THS4601: The same characteristics as the THS4601, but push the bandwidth to about 40MHz

I would also like the alternative to the THS4601 to be a soic with the same pin out

Morning Patrick, 

The transimpedance app depends strongly on the source capacitance - what is yours? 

Also, the normal situation is that you do not need much output voltage swing in this stage, so using the lowest supply consistant with that will open up a wider range of more power efficient solutions in the decompensated VFA world - which is preferred for transimpedance solutions. Min supply? 

Ideally I have a PCB with a fixed supply of +/-12V, the transducer supply is unwilling to confirm either the shunt resistance of series capacitance of the transducer. The current amplifier has two stages, the transimpedance amplifier and the output driver with a little gain, some of the better known current amplifiers use tha same configuration, I back engineered a broken one. I could switch of the power to the transducer and measure the discharge time, if I get a measurement that is in the region of photodiodes, i might be on to a winner, i could try that and get back to you.

from calculating back from the waveform discharge i would say about 80pF, that's my best estimate

I think 80pF even through large will be in the right ball park considering the number of Dynodes and charge of the transducer

So a quick calculation of required GBP will give 80GHz if you need 100kohm from 80pF with 40MHz F-3dB in a single stage - not possible. So yes, a 2nd stage gain will be needed. 

You will find faster devices at lower supplies (OPA818) but if you must apply +/-12V maybe the THS4631 - you are really restricting your solution if +/-12V is mandated. 

I can see the advantage of having all your gain in one stage it reduces your noise in the amplifier. I would like ideally at this stage to just replace the THS4601 and see what I get, one step at a time, then I could do a comparison between the performance of the amplifiers. Thanks for the THS4631, I thought of the THS 4211 as part of the equation

In theory yes, in practice though the 2nd stage adds very small input referred noise. Once upon a time I tried to detail all of that here, 

https://www.edn.com/controlling-spot-and-integrated-noise-in-a-two-stage-transimpedance-design/

The THS4601, 4631 and 818 are all JFET input - the THS4211 is bipolar requiring quite a lot more input bias current and having its noise as well. and again, for transimpedance, there are strong advantages going decomp. 

And then the starting point for transimpedance design, 

0550.Transimpedance design flow using high speed op amps.pptx

The transducer I'm using has a peak output current of about 500uA, and the THS4211 has an input impedance of 4MR, not 100K as you would get with bipolar TTL, that's why I thought it a good bet. I would also like to achieve a virtual input impedance near to 50R, with a lower input impedance between the differential inputs might help.

The op amp input impedance is principally important for the loop gain analysis but should be driven to zero by the loop gain as you look into the inverting node. So really only the C is important not the R normally for the op amp open loop input impedance. 

Your comment about 50ohms moves away from classic transimpedance to something else. Not sure where that is going?

The transducer we get has a short RG179 coax lead, not very long, but you get into transmission lines. My comment about the virtual input impedance should have been at DC. I can see the capacitance changing the input impedance of the opamp, as a two port network. I can see it being a comprimise as to the feedback resistor, compensation, and input capacitance when viewed as a four port network.

Kind of reading between the lines you have an existing box with a THS4601 input stage and 20MHz BW. You would like 40Mhz from a somewhat unspecified source. Perhaps an easy thing to try is to reduce the THS4601 gain by 4X (in theory doubling its BW for the same source in a Zt design) and recover that in the 2nd stage increasing its gain 4X perhaps changing that device if need be. 

Yea that's it. Gives me something to Mull over for a couple of days, thanks for your help, being the Technical prime here, I need to reach out to sort out what i'm doing.

By the way I tried changing the gain must be some internal poles in the THS4601, the gain is rolling off too fast, I can't push the gain out far enough for output driver

Hello Patrick,

  I would also highly suggest going with Michael's advice. Lowering your power supplies if you can, would lead to a more amplifier options. But here are more options for alternative to THS4601 but higher GBW: product list link.

Thank you,

Sima

Hello Patrick,

  I am glad we were able to help with some ideas! For TIA designs, I do want to emphasis Michael's statement from earlier. Most likely the concern at the input is the internal input capacitors of the amplifier rather than resistive.

  Generally, bipolar input amplifiers are more noise advantageous for lower transimpedance gains, while JEFT and CMOS input amplifiers are better suited for high transimpedance gains. This is because of the tradeoffs between current noise and voltage noise in regards of the feedback resistor.

Thank you,

Sima