I’m moving this question from the Rev7 features thread to here because I think it is best covered here.
Now the series resistors are 120 Ohms. Is this the optimal value or would for example 68 Ohms be better?
Regarding the wanted usage I mentioned at the beginning of this thread I see mostly these points affected:
- driving old-school TTL: a fanout of 10 is common, this means you must be able to drive 16mA and keep the high-output between 5V and 2V while doing so. When we account for marginal Vusb and drop in the voltage regulator, the drop in the resistor should be below 2V. 120 Ohms gives a drop of 1.92V - so we are fine with 120 Ohms or below
- driving I2C in Fast mode plus: This means we must be able to sink 20 mA and keep the voltage at the input below 0.4V while doing so. This means our whole resistance must be below 20 Ohms. Since the 1T45 already has an impedance in the range between 10 and 20 Ohms achieving this means we can’t use more than something like 5 Ohms as series resistor. This would severely affect any protection goals. So I think we have to scratch this goal and restrict us to I2C Fast Mode (400 kHz max.).
- fast digital IO: The 25 MHz I mentioned above shouldn’t be affected by 120 Ohms. Since Rev 7 will have a PSRAM one could want to fully utilize it for a fast logic analyzer or similar. But I’m not sure what kinds of speeds are realistic there and if they are really affected by 120 Ohms.
- fast PWM or clock output: a PWM that utilizes the max. freq of the RP2350 for very short LSB times. Or a clock output with high frequency. This could be intersting for some use cases and the 120 Ohms will affect you for example if you try to go to 75 MHz (half MCU freq.)
- The point of protecting the DUT is a bit difficult because it is a goal of the BP to interface with as many different kinds of DUTs as possible. So very different requirements and kinds of damage level.
So I don’t see a clearly defined & realistic feature that is not possible with 120 Ohms. But the lower the resistance, the more speed is possible, but the protection becomes more difficult.
To test the resistor effects on the protection, I used the protoboard circuit from yesterday and replaced the 120 Ohms with 100, 82 and 68 Ohms:
Interestingly the voltage the resistor started smoking at didn’t vary much. It was always about 24V. But the lower the resistance, the more power the TVS diode array and SMF5.0A had to carry. With 120, 100 and 82 this was still fine at 24V, so the resistor was the limiting factor. With 68 Ohms this changed and the TVS diode blew at 20V.
In the end I think there is no optimal resistor value we could calculate, it is guesswork and eyeballing. I think 120 Ohms should be fine, but it could also be slightly lowered to the more common value of 100 Ohms. But I wouldn’t go far lower unless some real issue with the series resistor comes up.