OPA3S2859-EP: OPA3S2859-EP: Clarification for datasheet specs

Hello Peter,

Just a quick update: your proposed 300 Ohm case on the lowest 1kOhm feedback path is fine according to design simulations.  However, there are issues with using low-resistance feedback resistors on the 'higher resistance' feedback paths, which nominally are 10k & 100k.

There are internal design choices involving the Ron for the switches which are sized to expect certain resistor values.   A lower-than-expected resistor value could enable higher currents than intended.  The switch Ron doe shave a max DC current of 1mA at high temp, which is a concern if lower value resistors are used in these high-resistance feedback paths.

The resolution so far for your first question is: it may be possible to a degree, but there are issues shifting away from the general magnitudes of the resistor values presented in the datasheet.  Scaling the whole range for all three feedback paths to be 300-1k ( difference of 700 Ohms) would likely push high currents across the switches.

Is there a different way to accomplish the scaling or customization you are looking for Peter?  I am still in contact with our team on figuring out what may or may not be allowed, but it is safe to say it is recommended to use the 1k, 10k, 100k selections until I know more about the workings of the design.

On your third question regarding currents, your use case is likely okay compared to datasheet spec.  I will update you alongside the first question if I find out otherwise.

In summary, I would not greenlight the approach with the scaled down resistors, but it is fair to continue designing with the input current setup you have outline in Question 3. 

As always, please ask if there is a need for clarity or share data if you find a contrary or supplemental data point.

Best,

Alec  

Hello Peter,

I understand the desire for more objective clarity on resistor use.  The technical reasoning is as such: the switches inside the OPA3S2859 have Ron value sized relative to the expected gain path resistors.  The current through the highest 100kOhm path is set by the expected Rf and the unchangeable Ron of the switch.   For the highest Rf path, the Ron is sized larger than the 1k or 10k paths to work with Rf = 100kOhms; the switches' limit for output DC current is 1mA.  There is effectively current limiting occurring with the Rf resistor, which enables the switch to operate as intended.  When a greater current is present, the voltage across the switch can modulate between COM and FB pin.  

With a much lower resistor value (either 600 Ohms or 4kOhms), there is a fair risk of reliability & rating issues.  The DC current would exceed the 1mA limit on the highest Rf path, if a lower than target (100kOhm) resistor is used.  

The use of higher than target resistors (say, for example, a 100kOhm on the 10k path) would not have the reliability issues, but may result in lower bandwidth.

As for the 'target resistance' zone and how much room there is to be above or below the target resistor value, I would guide users to design near the target values; the difference between 4kOhms and 100kOhms is significant; however the difference between 4kOhms and 10kOhms is less likely to be a definite current issue or reliability problem.  

If you can get the same order of magnitude and be half-as-much or closer, that may be a good rule of thumb for starting.  Now we do not have data to support what threshold of resistor choice 'works' and does not cause reliability issues or high current draw; any major deviation in use from the 1k/10k/100k model is an assumed risk.  As of right now, I cannot define what constitutes a major deviation.

I hope I was able to provide some more clarity; I know it is frustrating to not be able to scale down the resistors as would be possible in other amplifier applications.  The integrated switches have different Ron values for the different target Rf values; this inhibits the use of big changes in resistors for Rf.

Best,

Alec

Hello Peter,

Great timing on your end!  I actually just had my last discussion with the team members I have been working with.

There are two conclusive points of interest for you:

1) You can leave the highest gain setting path un-terminated/no-connect/NC and not switch to that path via the SEL (select) pins.  If the SEL1 & SEL0 pins do not transition to/through SEL1 = 1, SEL0 = 0, the amplifier loop will remain intact.

2) If there will be states where SEL1 = 1, SEL0 = 0, there will need to be termination of the high gain path pin.  In this case, you can use a 100kOhm resistor to terminate (but not use, as you described).  If not terminated and left open, the amplifier loop will be open.  Additionally, there could be antenna effects from the non-terminated stubs of the device pins, especially if traces are present at those pins for the presumably-used high gain path.  

The recommendation is the leave the package pins as no-connects and remove the S1 =1, S0 =0 state from the control circuitry.  If this is not possible, termination with the recommended 100kOhm resistor will assist with mitigating any unwanted effects.

Best,

Alec