as I understand it,
> the bias current modulation was supposed to extend the range
> of the buffers.
>
> I built a filter circuit using the LM13600 in a four pole lowpass.
> Basically the Electronotes standard four pole.
> As I mentioned, the output swings negative for very small control
> currents. I substituted LM13700s in the same circuit and they
> did not behave any differently. So I am confused as to the difference
> between them and what it gets ya.
>
> Anybody worked with these enough to know?
>
> Thanks,
>
> Grant
>
> ----------
> > From: Stewart Pye <stew at uq.net.au>
> > To: Paul Perry <pfperry at melbpc.org.au>
> > Cc: synth-diy at mailhost.bpa.nl
> > Subject: Re: tri lfo with 90deg lag?(blush)
> > Date: Friday, July 30, 1999 12:33 PM
> >
> > At 09:44 PM 30/07/99 +1000, Paul Perry wrote:
> > >Today I seemed to remember somethng slightly familiar fron the
> > >National Semi LM13600 data sheet, an oscillator with 3 OTA
> > >stages giving 60 deg in each stage.. which made me think if there were 2
>
> > >stages there would be 90 deg difference.. which made me thnk about what
> > >my LP filter (clone of the LM13600 Korg MS20 filter) must be doing..
> > >got out the scope (thanks Barry!) and viola, there it is, when
> > >it oscillates one of the OTA stages is 90 deg behind the other..
> >       <SNIP>
> >
> > >BTW the outputs aren't perfect sine/cos waveforms, but plenty good
> enough
> > >for LFO panning applications. And, putting a fet before the buffer stage
> > >(thanks JH) should give a nice big range..
> >
> > I presume the FET buffer would increase range when used as a filter also.
> > Correct? So, when building a 13600 based filter (AKA MS20) Do you think
> it
> > would be a good idea to use a fet before the darlington buffer. Also,
> does
> > anyone know if this arrangement would be lower noise than a FET input op
> > amp, TL072 for example? And ultimately is this (fet+darlington buffer)
> the
> > superior arrangement?
> >
> > Regards,
> > Stewart