Hi to all. I'm trying to figure out the current that can flow through a P-Poly 0p35 resistor. From the documentation I read that: "Currents up to 500 µA/µm of width (preferred use ≤ 100 µA/µm)": This seems quite strange to me; if I understand correctly, it is recommended to pass a current much lower than 500uA, despite this being the nominal value. In my circuit a current of 150uA should flow and I had done some calculations discovering that I could have a width of 300nm and therefore it would have been a good component for me (W= 350nm) However these calculations are based on the fact that the resistor can support 500uA/um myMinimunWidth = 150uA/500uA. AmI doing something wrong?

Does this feed into the lack of Safe Operating Area data and the queries around that may still be a factory secret. It is my understanding there maybe 2 current ratings, a never exceed rating and a current limit set that should achieve the expected service lifetime of the technology. Which again maybe unclear what lifetime limits to expect. To some degree the factory and PDK can only do so much, with the factory concern over fulfilling production and the customers problem with achieving yeild and IC longevity (as the specific design choices of layout and use affect it the most). Electromigration effect is a significant contributor to the reason the limits are set to where they are. Electromigration is always an effect (see current news with some PC processors and top of line running slightly over voltage and causing failure rates before expected warranty) The effect is not linear and the negative effect on circuit compounds over time. My interpretation of the two numbers are a never exceed and a lower current value (if never exceeded either) will result in expected service lifetime of the device. Everything in between is in the world of it depends and might need a physics/chemistry model to get an answer. That does not excuse the lack of some SOA graphs covering limits (voltage, current, lifetime) and expectations to allow the designer to use data to make choices. Instead we continue with the whichcraft. Hopefully overtime some of the open source nature will feedback and some entity will redo and publish experiments the factory has already done that can produce such data, which can be retested every 3 years to confirm, but it's an expensive business and squirrels like to hide their nuts.

Now I understand. I am trying to make a fairly complex design and I am trying to foresee all the possible problems that may arise so that my project can work. What you say about the currents could be a correct interpretation and I am sorry that more details are not known. It seems a bit like relying on luck and hoping that everything works or that it does not break after a couple of months. Of course if that is the case, it is quite difficult for those who want to make prototypes of projects of a certain weight to rely completely on this production model, unless something changes.