This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

OPA228: How to trade off gain and bandwidth?

Part Number: OPA228
Other Parts Discussed in Thread: LMP7717, OPA838

Hi ,

     Customer's product is a high accuracy ac power line quality  analyse equipment. They want to know whether the GBW is always constant in any condition?

They sch have 4 stage amp,every stage  gain is 10. The OPA228 GBW is 33Mhz ,Dose it mean  if the frequency less then 3.3Mhz the gain can be guaranteed by 10?

Or can I using the formula as below picture and the gain is 5 at input frequency equal to 3.3Mhz?  Which is right? 

Thanks

  • You might also look at the LMP7717 for a higher GBP decomp part, 

    A TINA sim will show nominal closed loop bandwidth. The GBP idea to predict closed loop bandwidth only applies if the phase margin is 90deg. Figure 4 in this article gives the bandwidth extension beyond that for Phase margin < 90deg. 

    https://www.planetanalog.com/author.asp?section_id=3404&doc_id=565056&

    Another option for you to consider if you don't need more than a 5V supply part is the OPA838 - really low noise for a 300Mhz 1mA device - not as good on DC though. 

    Be careful with cascaded high gain stages, normally I do pi filters on the supplies from the output stage back towards the input stage, but other folks prefer star filtering. 

  • Hi Jeff,

    The OPA228 is a decompensated Op amp and its open-loop gain-phase response deviates from a perfect first-order gain roll off and phase response. When you examine the datasheet open-loop gain/phase plot you can see an obvious ripple in the phase between 1 and 2 MHz, and there appears to be another pole breaking around 2 to 3 MHz, and a zero between 10 to 20 MHz. A ripple in the phase such is often an indicator of gain peaking in the response at that point.

    When I construct four stages of OPA228 each operating with a gain of 10 V/V using TINA Spice, the following gain vs frequency response of obtained.

    You will note the gain peaking just above 2 MHz, that becomes increasingly higher with each additional stage (blue is stage 1 output, olive is stage 2 output, etc.). Therefore, the peaking makes it more difficult to determine the -3 dB bandwidth. Using the brown curve it looks to be about 3 MHz. However, because of the overall gain of 10,000 V/V the output swing of the final stages is likely going to be large signal - volts peak-to-peak. If that is the case, the OPA228 slew rate of 11 V/us will likely become the limiting factor.

    Some of the High-Speed Op amps that Michael suggested would likely provide more satisfactory performances if speed and bandwidth are required.

    Regards, Thomas

    Precision Amplifiers Applications Engineering