The LF357 has 12nV/√Hz which is swamped by the Johnson noise of the 100k resistors on the input and feedback network which contribute nearly 60nV/√Hz.
Don't use series resistance if you want low noise, and keep the feedback network low impedance. Try losing the input 100k completely, and changing the feedback network to use 10k and 1k rather than 1M and 100k.
Don't use series resistance if you want low noise, and keep the feedback network low impedance. Try losing the input 100k completely, and changing the feedback network to use 10k and 1k rather than 1M and 100k.
The 3403C is a 100 MHz TRMS meter. Futzing with the inputs would most likely degrade its performance. The 100 MHz means the noise floor is necessarily high. Its not the instruent you would use to check for noise. Here is a picture of its insides https://antiqueradios.com/forums/viewtopic.php?f=8&t=369437
The LF357 is not supposed to be stable at a gain of 1.1.
For low noise, resistors in series with the signal source should be small and those in parallel should be large, so making a large input impedance is no problem: just use a small or preferably no series resistor and a large resistor to ground.
Protecting the amplifier against excessive input voltages is another matter, series resistance could help for that.
For low noise, resistors in series with the signal source should be small and those in parallel should be large, so making a large input impedance is no problem: just use a small or preferably no series resistor and a large resistor to ground.
Protecting the amplifier against excessive input voltages is another matter, series resistance could help for that.
I conclude that swapping opamps is only marginally going to solve the problem. The 10mV noise I mentioned above should be 1.5mV.
The board is made from SMD's and hard to get by and changing resistors is challenging.
Only the opamps are on sockets.
In the end, it is decided to live with it.
Thanks everyone.
Hugo
The board is made from SMD's and hard to get by and changing resistors is challenging.
Only the opamps are on sockets.
In the end, it is decided to live with it.
Thanks everyone.
Hugo
Sure Mark, but my friend said it wasn't worth the trouble anymore, for this little meter has become more of a museum piece. 
He would have taken the trouble to swap opamps but not alter hard to reach parts because he has to take apart
the pieces that are build on top of each other.
He is grateful with your help and fully agrees with the resistor theorie.
Hugo
He would have taken the trouble to swap opamps but not alter hard to reach parts because he has to take apart
the pieces that are build on top of each other.
He is grateful with your help and fully agrees with the resistor theorie.
Hugo