Designer’s Guide Community

taofeng Junior Member Offline Posts: 22 Belgium	GBW and UGF Jun 23^rd, 2006, 1:07am are they the same in value ? can anyone make it clear for me ? thanks a lot
Back to top	IP Logged

Jerome New Member Offline Posts: 6 Toulouse, France	Re: GBW and UGF Reply #1 - Jun 23^rd, 2006, 12:06pm Hello, In general GBW=UGF. It is true for a system with a dominant pole and all other poles and zeros located well beyond GBW (or UGF). For example, it is the case for a simple Miller opamp (with PM>60deg). If you have poles and zeros within the GBW, then the relation GBW=UGF does not hold. I hope this helps, Jerome.
Back to top	IP Logged

Croaker

Senior Member

Offline

Posts: 235

Re: GBW and UGF
Reply #2 - Jun 24^th, 2006, 6:04pm

Jerome wrote on Jun 23^rd, 2006, 12:06pm:

Hello,

In general GBW=UGF. It is true for a system with a dominant pole and all other poles and zeros located well beyond GBW (or UGF). For example, it is the case for a simple Miller opamp (with PM>60deg).
If you have poles and zeros within the GBW, then the relation GBW=UGF does not hold.

I hope this helps, Jerome.

Yes, for a single pole Adc*f3dB=unity_gain*funity. This is true because a 20 dB/dec decrease means the gain is going down 10x for every 10x increase in frequency. So, a DC gain of 100 and pole at 10 Hz gives GBW=1000 and it also gives a gain of 1 at 1000 Hz, so UGF=1000.

Back to top

IP Logged

aamar

Community Member

Offline

Posts: 57
Germany

Re: GBW and UGF
Reply #3 - Jun 26^th, 2006, 1:23am

Hello,

This question has confused me once before, because as I tried to calculate the GBW as a UGF from the AC analysis and then compare it to calculator output from the GBW product function I found that the results were different . As I read in the calculator manual it calculates the GBW product as the product of the DC gain and the first pole frequency and I was using the 0dB intersection with the gain curve.

What to be noticed that although the PM is above the 45 degrees but its effect on the gain curve starts taking place at a frequency decade before, and it makes the curve bends with additional 20dB/decade and so it will intersect with the 0dB earlier than what expected, this makes the unity gain frequency in most of the times smaller than the gain bandwidth product.

So inorder to make them equal the second pole should come at least one decade after the GBW product (PM>=90 degree), so that to ensure that the gain slope was constant at -20dB/decade till it intersects the 0dB level.

Best regards,

aamar

Back to top

IP Logged