TL074-EP: Output Impedance (Zo)

Hi Brian,

a TINA-TI simulation yields this result:

brian_tl074.TSC

Keep in mind, though, that the TL074's model is an out-of-date Boyle macromodel which doesn't give very precise results.

Kai

To Ron,

1. In your expression below about the total Zo, what is the corresponding frequency with?

It is approximately 128 + 64 + 38 = 230 ohms  [output resistor, 2nd resistor, emitter output] 

2. As unit gain buffer below, with 1uF capacitance load, What's corresponding frequency_Pole to get the RO?

Brian,

That is the great thing about low impedance transistor outputs, the impedance is mostly resistive (flat over frequency). Most of the impedance in TL074 is resistors which are resistive and flat over frequency. So I did not, and need not pick a frequency in the process. Your Fpole formula assumes that Ro is fixed resistor. 1uF is bad idea in practice. 

Here is an answer for the question you did not ask. For unity gain TL074, 100pF is generally safe. 200pF or more can be problematic.

Thanks. Customer understood the 1uF output cap is not stable with TL074. However, they would like to implement the more precise simulation model for the phase margin to judge if OPAs are stable or not? 

So What's the output impedance you suggesting for simulation? 230-Ohm, 300-Ohm or 150-Ohm? In the below plot of the Zo vs. frequency. When Zo is 230Ohm, the frequency is at 25Hz. i'm a little confused as you mention the 230-Ohm is at flat-over frequency. but in below plot, where is the flat-over frequency?

also, we found a material from internet. it mentioned the resistive   region is in the middle of the curve. why? how to define it?

Brian, 

The TL074 sim is flat after 100Hz. Does Zo below 100Hz even matter? 

I still suggest a flat 230 ohms.

I'm still a little confuse about your explanation. 

1. In the below sim plot, the ZO is 230 Ohm at 25Hz, correct? and you mentioned that "after 100Hz" does it mean that freq. is lower than 100Hz?

2. As slide i attached in the previous thread, i'd like to know how to decide the resistive, capactive and inductive region in the Zo vs. Freq sim.   

Since customer would like to build up a precise simulation model with stability analysis to know which device/parameter is good in their particular circuitry before design_IN, please help to support this. Very thanks. 

Brian

You can use the sim results.

Hi Brian,

I fully agree with Ron. From my experience with the TL07x I can say that this OPAmp doesn't like capactive loads, being directly connected from the output to signal ground, at all.

Kai

Hello Kai,

Sorry, i'm really confused with what you said. In the plot below, we still doesn't figure out how to get the 230-Ohm? You mentioned that the TL074 doesn't work in Capacitive behavior but resistive. From below sim plot, the ZO is 150.02 ohm in the resistive. is it NOT 230-Ohm, why?

Since customer would like to do the simulation and build up a model, could you offer more detail one! Very thanks.

Regards

Brian  W

Hi Brian,

I cannot help you with a better Spice Model of TL074.

If doing simulations is of crucial importance for the customer, I would recommend to use the OPA1644. It's well known as heavily improved version of the TL074 and is available at a reasonable price. On the product site you will find a OPA1641 PSpice Model (Rev. B) from 2019 and a OPA1641 TINA-TI Reference Design (Rev. C).

Kai

Brian,

I measured a TL074 with the same design as the EP device to get a Zo measurement.

I measured gain/phase with 3 different resistive loads 10k, 2k, 733 then I calculated the output impedance using all three combination pairs of resistors with the gain difference between the loads. As expected, all three combos gave the same result.  

260 is pretty close to my estimate , the TL074C runs at 200 ohms so my estimate is right in the middle of the two TL074 designs.

Hi Ron,

very interesting! How have you measured this? Can you show a schematic, please?

Kai

Kai,

My favorite method is using the AOL over frequency data tested twice, 2 different resistive loads. This is a forward method. This could also be tested backwards by injecting a signal into the output.  

Then I use complex numbers and math to solve for ZOUT term for both voltage dividers. If I know the result will be resistive then complex numbers are not needed. However the Zout dip near 100Hz wasn't accurate because I skipped this step. 100Hz is 290 ohms if I had used gain and phase data. From 1Khz and above the results are the same using standard and complex numbers. Most of the time I use three loads to make sure all the load pairs come up with the same result.    

Thanks, Ron!

Kai

Hello Ron, 

Thanks. As you mentioned in the previous discussions, i'd like to check which value is correct one? 230 or 260 Ohm for TL074. Customer would like to do the SIM model with this ZO. 

And also, i'm confused about the meaning of the " TL074C runs at 200 ohms so my estimate is right in the middle of the two TL074 designs"

Regards

Brian 

Hi Brian,

to have some headroom in the phase stability analysis (worst case) I would take 300R. Keep also in mind that the TL074, like any other OPAmp, will show some manufacturing tolerances. About +/-20% are realistic.

But let's wait for Ron.

Kai

Brian,

I implemented complex number equations and used a 2nd AOL test board that is engineered for lower frequency measurement. This is helpful for below 3 kHz data which I normally have no interest. (unless the op amp was very low power / bandwidth)

Here is all the data I have in one chart. This is as good as it gets. If have nothing else to provide.