If i plan to operate the 1.8V device with supply v...
# sky130
If i plan to operate the 1.8V device with supply voltage as 3V/5V , the doc below says for gate oxide and D/S : The maximum voltage across the gate oxide (gate to channel voltage) is restricted to:
Any HV NMOS device: 7.3 V @ 25C.
Any HV PMOS device: 8.1 V @ 25C.
The maximum source/drain to substrate junction voltages are restricted to the following: 1. Any HV NMOS device: 11.0 V @ 25C. 2. Any HV PMOS device: 11.0 V @ 25C. So is this the correct undertanding that i can still be able to operate a 1.8V device to 3.3V ? Although the spice might be out of range ? https://skywater-pdk.readthedocs.io/en/main/rules/hv.html
In my experience, the thin-ox low voltage devices are not used outside of the rated voltage range. This is not a simulation issue, but rather a reliability issue. Using higher voltages on 1.8V devices can lead to hot carrier migration which will change the operating characteristics of the transistor with time or in the worst case, device destruction with the gate shorting to the source/drain. https://skywater-pdk.readthedocs.io/en/main/rules/device-details.html#v-low-vt-nmos-fet
@User The keyword ‚HV‘ refers to the thick oxide MOSFET. Don't use 1.8V devices beyond ca. 2V. They will sustain a higher voltage for sure for a limited time, but no guarantees how long this is.
Because I have been mostly a mixed-signal designer all my career, I have had my share of chips that accidentally contained a circuit driving a 1.8V device with 3.3V. Some of them worked, some of them didn't. It's right at the edge between being a long-term reliability issue and just flat-out blowing a hole in the gate as soon as you power it up. I have done some tests on sky130 parts driving all of the 1.8V logic up to 2.75V and the parts continued to work reliably. I did not dare to go higher than that because I had a limited number of working parts on my desk and I didn't want to fry any of them.
😆 1
@Tim Edwards that was good insight. I was actually looking for 3v3 enchancement type mos , but in 3v3 sky-pdk has native vt transistors which has near zero threshold (not desirable) , or use nvt device with -ve on voltage for complete turn-off. I would need upto 2.6V . Now i need to rethink what is possible. @Mitch Bailey @Harald Pretl
@Ryan R oscillators and power amplifiers drive MOSFET up to 2..3xVmax, but that is for a sinusoidal waveform, and absolutely non advisable without a reliability calculation.
👀 1
These point makes me think of a book i read recently "Think again" by Adam Grant
@Tim Edwards 15 years ago i was tempted to add a test mode to a flash device, to allow driving the (internally generated) program/erase voltage (~9V) onto some LV sense amplifier transistors, to smoke them out and make it almost impossible to read out the data any more... This test mode could later be implemented as a command to instantly shred all the data. Of course this was way too dangerous and this thing has never ever been tried. It was however doable with only a ROM firmware change.