Accurate SPICE macromodels for some op amps

[Nazar_lv]

OP07 model

• OP07_SN - TI/AD +-15V
• OP07_5V_SN - TI/AD +-5V

TL071 / TL072 / TL074 model

• TL071_SN - Texas Instruments

LF356 model

• LF356_SN - National Semiconductor

NE5534 model

• NJM5534_SN - JRC
• NE5534ON_SN - ON Semiconductor
• NE5534PH_SN - Philips
• NE5534TI_SN - Texas Instruments

NE5532 model

• NJM5532_SN - JRC
• NE5532TI_SN - Texas Instruments

LM318 model

• LM318NS_SN - National Semiconductor
• LM318TI_SN - Texas Instruments
• LT318A_SN - Linear Technology

OP467 model

• OP467_SN - Analog devices

THS4011 / THS4012 model

• THS4011_SN - Texas Instruments

AD8055 / AD8056 model

• AD8055_SN - Analog devices

LM6171 model

• LM6171_SN - National Semiconductor

LM6172 model

• LM6172_SN - National Semiconductor

LM7171 model

• LM7171_SN - National Semiconductor

Download page

There will be more models when i have free time

[Dave Zan]

Originally Posted by Nazar_lv

...There will be more models when i have free time

That looks like nice work, my compliments, I liked the reference paper.
When you do more models I hope they include the AD797.

Best wishes
David

[Monte McGuire]

Analog Devices offers an AD797 model that seems pretty accurate at least in terms of stability and dynamical response: License agreement for Spice Models | Analog Devices

The AD797 product page has a few other variant models, probably to handle slightly different expected offsets or other parameters.

[Dave Zan]

Originally Posted by Monte McGuire

...Devices offers an AD797 model that seems pretty accurate at least in terms of stability and dynamical response...

A lot of the manufacturers models have the dynamic response OK but are otherwise fairly poor, don't model power draw from the rails, don't have correct PSRR because they assume a (non existent) internal earth connection, and so on.
I don't know if the AD model is acceptable or not but Nazar's paper shows that he has included many aspects that are often not correct.
So I would be more confident of his work than the AD model.
But the AD model may be fine, what have you checked?

Best wishes
David

[Nazar_lv]

Thx David.

Often datasheet data are far away from reality, thats why i measure open loop gain(with different loads), CMRR, open loop output impedance as most important parameters.

[Dave Zan]

Originally Posted by Nazar_lv

Often datasheet data are far away from reality...

I checked the AD797 model and it sure doesn't match reality, or even the datasheet, on PSRR but at least it draws power from the rails.
I didn't notice temperature effects modelled in the net-list, but I didn't check this because it's mostly irrelevant for audio.

It is curious that the AD797 is not included with LTSpice, since AD owns LTSpice.
I assume Mike E. does not want to include components unless their models meet his standards, and that the LTspice staff don't yet have an acceptable model.
In the meantime I have a micro-model that is usable.
Thanks to Monte for the prompt to look at this.

Best wishes
David

[Monte McGuire]

I think the reason that the 797 is not in the LTspice distribution is because historically, only Linear components were installed. The merger between Analog and Linear is somewhat recent, and the Analog models are not of Linear origin. So, it's probably an artifact of all of the consolidation that's been happening.

I've never tested many aspects of the 797 model - thanks for doing that! I use LTspice primarily for stability and noise analysis, and also DC offset, so most 'fake' models using poles zeroes and pure gains with some pseudo-real IO components seem to work well enough. If I can get the circuit and surrounding passive component values to be at least somewhat plausible, then it can go into a PCB and be tested for real.

[Nazar_lv]

updated some models