I have downloaded the SPICE model file for the OP177 but found that it actually contains the model for the OP1177.
Where can I get the proper SPICE model file for the OP177 ?
We have done OP-07, OP177, OP777, OP1177, AD8677, ADA4077-1, ADA4177-1, so six generations.
All of them are very close. The 1177 has higher SR and GBW, so it depends on what you are doing as
to whether or not it would make a difference. I would not do a design with the 177 or the 1177, but rather
the ADA4077-1 or ADA4177-1. If it is an existing design, just use the OP1177 model.
thank you for the reaction on my posting.
The background of my question is: I was trying to figure out differences between the OP77(E) and the OP177, but could not make out anything significant in the datasheets, so I wanted to compare the SPICE models by inspection of the files and simulation of a high voltage (350V) current monitor circuit (in a vacuum tube curve tracer, DIY project).
Inspection of the OP1177 model contained in the present OP177.cir showed at first glance very different pole and zero frequencies. So my first thought was "that's something totally different". In the meantime I have found an OP177A model in the internet of 12/90 Rev.B which has 4 of 5 pole/zero frequencies equal to those of the OP77E model. Only the common mode gain zero ist at 200 Hz as opposed to 63Hz for the OP77E. I have not yet compared other detail although my main interest is to acchieve better than 74 dB common mode suppression at an internal gain of 34 dB, overall gain of 0 dB (because of an initial attenuation of 34 dB to get the 350V downto 7 V) with a bandwidth >100kHz. So the primary question was : can I get more precision with an OP177 than with the OP77 (of which I have still a few amples). I plan to keep almost the zero gain bandwith of the OP77 or 177 by incorporating a less precise but faster JFET OP as an amplifying buffer in the overall differential amplifier setup.
A spice model is a macromodel; the average op amp has 50-100 transistors. Most op amps are
junction-isolated, so every transistor is diffused into a silicon substrate, creating a parasitic,
reversed-biased diode to the substrate. In the old days of Intel 286s, if we gave you the actual
netlist and you used it for an 8-pole filter, it would take an hour or more to run a simulation. So the idea of
a very simplified model was born. Here is a good historical overview:
You are assuming that the macromodel exactly matches the actual silicon. Sometimes we get
close, but they are simplified. Some models don't model model voltage noise for example.
Some model flatband voltage noise, but not 1/f noise for example.
I would not trust any model from any manufacturer for what you want to do. Take the min/max numbers
from the spec table and use a spreadsheet.
-- Are you building the classic four resistor difference amp? If so, the resistors are more of a limiting
factor than the op amps.
-- Can you post a simplified schematic?
thank you for your continued interest and support, especially for the hint to AN48.pdf. That really looks like a valuable and informative paper.
My inquiry for an OP177 SPICE model was just another attempt to find out (by comparison of the models) possible significant differences to the older OP77, which I could not spot at first glance in present day datasheets available on the internet.
In the meantime I have dug out an original copy of the old PMI volume 10 Databook dating back to 1990, where the OP177 appeared as a new device. Comparing those datasheets it seems that the two devices differ only by an additional selection grade of the OP177 with only 10µV offset voltage, a better bias current cancellation, less thermal bias current drift and somewhat better noise specs. The "rest", especially CMR and PSR seem to be the same.
I have included two simplified schematics that illustrate my first approach to plate current tracing and the next version that should provide better CMR.
Plate voltage is supplied by fullwave rectified 250Vrms 50Hz derived from the mains power lines as proposed in other tube curve tracer circuits for economical reasons. Peak voltage reaches up to 350..370 V. The current sensor consists of a half bridge of 100 Ohm resistors plus two 50:1 attenuators consisting of 0.1% resistors of 100k91k3.9kOhm. They reduce the maximum common mode voltage from ~350 to ~7 V. My first idea was to amplify and buffer the differential output voltage with one of the cheaper instrumentation amplifiers of the INA126 variety. Its architecture is sketched on the right side, set-up for a gain of 5.
After assembling and testing the circuit it turned out that this type of circuit has a major flaw (compared to the classical 3 OpAmp instrumentation amplifier):
The two OpAmps are driven to very different output levels. The lower one amplifies the common mode signal by 5/4 to ~8,75V which is then attenuated back to original level to compensate the common mode signal at the upper OpAmps positive input, resulting in almost zero output. However, the slight delay and nonlinear distortion generated by the highly driven lower OpAmp are not cancelled as in the symmetrical 3 OpAmp instrumentation amplifier. In this case, after trimming the CMR with Rtrim a residual signal with ~5 mVpp amplitude was left. With the present plate current scaling this corresponds to a resolution of 0.5 mA which is worse than what I hope to acchieve.
Actually there is no need to employ an instrumentation amplifier with high impedance inputs behind the attenuator circuit. Considerably better performance can be acchieved by replacing it with a classical differential amplifier. The second schematic sketches what I am going to test next. This circuit avoids the extra delay and distortion in the common mode cancellation bridge. The version sketched has been additionally modified to provide an overall gain of 1 (instead of 1/10) .
The second OpAmp serves to increase the open loop gain to acchieve higher bandwidth. Precision OpAmps OP77/OP177 have a gain-bandwidth product of only 0.6 MHz. At a gain of 50 the bandwidth would be only ~12 kHz. The 2nd OpAmp, set-up for a gain of 13..15 increases the bandwidth to ~240 kHz, which is even slightly more than that of the INA126 at a gain of 5 (200 kHz).
Hopefully this circuit will perform in reality just as well as in simulation.
I knew the AD629 but not the newer AD8479.
In fact my 2nd version is modelled after the AD629, which however has only +(-)270V common mode range. So for 350V with some headroom additional series resistors woud have to be used, reducing the overall gain correspondingly.
Yes the noise gain would be 50 which is comparable to the AD629's behavior. Actually it is somewhat more, but that is due to a generous common mode range setting of roughly 700 V at supply voltages of -9 & +18 V, assuming a maximum common mode voltage of ~15V at the OpAmp inputs. It might be reduced to a lower value between 25...30.
The AD8479 looks very interesting with its +-600 V common mode range and still quite high bandwidth, but I can't use the SOIC-8 package, can only handle the old DIL-8 stuff.