LT1013: Can anybody tell me the output impedance of the LT1013 op amp?

Hi Jerry,

you can estimate it from a TINA-TI simulation. Inject a sine current of 100µAp at unity gain frequency into the output of LT1013 and measure the output voltage:

This gives about 26R. So, the open loop output impedance of LT1013 is about 26R.

Kai

Hi Jerry,

The LT1013 is one of the op amps that we haven't obtained measured open-loop output impedance Zo just yet. We took a look at the overall design and in particular the output stage design. That information gives some idea what to expect in terms of Zo and its behavior across frequency. Doing a few calculations allows us to estimate the Zo and check the simulation model for correctness. Although this method isn't as exact as having measured data this method does get us in the ballpark pretty quickly.

The LT1013 model Zo was tested and it passes a reasonableness test for its design and biasing. It should be sufficient for a first-order analysis. If you find that you are receiving questionable results from your analysis, let us know and we can try and dig into this further. 

Regards, Thomas

Precision Amplifiers Applications Engineering

Hi Thomas,

my TINA-TI has two models of the LT1013, one from 08/31/89 / release 4.01 / "LT1013/301" and one from 07/18/90 / release 4.03 / "LT1013/5_1". Both give different results.

I see you have another model. Can you tell me which one you use?

Kai

Hi Kai,

I tested both of the Texas Instruments LT1013 models listed in the TINA macros and obtained the same Zo results. The key is to set the vertical Voa scale to the logarithmic scale option. Then re-label the vertical scale as Zo (Ohms).

Here's my TINA file:

LT1013_30V Zo Test.TSC

Regards, Thomas

Precision Amplifiers Applications Engineering

Thanks, dear Thomas! I will check it...

Kai

Hey Jerry,

The LT1013 is an old device from before we characterized the output impedance so unfortunately we do not have that data. However, given the architecture i believe it is going to have a dominantly resistive impedance over frequency around 50 Ohms.

The model for the LT1013 is actually quite old and if you look inside it is a very simple model. The output impedance would not have been explicitly modeled, given that it was not a parameter we characterized.

Perhaps if you share some information about your application and why you are concerned about this parameter we can offer some additional guidance.

Kai,

It is not generally good practice to rely on a model, especially an older model, for characterization data not given in the datasheet. Keep in mind that this is not a true transistor level model of the device, and only emulates the behavior of certain blocks. Models will generally specify in the netlist what is included, and if a parameter is not specified you should not assume that it has also been modeled. Without the ability to verify the model against the datasheet I would shy away from relying on it.

That being said, the ~50 Ohm resistive output impedance (remember Zout = Zo/(1+AolB) so at the 0dB point Zout = Zo/2) is pretty standard for op amps of that time so this particular model is probably not far off.

Hi Thomas, hi Zak,

the model of LT1013 is old, yes, but nevertheless it should be able to simulate the AC output resistance. At least an estimated value, as I mentioned.

This is an extract from the 30V model:

And this is an extract of the 5V model:

By the way, please notice that the values of RO1 and RO2 are just swapped in the two models. A typo?

And I get different results, when I simulate the open loop output impedance.

First, the model Thomas is using:

The 33.98dB reading is the quotient of output voltage divided by output current, given in dB. Reversing the dB logarithm gives 50R.

Then, this is the result with the 30V model of my TINA-TI, which should be the same as Thomas used:

It also gives 50R.

And this, finally, is the result with the 5V model:

Yes, this model does not give tight results but only allows to estimate the open loop output impedance. But at least it should be correct in saying that the output impedance is somewhere under 100R. Such an estimated result may also be very useful.

To be honest, I actually never know whether I can trust a model or not. Many simulation tests have left me with a bloody nose. So, I do always assume huge tolerances when doing a simulation and do always carry out very thorough measurements before going into series production. And I have learned that only a small fraction of parameters is modelled at all...

Kai