[sdiy] Using both OTAs in 13700 as one OTA

Sun Feb 28 18:41:58 CET 2016

Hi Steve,

I don't use window, Mac OSX here...  But hey, I'm not an engineer (my
partner Vinicius Brazil is, he does the precision testing on our
actual products). My ears are all the test equipment I really need for
my sdiy musings, since this is all about sound anyways. Thanks though.

Right now I'm trying to solve the THUMP from CV bleedthrough with fast
attack and decay settings, via a biasing trimmer at the audio input,
or a cap to ground (which slows response down a bit, but maybe I can
find a good compromise).  This may be where Aaron's original idea
might work, as the inverted OTA's might cancel out the DC offset.  I
think I'll try breadboarding it.  Aaron, did you ever test your idea?

Cheers everybody

On Sun, Feb 28, 2016 at 7:34 AM,  <sleepy_dog at gmx.de> wrote:
>
> Assuming you use windows, if you have a not too shabby ASIO soundcard, it
> means a program can get sole ownership of the card and windows mixer not
> polluting your output signal. So you could do audio range SNR and THD
> measurements with looping the soundcard output, going through your test
> hardware, back to the input, and using appropriate software. In the confines
> of the distortion & noise of the soundcard used itself, obviously, but I
> suspect that DIY analog synth circuits will be noisier than a decent ASIO
> soundcard / external USB audio interface, my Roland/Edirol UA-25ex served me
> well enough.
>
> Other than using existing software, you could implement it yourself, I
> forgot where, but there are some app notes by TI or National explaining how
> to use an FFT and generated test signals to measure this. I have done it,
> but am not sure how well I did it as to let my crappy little program loose
> on you :-D Unless you insist, hah.
> It was done on one weekend, though, using FFTW library C# wrapper. So not
> too difficult. If you are into software dev that is.
>
> There is a better option: one nice, free of charge program which does all
> those things, and generate frequency response curves etc - If you are
> interested, I will ask my colleague next week, because I forgot its name,
> alas.
>
> You would need some level conversion for testing synth hardware that way, I
> guess. At least protection for the soundcard input. But (warning, I'm an
> analog semi n00b) I guess that operating typical synth circuits with inputs
> of only +/- 1V or whatever comes out of a soundcard might not do the noise
> figures justice?
>
> I'm sure some of the experienced people here can comment on that, how to
> best do this. I'd be interested myself, for I have not build any test bench
> for synth DIY circuits in this regard (interfacing with PC soundcard *the
> proper way*) myself, but would like to.
>
> As for many circuits with LM13700 not using the diodes, I would guess that
> many were designed for the CA3080 and when that became unavailable, people
> dropped in the 13700 as-is? (please do correct me ;) )
>
> Steve
>
>
>
> Am 28.02.2016 um 09:13 schrieb Quincas:
>
> Hi guys
>
> This thread got me thinking a lot about LM13700 VCAs, so I went to the
> datasheet and noticed they suggest using the linearizing diodes to reduce
> harmonic distortion. So I breadboarded it, changed some values to fit synth
> levels and added a voltage to current converter as seen in Thomas Henry's
> VCA1. I don't have adequate tools to measure THD or SNR, but it sounds dead
> clean to me! No noticeable noise and no audible distortion using the nice
> sine wave from my Norton filter in self oscillation. Is there a reason none
> of the designs discussed above use the linearizing diodes? Here's my shabby
> hand drawn schem....
