Quietest Lm13700

Started by Ed Rembold, September 27, 2003, 03:06:57 PM

Previous topic - Next topic

Ed Rembold

If any of you vetrans has ever done a noise-test shootout with the various "brands" of the LM13700, and have an opinion as to which is "quietest". I'd love to hear about it.  (I bought some JRC and could not believe how noisey they were!)
Thanks, Ed R.

Mike I.

Hi Ed,
What other manufacturers 13700 were you comparing with (presumably National)? Are they all noisy? OTA's have never had a reputation for being quiet... the on-chip  linearizing diodes help a bit.
Regards, Mike

gez

I haven't done a 'shoot out', but I have one made by National Semiconductor on the breadboard at the moment and it's exceptionally quiet.

I 'think' it might be CMOS (part of it anyway) as it came in a anti-static pack if I recall (it's been on the board so long I could be wrong about this).
"They always say there's nothing new under the sun.  I think that that's a big copout..."  Wayne Shorter

Mike I.

Ed,
The noise specs could be degraded if the peak-peak signal swing between the (+) and (-) inputs ever exceeded 5 to 7 volts... same as with any normal bipolar transistor... emitter-base reverse breakdown should be avoided. The OTA uses bipolar transistors at the inputs. In normal operation without the linearizing diodes connected you don't want the peak-peak level to be much larger than 20-25 millivolts if distortion is to be avoided - with the linearizing diodes this can be increased to about 100 millivolts peak-peak... and then hard distortion starts quickly. So no need to go much above these levels in normal use. A static discharge could degrade the noise specs... really, all chips should be shipped in antistatic packages. Do the rest of the JRC chips also exhibit high  noise level? (also remember to keep Iabc under 1 mA else damage can occur).
Regards, Mike

Paul Perry (Frostwave)

Anyone doing comparisons, don't forget the NE5517 (13700 copy, but replaces 13600 in most circuits).
personally, I havn't seen a noticeable difference between manufacturers.

Ed Rembold

Thanks for the tips guys,
... and yes, I'm using the linearizing diodes Mike.
I think I'll try a couple of different brands,  just in case.
I remember that OTA's are noisey,  but I remember that R.Keeley
said a certain brand of 3080 was quieter,  when he was doing his Ross comp. clone.
we'll see...
Thanks, Ed R.

gez

I just added a bit more to my circuit today and had to change a few things around, and now this chip  hums like crazy!  

It's happened to me before so I'm not too bothered, I'll just have to change a few things round again (sigh!) - layout seems to be everything with these little buggers (perhaps that was your problem too?).
"They always say there's nothing new under the sun.  I think that that's a big copout..."  Wayne Shorter

Paul Perry (Frostwave)

I've never had hum probs with any LM13700 type, but I always use fully regulated power.
Normal op amps are designed to reject a lot of ripple on the DC supplies, but most OTA circuits have the control current coming (usually) from the DC supply somehow.. so they might be likelier to pick up hum from the rails, Put it on a battery & see how it goes, if it isn't already.

puretube

gez: they`re not Cmos;
sounds like yoùr power supply is loaded to much by now, or you use a series-resistor to the B+ ? - make it a little smaller - measure the + rail-voltage from time to time after having made circuit changes.
or extra-decouple (filter) the control-voltage source with an extra couple uF.


//www.puretube.com

gez

I'm in the middle of tweaking a CMOS envelope follower to connect up to the chip and I forgot to connect up some caps for filtering  :oops:

Having said that, I have had problems with slight noise increase in the past when I've crowded the circuitry around the OTA (on a breadboard that is)
"They always say there's nothing new under the sun.  I think that that's a big copout..."  Wayne Shorter

puretube

probably just "breadboard-noise"...

Ed Rembold

Why "researching" the web for OTA stuff,  I see that the "synth-guys" make use of the CA3280, alot.   While digging up info on that IC,  I can across an article from EDN april 17 2003,  by our very own Mike Irwin!!
Not my app,  but very creative Mike!
Perhaps an opinion on the use of a CA3280 as a VCA,  for lower noise than the LM13700??
I've also asked Steve D. to get these chips- Any interest guys?
Thanks, Ed R.

Mike I.

Hi Ed,

The CA3280 is a superb chip but three "problems" - they usually cost around 5 US apiece, there is only one manufacturer, and there are no built-in buffers (unlike 13600 etc). So you would have to use an external op amp/FET/Darlington for many of the standard circuits. According to the specs, the linearizing diodes work more effectively than with the 13600 etc allowing better S/N. It also has much better offset specs than the others, making CV thump less of a problem.

The Jan.19, 1995 EDN shows a very nice VCA design by Mike Sims that uses both sections of a LM13600 to make one good OTA. Worth considering as it is less expensive than CA3280.
Regards, Mike

Ed Rembold

Thanks for the reply Mike,
In my circut (it's not a trem- rest easy R.G.)
there is "thumping" and hiss with the LM13600.
subbing in an LM13700 gets rid of the thumping, but not the hiss.
I can't stand the noise-  
So I'm gonna mod the circut for the CA3280,  I'm already using
opamps following the 13700,  so buffering the output of the CA3280's is not a problem.
Ed R.