Wideband 1Mohm to 50ohm unity gain Amplifier circuit

[Win_Hill]

Sorry for the intrusion in this thread, but when I saw Win_Hill posting, I have to ask:

Will there ever be an update on "The Art of Electronics"?
https://www.amazon.com/The-Art-Electronics-Paul-Horowitz/dp/0521370957

Oh, OK. Yes, We've finished the entire 3rd edition and it's been copyedited and is now passing by us, going back into the hands of our publisher where it will progress towards printing; it should be out early next year. It's got lots of awesome stuff, even if I do say so myself. After it's published there'll be a second more advance book, with extra material, that we're calling "x Chapters", so it can have the same chapter numbers as in the 3rd edition. E.g., Take Chapter 4 on Opamps, there's a chapter 4x with extra advanced material about using opamps. This advanced-material book will come out a bit later.

 

[E-design]

It sounds very good. Looking forward to it. I hope to see you posting here more often.

 

[Win_Hill]

Yes, the ths4631 is a respectable part. But in this 100MHz territory it may be better to stick to low-voltage parts, to keep the power dissipation down if nothing else. I do not have direct experience with either one, but looking at the graphs on the datasheet, the response plot of the opa659 I suggested looks a little better. One advantage for the ths4631, having an soic-8 package, it can be placed on TI's evaluation board.

Judging by the part number, the OPAxxx part was designed by engineers at TI's Burr-Brown division, whereas the THSxxxx part was designed by its local staff. If we were also considering an LMxxxx op-amp, it'd have been designed by the NSC engineers in California. Let's hope the TI managers don't decide to downscale, and that they keep these three groups hard at work designing good linear stuff for us to choose from.

 

[audiomik]

Well, I do have a copy of your Second Edition here. Comes in very useful when trying to explain bits of basic theory to others, especially those I've forgotten and have become more 'assumed knowledge' over time.
Looking forward to your next edition.

Back to the Thread title, my choice of circuit does relate to the potential for 'home construction' in post #18.
Also, having looked into it further, it should be very possible to use a single supply rail with AC coupling together then with the tab of a TO220 package for cooling the BUF634 at 0 Volts also doubling as a mounting point for the PCB into the case.

Purpose of this device is to be able to have a very low load on a circuit on test whilst shareing it's output between various instruments as in this pic of my 'Bench':


Mik

 

[FvM]

Yes, the ths4631 is a respectable part. But in this 100MHz territory it may be better to stick to low-voltage parts, to keep the power dissipation down if nothing else. I do not have direct experience with either one, but looking at the graphs on the datasheet, the response plot of the opa659 I suggested looks a little better. One advantage for the ths4631, having an soic-8 package, it can be placed on TI's evaluation board.

OPA659 is a newer part, so I didn't yet use it. It looks interesting indeed. The supply current is rather high, however.

If pulse transmission respectively time domain performance is an objective, none of the discussed voltage mode OPs might be sufficient.

Purpose of this device is to be able to have a very low load on a circuit on test whilst shareing it's output between various instruments

Signal distribution at 100 MHz will usually require impedance matching, e.g. a low impedance buffer with individual 50 ohm series resistor for each signal output. BUF634 would be able to drive multiple 50 ohm outputs, but the distribution amplifier will hardly fit a small BNC adapter case.

 

[audiomik]

Re: "BUF634 would be able to drive multiple 50 ohm outputs, but the distribution amplifier will hardly fit a small BNC adapter case."

A distribution Amplifier might be a useful variation to the original idea, especially if it can be done using the same circuitry and PCB artwork as the for the in-line BNC box.
Will add that as a 'possible' to the design

thanks
Mik