LM7171: RF amplifier osicllation

Hi Fred,

Unfortunately, I do not think corner case testing or some kind of Monte Carlo sim will show much but we can definitely get back to this if we can't figure something out. I was curious if the customer can share an image of the units that are working in this same circuit.

Best Regards,

Ignacio

Hi lgnacio,

sure, I will try to get some top marking picture from them

would you help me run a simulation on the phase margin of this system with respect to our IC's variation?

we're having a meeting with customer this Thursday night, hope we could have some TINA or SPICE simulation can help back up 

Hi Fred,

This is not something we could capture. Our models are not designed to adjust a value without basically redoing the whole model for any variation you could see in the real device. The concern I see with this circuit is the feedback capacitance which could cause oscillations as it would ideally turn this circuit into a buffer at higher frequencies. This device is not stable in a gain of 1V/V, which this feedback capacitor will cause at higher frequencies. Could the customer remove these feedback capacitors and see if the oscillation goes away.

Best Regards,

Ignacio

Hi Ignacio,

is there a equation that could show the potential of this 10pF would bring stability ?

to be honest, I don't know how they achieve negative feedback with R28 connect to positive input as well, hope you could help me clarify it more

Hi Fred,

The concern with this feedback capacitor causing stability issues is because this device is not unity gain stable. At higher frequencies this capacitor will behave like a short and will effectively turn this circuit into a buffer at higher frequencies. The image of the units that are working in their system would be useful to confirm what might be happening. The circuit itself is very interesting to get differential outputs, but I was able to get the expected results using a standard model in TINA. I was wondering if they have considered trying the LM6171, as this is a unity gain stable amplifier in the same family as the LM7171. There is also a dual version of the LM6171. 

Best Regards,

Ignacio

Hi lgnacio,

sorry, may I know why LM7171B is not suitable for unity buffer in terms of high frequency?

and is LM7171A also not suitable either?

and I'll forward to customer

BTW,

the good marking at right 

Hi Fred,

This device was designed to work in gains of 2V/V or better. The A and B version are the same core device. The LM6171 version is unity gain stable which should behave better with a feedback capacitor.

Best Regards,

Ignacio

Hi Lgnacia,

from the quality perspective, may I know how the good working part top marking helps to clarify this issue?

Hi Fred,

This device was redesigned in one of our new processes and the older device does have slight differences in behavior although it is a direct drop-in replacement. I am curious if maybe the devices that do work as expected are from the newer revision. If so, then it is much easier to understand why this has been happening to the customer. 

Best Regards,

Ignacio

Hi Ignacio,

Goi it， but they said this medical design has been used and fixed for more than 7 years with LM717B without abnormal until recently, not sure if this give you more clue on quality perspective.

On the other hand , this design was not meant for unity buffer , they did want to amplifier the 1.5MHz sinewave instead of buffering, and the feedback capacitor was for stability , does it make sense to you? And is it ok they change to use LM7171A on this design from now?