Operational Transconductance Amplifier Layout

[blue2004STi]

Hey all,

So I am a Senior in college, enrolled in a microfabrication lab. We have a final project to do and we're already being asked to come up with all of the design and physical sizes along with beginning simulations. The reason is so we will for sure have everyone's design by the time we have to start fabrication. Well I chose to model after a operational transconductance amplifier that I'm using for a different project. The background for this class is basically I have to implement everything in pMOS. Our professor has outlawed nMOS, bipolar, diodes, and the like. Now I'm looking over the equivalent circuit and I'm having trouble understanding a few things. I drew up a circuit diagram(yes in MS paint, I'm in a lab that doesn't have one single circuit schematic program).

The black portions are still in BJT as of yet and are what I believe I need to keep. The grey stuff is what I am omitting, but is there to show other connections that may be relevant and the filled red squares are where diodes used to be. So I understand that the red boxed area is a BJT differential input and I obviously the green boxed area is a simple BJT buffer amp. My problem lies with the light blue and purple boxed areas. I can see that they are similar except the purple are npn where the light blue are pnp, so my first thought was "CBJT"(?)lol but that didn't fit in my head. So I thought maybe it's an active load used as a pull up/ pull down network but once again that doesn't seem to fit. I know that there has to be a high gain stage, but contrary to what my professor said I don't see how these are in a cascode configuration. I'm really lost right now. Any thoughts would be helpful.

Thanks,
Matt

 

[Adjuster]

I think these are Wilson current mirrors (look up on Google etc.). They include a cascode element, and one could say that they act as active pull up/down.

 

[blue2004STi]

But if you wanted a pull up, pull down network why wouldn't you use simple enhancement loads? Only two things about these things concern me, one is that only the purple boxes look like any version of a Wilson current mirror I could find. The blue box on the right looks like it is being biased by the "current mirror" below it. Other thing is why are they all connected together in some way or another. I mean I understand that the purpose of a OTA is to be a voltage controlled current source (for lack of better termage), but if they all are current mirrors, then does it mean that the inputs drive the upper two current mirrors and the upper left one drives the lower right and the two right mirror act as a p-up p-down network? Then the only purpose for the one in the lower left is to bias the differential input? Did that make any sense? I guess I'm just a little bit tentative as to why you'd do it that way, but maybe that's how come I didn't invent the OTA,lol

Thanks again,

Matt

 

[Adjuster]

Actually I think you have pretty much got it.

The left hand side differential pair transistor collector current is mirrored once off positive rail V+, then a second time off negative rail V-. This gives a current tending to pull the output negatively. The right hand side differential pair transistor collector current is mirrored once off positive rail V+. This gives a current tending to pull the output positively. The balance between the two currents provides the required current output. The other mirror is for bias.

Look up "Wilson current mirror" and you should be able to find equations and a functional explanation of the mirror.