Best 8 pin DIP Single OpAmp for Receiver Front End

[baum7154]

I am looking to improve the receive front end amplifier of the CHANCE PI detector that is currently spec'd to use an OP37 amplifier. Many improvements have evolved since the OP37 was created and I think there is room for improvement here. I have been looking at the Analog Devices ADA4895, 4897, & 4899 products. Priorities are High Speed, High Bandwidth, Low Noise, low offset voltage, and secondarily low power consumption from a single supply. What other parameters should I be considering? What other devices would anyone recommend for consideration? I appreciate your assistance!

Thanks,

Dan

[Davor]

Pick just about any. I'm not kidding.
Trouble with Chance front end is that it is noise limited by resistors at it's input. You have waaaaaay over 1k resistors in series with signal source, so it's input referred noise is over 4nV/sqrt(Hz). So just about any op amp declared as "low noise", including TL071 (!) will have sufficient low noise response. There is a servo taking care of the offset, so even offset is not a problem.
To get a better noise you'd have to do some serious makeover of the front end.

[baum7154]

Thanks for the response. Is the schematic view you included from the CHANCE PI or some other detector? I don't recognize it as the CHANCE front end.

[Davor]

I stored it for my reference and it is indicated as Chance. The king of this design is ATmega8-16PC, and schematic date is 29.09.2010

I guess there might have been some changes to this design, so whatever you have just put it on here so that we may see if there is anything else you can do.

In short, with PI you are time limited by coil discharge that needs to be short (requiring high impedance) followed by a period of low noise sampling (low impedance). Easy designs use constant impedance, and they all suck at both speed and noise. Every design that does not do either input switching, or some sort of active input impedance ... thing, will not benefit from extra low noise front end. BTW much cheaper than op amp, the extra low noise front end is achievable with garden variety BJTs

[WM6]

Hi Davor, I guess that baum7154 referred to this Chace schematic:

http://www.silverdog.co.uk/images/ChanceA.pdf

[Davor]

OK, this one is a bit better, yet it is also limited to 4nV/sqrt(Hz) because of a series resistor of 1k. So OP-37 is as good as it gets. It is also fast and has input bias compensated, so it is a very good choice.

[baum7154]

Hi Davor, I guess that baum7154 referred to this Chace schematic:

http://www.silverdog.co.uk/images/ChanceA.pdf

Yes this is the schematic of Chance with only the 1K resistor at the input from the coil to the opamp.

Dan

[baum7154]

OK, this one is a bit better, yet it is also limited to 4nV/sqrt(Hz) because of a series resistor of 1k. So OP-37 is as good as it gets. It is also fast and has input bias compensated, so it is a very good choice.

So there is nothing more to be gained with faster slew rates or the greater bandwidth of a better opamp?

[baum7154]

Pick just about any. I'm not kidding.
Trouble with Chance front end is that it is noise limited by resistors at it's input. You have waaaaaay over 1k resistors in series with signal source, so it's input referred noise is over 4nV/sqrt(Hz). So just about any op amp declared as "low noise", including TL071 (!) will have sufficient low noise response. There is a servo taking care of the offset, so even offset is not a problem.
To get a better noise you'd have to do some serious makeover of the front end.

The schematic you have in file for CHANCE is definitely not the CHANCE, perhaps one of the other designs from Fandy as he has several out there.

Thanks

Dan

[baum7154]

Yes this is the schematic of Chance with only the 1K resistor at the input from the coil to the opamp.

Dan

The link for the schematic above does not include the five modifications required for firmware Version 1.2.1. The modifications can be found at:

http://fandy.hut2.ru/Chance.htm

[Davor]

The noise limiting resistor is still there, so it is not changed. 1k resistor is quite large for an average PI with simple front end, so I guess it would not work well unless a coil is made well. A diode that is added in parallel with 470k has more impact on speed of recovery than the speed of the very op amp and it is a nice catch. Otherwise, this is the most you can possibly squeeze from this topology. Front end switching would give you more, but at cost of complexity.

[6666]

Davor any chance that we could see the complete schematic from post #2 ?, thanks.

[Davor]

I have two, but I have no idea if you'll be able to access these - forum acts oddly. I have no Idea where I picked these because I failed to include url of the source.


Chance_AVR.pdf

[6666]

Thanks Davor will try and have a look when forum is back to normal, cheers.

[baum7154]

The noise limiting resistor is still there, so it is not changed. 1k resistor is quite large for an average PI with simple front end, so I guess it would not work well unless a coil is made well. A diode that is added in parallel with 470k has more impact on speed of recovery than the speed of the very op amp and it is a nice catch. Otherwise, this is the most you can possibly squeeze from this topology. Front end switching would give you more, but at cost of complexity.

Thanks for the input. I have seen in other posts, possibly one from BB Sailor, that lowering the input resistor value may yield some performance improvement assuming a good coil. I think I have a pretty good coil and may try a drop of the resistor to about 500 ohms. Small gold is my target and any performance improvement is welcome.

Dan