One more LMx08 equivalent: the most obvious of all, the LT1008. I
didn't think about it earlier because I assumed it would have been
discontinued long ago.
The LT1008 was designed specifically to upgrade LMx08 sockets, so it
uses Pin 8 for overcompensation -- just like the LMx08. It is one
generation older than the LT1012/OP97, so its DC specs are not quite as
good as theirs' (but still *much* better than the LMx08).
If I were choosing an opamp for this type of circuit today, with a new
PC layout, I would use either the LT1012 or OP97. On the other hand, if
I were fixing an HP10811 that had a defective U3, I'd use an LT1008 to
avoid the need to re-route any traces.
Best regards,
Charles
On 9/13/2019 6:04 PM, Charles Steinmetz wrote:
You are very welcome, Gilles. One point of clarification, and one
additional note:
I mentioned the LT1012 and OP97 together, then said I've settled on the
OP97 without further comment. My choice had nothing to do with
differences between these two parts. As far as I can tell from using
them extensively, there are no differences -- for all I know, they could
be made using identical masks. I chose the OP97 for supply-chain
reasons that won't be relevant to anyone else.
Looking at the 10811A/B schematic, the bridge sampling resistors (R22,
R23) are 1Mohm. This is what accounts for choosing U3 with extremely
low input bias current. Note that the LT1012/OP97 are roughly 1000
times better than the LMx08 in this respect (pA for the LT1012/OP97, nA
for the LMx08) as well as input input offset voltage (mV for the LMx08,
uV for the LT1012/OP97).
One thing to note when replacing an LMx08 with an LT1012/OP97: Notice
C19 (0.01uF from Pin 8 to Ground). This is an *overcompensation*
capacitor, used to slow U3 down for lower noise and better transient
response of the oven control loop. The LT1012/OP97 use *Pin 5* for
overcompensation instead of Pin 8, so C19 should be connected from Pin 5
to Ground.
Best regards,
Charles
On 9/13/2019 11:17 AM, Club-Internet Clemgill wrote:
Charles,
Thank you for this very useful information !
My plan is to « clone » as close as possible the initial 10811 design,
as it is well proven.
Except changing the crystal and possibly use modern components if they
bring any improvement, and adapting the « mass » to the crystal format
and manufacturing it through 3D printing.
Did anybody tried something similar ?
Best,
Gilles.
Le 11 sept. 2019 à 22:05, Charles Steinmetz <[email protected]> a
écrit :
Didier wrote:
The LM208 is a single op-amp with low offset (<3mV) and low bias
current.
Even 3mV offset is verging on horrible for a "precision" opamp these
days -- it's hard to find any respectable opamp that doesn't do lots
better.
The secret sauce of the LMx08 series is the super-beta input pair,
which matches FET opamps for input current but also holds that low
(room-temp) input current over the full temperature range (the input
current of FET opamps generally rises by at least 10x to more than
1000x at warmer to hot temperatures).
There is no need for extraordinary measures -- there are a number of
more modern opamps specifically designed to upgrade LMx08 sockets.
The OP77 is one:
<http://pdf.datasheetcatalog.com/datasheet/analogdevices/206654345OP77_c.pdf>
The LT1012 and OP97 are two more. I've used both extensively, and
finally settled on the OP97 for this sort of application:
<http://pdf.datasheetcatalog.com/datasheet/analogdevices/140679914OP97_d.pdf>
Best regards,
Charles
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
