TL071/TL081

[Martin Czech]

I think there are many cases were this is right: "binning".
OTOH noise measurements are quite costly, 
so as a manufacturer I'd try to design low noise parts.
And the lower supply current of the 061 makes me also think that
at least internal biasing is changed.

The problem with distributions and binning is that you might get
less parts of the wanted bin then the market needs, now, what
do you do with the unwanted rest of the wafer?
Not very likely for a low cost product like this,
if you ask me. Is there a chip foto in the data sheet?
I've to take a look.

For expensive parts like top notch microprocessors
there must be some speed binning.
And a OP27 is also quite expensive (I think so) because
noise is measured, not because of die size.

I must admit that this 051 061 071 081 stuff is one 
of the most confusing things when dealing with from the shelf
op amps. I'm only guessing, not knowing.

m.c.

:::Date: Thu, 09 Nov 2000 11:51:49 -0600
:::Subject: Re: TL071/TL081
:::From: "Grant Richter" <grichter at asapnet.net>
:::To: owner-synth-diy at node12b53.a2000.nl, Tim Curtis <sexsymbol at execpc.com>
:::CC: AH <analogue at hyperreal.org>, synth-diy at node12b53.a2000.nl
:::Content-transfer-encoding: 7bit
:::X-MDaemon-Deliver-To: synth-diy at node12b53.a2000.nl
:::X-Return-Path: grichter at asapnet.net
:::
:::As far as semi-conductors go. There are only so many die types, and doping
:::is still a somewhat uncertain process. A number of different parameters are
:::going to have Gaussian (bell curve) distributions around the center target
:::point. But you don't want to have to throw away dies outside some
:::percentile. You just assign them a different part number with a relaxed
:::spec.
:::
:::In the old days, they would take a batch of transistor die and "test" them
:::into part numbers. That is they would check the die test results against
:::multiple specs and slap a part number on the ones that passed for that part
:::number. So the same die might appear as several part numbers.
:::
:::I could be wrong, but I believe that's how 2N3905/2N3906 (higher betas are
:::3906) CA3045/CA3046 (higher betas as 3046) and so forth. Basically splitting
:::the bell curve and assigning two different part numbers to the upper and
:::lower half, and charging more for the better results.
:::
:::Maybe the 1N400x series are also like that, a 1N4001 is a 1N4002 die that
:::tested below 100 volts, so now it's rated at 50 volts but might actually
:::have a breakdown voltage between 50 and 90.
:::
:::I had always assumed the TL07x were the lower half of the bell curve for
:::TL08x sorted for noise figure. The noisier die that still pass DC spec are
:::labeled 08x and the quieter ones 07x, but otherwise are the same die?
:::
:::----------
:::>From: Jim Patchell <patchell at silcom.com>
:::>To: Tim Curtis <sexsymbol at execpc.com>
:::>Cc: AH <analogue at hyperreal.org>, synth-diy at node12b53.a2000.nl
:::>Subject: Re: TL071/TL081
:::>Date: Thu, Nov 9, 2000, 8:54 AM
:::>
:::
:::>     In my opinion, the two are pretty much interchangeable.  The 07x 
seriers
:::> has a slightly quieter front end, and I believe the output stage is biased 
a
:::> little heavier to give better distortion.  In general, I have used the two
:::> interchageably.
:::>
:::>     -Jim
:::>
:::> Tim Curtis wrote:
:::>
:::>> Can anyone really explain to me what the difference is between these two
:::>> op-amps?
:::>>
:::>> One of my MiniMoog Oscillator boards was diagnosed as having a bad
:::>> TL081.  I replaced it with a TL071because I have some of those around.
:::>> Nada.
:::>>
:::>> Is this normal?  I was always told that the two were pretty much
:::>> interchangeable.
:::>>
:::>> Or is it just a problem with them in that circuit?
:::>>
:::>> Tim
:::>
:::>