> http://s1070.photobucket.com/user/Quincas_Moreira/media/image_zpsu5cm6fz8.jpeg.html
>
>
>
> Sent from my iPad
>
>> On 10/02/2016, at 6:57 a.m., Simon Brouwer <simon.o at brousant.nl> wrote:
>>
>> Hi Mattias,
>>
>> I mean the circuit like this:
>> http://www.bernacomp.com/elec/lockbox/lockbox_vca2.gif
>>
>>
>> So, no.
>>
>> Best regards
>> Simon
>>
>>
>
>> > Op 10 februari 2016 om 12:31 schreef Mattias Rickardsson
>> > <mr at analogue.org>:
>> >
>> >
>> > Simon,
>> >
>> > what about the temperature dependencies in the transistor case?
>> > Is that control circuit more sensitive for temperature variations?
>> >
>> > /mr
>> >
>> >
>> > On 10 February 2016 at 09:08, Simon Brouwer <simon.o at brousant.nl> wrote:
>
>> > > Yes it was.
>> > >
>> > > However, this circuit still uses one opamp, a resistor and at least
>> > > two
>> > > diodes.
>> > > The more straightforward approach adds two resistors, and replaces one
>> > > diode
>> > > by a
>> > > transistor.
>> > > So this doesn't add much in complexity and cost, and you can control
>> > > both
>> > > halves of an LM13700 independently.
>> > >
>> > > Best regards
>> > > Simon
>> > >
>> > >
>
>> > >> Op 9 februari 2016 om 20:20 schreef Mattias Rickardsson
>> > >> <mr at analogue.org>:
>
>> > >
>
>> > >>
>> > >>
>> > >> Aha!
>> > >> It *was* a beautiful trick. :-)
>> > >>
>> > >> http://electronotes.netfirms.com/AN399.pdf
>> > >>
>> > >> /mr
>> > >>
>> > >>
>> > >> On 8 February 2016 at 19:09, Ian Fritz <ijfritz at comcast.net> wrote:
>
>> > >> > The reference to the original Jung article is JAES, 23, 1975. This
>> > >> > method
>> > >> > was widely used in EN projects. I looked at using this method with
>> > >> > the
>> > >> > LM13700 to aid people who want to build off of these projects, but
>> > >> > who
>> > >> > can't
>> > >> > obtain the CA3080. It's also an easy way to improve S/N by 3dB, as
>> > >> > pointed
>> > >> > out by Serge T. in his EN article from when the chip was new.
>> > >> >
>> > >> > I couldn't wade through Aaron's article either. It sounded like
>> > >> > negating
>> > >> > then subtracting? Isn't that the same as adding?
>> > >> >
>> > >> > Ian
>> > >> >
>> > >> >
>> > >> >
>> > >> > On 2/8/2016 2:41 AM, Mattias Rickardsson wrote:
>
>> > >> >>
>> > >> >> On 4 February 2016 at 05:52, <ijfritz at comcast.net> wrote:
>
>> > >> >>>
>> > >> >>> Good grief! Just connect them directly in parallel. An example,
>> > >> >>> complete
>> > >> >>> with the W. Jung trick:
>> > >> >>>
>> > >> >>>
>> > >> >>>
>> > >> >>> http://electro-music.com/forum/viewtopic.php?highlight=jung+ota&t=50884
>
>> > >> >>
>> > >> >>
>> > >> >> What does it do? What is the "trick"?
>> > >> >> I feel like one of the answers in that thread:
>> > >> >> "I hadn't heard of this trick before, is there an article
>> > >> >> somewhere
>> > >> >> that explains the benefit of drawing bias current from V-? I feel
>> > >> >> like
>> > >> >> there's something obvious that I'm missing."
>> > >> >>
>> > >> >> I spontaneously found Aaron's idea potentially better than
>> > >> >> nothing,
>> > >> >> but I haven't plunged into it.
>> > >> >>
>> > >> >> /mr
>> > >> >>
>
>> > >> >
>
>> > >> _______________________________________________
>> > >> Synth-diy mailing list
>> > >> Synth-diy at dropmix.xs4all.nl
>> > >> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>
>> > >
>> > > _______________________________________________
>> > > Synth-diy mailing list
>> > > Synth-diy at dropmix.xs4all.nl
>> > > http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>> > >
>
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at dropmix.xs4all.nl
>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>
>
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>

-- 
Quincas Moreira
Test Pilot at VBrazil Modular