[volt-nuts] 3458A integrator

Bruce Griffiths bruce.griffiths at xtra.co.nz
Thu Nov 10 18:28:21 UTC 2011


Bruce Griffiths wrote:
> Randy Scott wrote:
>> Since it seems that more people now have access to the 3458A CLIP, I 
>> have a question about the integrator amplifier being used.  In the HP 
>> Journal article on the 3458A, they list a set of specifications for 
>> the integrator and then go on to say that it required a custom 
>> amplifier design.  I had always taken this to mean that they had 
>> constructed a custom op amp IC for the task.  I was then surprised to 
>> see that the schematic used a set of off-the-shelf op amps in the 
>> integrator loop: one each of LT1001, LM6361, and AD848, plus a 2N4392 
>> JFET and various passives.
>>
>>
>> It bears some similarity to the integrator in the 34401A that uses an 
>> OP-27 and an AD711.  In the notes on the 34401A, they mention that 
>> the OP-27 is used to improve the input offset characteristics of the 
>> AD711 (a JFET amp with large Vos).  I suspect that the LT1001 in the 
>> 3458A serves a similar purpose.  Both the OP-27 and LT1001 have low 
>> Vos, but are too slow to be useful as the integrators in either circuit.
>>
>>
>> The problem is that I am failing to even begin to understand how 
>> either of these two designs work.  Can anyone shed some light on this 
>> or maybe point me to a reference that I could use to help figure it 
>> out for myself?
>>
>> Thanks.
>>
>> Randy.
>>
> The OP27 in the 34401A integrator reduces the effective input offset 
> of the AD711 (including the finite (non negligible) input required to 
> produce the desired integrator output slew rate).
> Since the OP27 output swing is small only its small signal 
> charateristics and not its large signal characteristics (slew rate 
> etc) are important. The resistive divider network at the OP27 output 
> adjusts the effective gain bandwidth of the OP27 so that the 
> integrator loop remains stable.
> Higher slew rate integrators with low input offset using this 
> technique are feasible with the faster JFET opamps that are now 
> available.
> LTSpice simulations will quickly show how such integrators behave and 
> the necessity for the OP27 output divider.
>
> The more complex integrator of the 3458A is an earlier attempt to 
> achieve a high slew rate integrator with a low effective input offset.
>
> Bruce
>
In the 3458A integrator the AD848 plus the 4392 source follower form the 
uncompensated integrator with large input offset (volts ? due to the 
source follower) and a significant slew rate induced offset.
The LM6361 corrects the slew rate induced offset variations by adjusting 
the input at the noninverting input of the AD848 whilst the LT1001 
corrects the input offset (at the 2N4392 gate) by adjusting the input 
signal at the LM6361 non inverting input. The additional passive 
components ensure that the resultant integrator is stable with a well 
damped transient response.

Bruce




More information about the volt-nuts mailing list