[Amps] a 'circuits' question

[Dan Sawyer]

Steve, thank you for your reply.

I have 8 working parts. 3 are nearly identical in bias and measure a 
gain of 4.5 cm +or- .1 cm. A fourth part has an identical gain but a 
bias that is significantly different. The other parts have very 
different bias.

Are 2 cycles sufficient to determine a common gain?

If the gain is common will a bias that is off by a cycle step create a 
problem?

Would selecting a part with a 10% difference in gain but similar bias 
create a better balance?

Thanks,
Dan

Steve Thompson wrote:

>On Thursday 09 September 2004 03:21, Dan Sawyer wrote:
>  
>
>>Steve,
>>
>>Thank you. I was trying to drive the MOSFET in the linear region. I
>>figured gm would be accurate for the linear region and that bias
>>adjustment would be accurate for a particular initial condition. I was
>>speculating that even given those two points the shape of the drive
>>curve might not be identical. Is this true??
>>    
>>
>
>It could be true although, in my experience, mismatch has to be very 
>significant before there's any appreciable effect on overall performance 
>(assuming matching for bias voltage, or separate bias controls). If you do 
>the threshold and gm tests, then any pair with gm within 10% from the same 
>batch code is likely to be fine. For dynamic matching, a gain selection at 
>the intended operating power is probably all you need.
>
>Steve
>
>  
>