quantized
Member level 2
- Joined
- Jul 6, 2012
- Messages
- 51
- Helped
- 1
- Reputation
- 2
- Reaction score
- 1
- Trophy points
- 1,288
- Activity points
- 1,887
I've been banging my head against the wall on this for a day now and stumbled across the solution, but I can't figure out why it works.
The foundry supplies SPICE models; unfortunately using them in the standard "Dname pos neg" style doesn't work; when I swap the terminals everything works perfectly. So I look at the diode model in the SPICE deck and it has this mysterious parameter "type=-1". I remove that and everything works with the diodes connected the way SPICE says they should be.
Unfortunately I can't seem to find any explanation of this "type" parameter in the **broken link removed**. This worries me.
For what it's worth, the foundry deck only has this weird parameter for the P+/N-Well diode model and none of the others.
If it matters, this is for a quick+dirty+crude bandgap reference so I can measure on-chip temperatures. It's the only truly analog circuit on the chip, and it doesn't have to work well… just well enough to assist with troubleshooting. I'm using the circuit in Baker's book, but I'm modeling the reference diodes using the foundry's P+/N-Well model rather than the PNP BJT model (P+/Nwell/substrate), mainly because their LVS deck's support for BJTs is pretty poor. Since the BJTs in the bandgap reference all have their base+collector tied to ground they're equivalent to a diode (at least as far as SPICE is concerned), right?
Thanks,
The foundry supplies SPICE models; unfortunately using them in the standard "Dname pos neg" style doesn't work; when I swap the terminals everything works perfectly. So I look at the diode model in the SPICE deck and it has this mysterious parameter "type=-1". I remove that and everything works with the diodes connected the way SPICE says they should be.
Unfortunately I can't seem to find any explanation of this "type" parameter in the **broken link removed**. This worries me.
For what it's worth, the foundry deck only has this weird parameter for the P+/N-Well diode model and none of the others.
If it matters, this is for a quick+dirty+crude bandgap reference so I can measure on-chip temperatures. It's the only truly analog circuit on the chip, and it doesn't have to work well… just well enough to assist with troubleshooting. I'm using the circuit in Baker's book, but I'm modeling the reference diodes using the foundry's P+/N-Well model rather than the PNP BJT model (P+/Nwell/substrate), mainly because their LVS deck's support for BJTs is pretty poor. Since the BJTs in the bandgap reference all have their base+collector tied to ground they're equivalent to a diode (at least as far as SPICE is concerned), right?
Thanks,
Last edited: