Give your gizmo the gift of sight
I have a question regarding the Stonyman Vision Chip.
If I run the chip with 3.3V will I have any problems or disadvantages?
According the datasheet:
Power supply voltage - VDD=3.0V to 5.0V - Operation below about 4.0V requires the use of the amplifier to implement level shifting
Is there any disadvantage when I use "Level shifting" ?
It is better to use the Stonyman chip at the highest voltage, generally 4.5V to 5V.
When operating from 4.0V to 4.5V, you don't need to use level shifting but the output range reduces- you will get less dynamic range.
When operating below 4.0V, you have to use the amplifier to do level shifting. The disadvantages of this mode are that you have to adjust VREF if you move from a very dark to a very bright environment, or vice versa, and operating the amplifier slows down the chip- you need a few microseconds additional to operate the amplifier so this will slow down the maximum pixel acquisition rate. This doesn't make it "impossible"- it is just a few more things to worry about.
Future chips will be able to operate at 3.3V without the level shifter- it is an easy change to the circuit.
Which part of the stonyman circuits have the highest power consumption? The amplifier?
Even if you bypass the amplifier on the stonyman chip, it looks like that the amplifier is always powered. Bypassing the amplifier does not contribute to lower power consumption of the chip. Correct?
Another thing I do not understand is Bias generator. From the stonyman's datasheet, it is not clear what's the function of NBIAS and AOBIAS. Could you give me more hints here? Thanks.
The output amplifier (the one that buffers for the output, not the one that provides gain) has the potential to draw the most power if you put a low impedance load on it. Otherwise I believe it is the focal plane readout circuitry itself, though I have not tested it.
AOBIAS vs. NBIAS: NBIAS is used to bias the column readout circuits. This is how the pixel signals get from the focal plane to the gain amplifier. The AOBIAS is used bias the final output buffer amplifier. So there are actually two bias generators, one for each bias.
Generally the purpose of the bias generators is to modulate a current source that causes an amplifier to operate "properly". The default values set should be adequate. But you can sometimes speed up the chip by raising one or both bias voltages (by lowering the numbers stored in the bias generator register circuits). This comes at the cost though of reducing the dynamic range of the chip by reducing voltage swing.
Do you mean that the NBIAS is used to power the column readout circuits?
Because the bias generator is used to modulate a current source to power the amplifier, I can reduce the system power consumption by setting small bias values. Is this assumption true?
Yes and Yes- though remember that the bias generators generate a SMALLER voltage when given a LARGER number. 0 produces the highest voltage = highest current, while 63 produces the lowest.
The amplifier on the stonyman chip can be bypassed by setting selamp to 0. Is it possible to shut down the amplifier module on the stonyman chip? Thanks.
The amplifier was not designed to be shut off, but it shouldn't draw very much current- probably no more than 100uA.
You *might* be able to reduce the current a little as follows:
1) keep selamp = 0
2) set VREF = 63
3) hold INPHI high.
I haven't tried the above yet. If you do try that, let me know how it goes.
When we keep selamp=0, set VREF=30, and hold INPHI high, the current consumption of the Stonyman chip is 0.826mA. Then we hold INPHI low and keep the rest the same, the current consumption of the Stonyman chip is 0.71mA. Keeping the camera in reset state increases its current consumption...
Wow- Thanks for the detail! I took a closer look at the chip schematic and what you report is correct- the chip draws more power with INPHI high.
A future chip we are working on will draw less power.
Thanks Geof. We are investigating the Stonyman's behavior when it is operated in sub-threshold regime. For example, how the camera behaves when its voltage is below 3V. The sub-threshold operation reduces the camera's power consumption at the cost of slightly worse image quality.
I'd be interested in seeing what you get, if you are willing to share it (openly or closed).