Does an LM285 voltage reference IC's noise scale with revers

[Lostgallifreyan]

I'll try to find out for myself, but I think having looked at the data sheet
and found no obvious answer, I can justify asking here..

The LM285-2.5 reference has wideband noise (10 Hz to 10 KHz) at 120ľV when
its reverse current is 100ľA, according to both the National Semiconductor
and Linear Technology data sheets. Neither says what happens at different
reverse currents, and the graphs don't allow me to work it out.

Either I'm missing something obvious thart I ought to know (and can't hwlp
thinking ought to be spelled out in the data sheet too if it's that
important), or it's not different enough at any operating curent to matter.
Anyone know which?

 

Lostgallifreyan schrieb:

I'll try to find out for myself, but I think having looked at the data sheet
and found no obvious answer, I can justify asking here..

The LM285-2.5 reference has wideband noise (10 Hz to 10 KHz) at 120ľV when
its reverse current is 100ľA, according to both the National Semiconductor
and Linear Technology data sheets. Neither says what happens at different
reverse currents, and the graphs don't allow me to work it out.

Either I'm missing something obvious thart I ought to know (and can't hwlp
thinking ought to be spelled out in the data sheet too if it's that
important), or it's not different enough at any operating curent to matter.
Anyone know which?

Noise will be dertermined by the input stage, whose operating conditions
do not depend on reverse current of the LM385; that is felt by the
output stage only. Filtering capacitors integrated on the chip
apparently shape the frequency characteristics of the noise, they appear
too cut it down above 1kHz to 10kHz. This can be seen from curves in the
corresponding LT1004 data sheets. The LT1004-1.2 and -2.5 made by LT and
also by TI are the same device with tighter voltage tolerances (other
specs and circuit diagram agree) while still affordable. One typically
gets +-1mV at room temperature.

If unsatisfied by the actual noise performance of the LT385 load with C
or R-C.

Martin.

 

[Lostgallifreyan]

clicliclic@freenet.de wrote in news:4C44C068.84B7558@freenet.de:

Lostgallifreyan schrieb:

I'll try to find out for myself, but I think having looked at the data
sheet and found no obvious answer, I can justify asking here..

The LM285-2.5 reference has wideband noise (10 Hz to 10 KHz) at 120ľV
when its reverse current is 100ľA, according to both the National
Semiconductor and Linear Technology data sheets. Neither says what
happens at different reverse currents, and the graphs don't allow me to
work it out.

Either I'm missing something obvious thart I ought to know (and can't
hwlp thinking ought to be spelled out in the data sheet too if it's
that important), or it's not different enough at any operating curent
to matter. Anyone know which?

Noise will be dertermined by the input stage, whose operating conditions
do not depend on reverse current of the LM385; that is felt by the
output stage only. Filtering capacitors integrated on the chip
apparently shape the frequency characteristics of the noise, they appear
too cut it down above 1kHz to 10kHz. This can be seen from curves in the
corresponding LT1004 data sheets. The LT1004-1.2 and -2.5 made by LT and
also by TI are the same device with tighter voltage tolerances (other
specs and circuit diagram agree) while still affordable. One typically
gets +-1mV at room temperature.

If unsatisfied by the actual noise performance of the LT385 load with C
or R-C.

Martin.

Thankyou. That looks complete. I can't help wondering why they specify the
noise at a given reverse current though, if it does not depend on it.

Anyway, I'll be filtering with 10K and 1ľF before the low noise amp I'm
already using (because several inputs will be fed with this reference).

I'll also look at the LT1004, though for this current task the differences
between voltages are ok, no worse than between existing references in the
rest of the system.