OPA445 acceptable replacement for HA2-2645-5 ?

[rigrunner]

I'm looking for a suitable replacement for a HA2-2645-5 as used in the Datron 1271 input amplifier.

The OPA445 is half the GBW at 2MHz, a higher slew rate and being FET input a lower input bias.

If this isn't a good substitute any suggestions?

[Kleinstein]

The compensation for the JFET amplifier part (the 82 pF capacitors and  G of the JFETs) looks quite fast (e.g. 2 MHz). So the speed of the OP may be relevant in that circuit. A slower OP may need some adjustment to the compensation (e.g. about double the resistors in series to the 82 pf capacitors).  It looks like the circuit uses a bootstrapped supply and such circuits are often a bit prone to ringing / oscillation.

There are not that many high voltage OPs around and the OPA445 is not that far off.

[rigrunner]

The compensation for the JFET amplifier part (the 82 pF capacitors and  G of the JFETs) looks quite fast (e.g. 2 MHz). So the speed of the OP may be relevant in that circuit. A slower OP may need some adjustment to the compensation (e.g. about double the resistors in series to the 82 pf capacitors).  It looks like the circuit uses a bootstrapped supply and such circuits are often a bit prone to ringing / oscillation.

So with a bit of tweaking it should at least work.

There are not that many high voltage OPs around and the OPA445 is not that far off.

This is the problem with finding a replacement. There aren't many op amps about that can happily hang across 70V.

The 1281 has a different approach. It uses an OP27 with a GBW of 5-8MHz ( maybe you're right about the speed being important?) with a push-pull circuit following it.
I have thought about replicating that on a piggy-back board, but that push-pull circuit gets quite toasty.

[Kleinstein]

The DA1281 amplifier is different in that it has only a chopper path an no parallel higher frequency path. The extra dicrete build amplifier stage also has some gain, which complicates the compensation. The Chopper stage has a somewhat limited gain and the amplifier after that should still have a reasonable low drift.

The time constant from the 220 Ohms and 82 pF is likely to counteract some slow down from the OP. So there is a chance that this may need some adjustment with a different OP. This does not look like it is directly for the GBW, but it could be a 2nd pole that effects the phase at around the GBW and beyond.
I see a good chance to get it working with a few tweaks. Worst case one could switch the compensation to handle either the x1 or x 50 case separately.

[rigrunner]

Just checked the values of the installed parts and they don't match the schematic. 82pf and 221r have been updated since the schematic to 68pf and 100r.
Do the updated values shed any further light on what they are trying to achieve with them?

[Kleinstein]

The 82 / 68 pF are the main compensation for the amplifer loop. They act together with the FET transconductance.
The 220/100  Ohms are likely to counteract one pole in the amplifier, to give it a little phase boost so that one could push the frequency limit a little further. My fist though was this could be related to the OPs GBW, but this would be at lower frequency. Usually there is some unwanted pole somewhere above the GBW of the OP and my guess is that this is the one to counteract. This tweak likely depends on the details of the OPs and maybe other parts used - so even if one would get another OPwith 4 MHz GBW, one may still need to rethink the resistors and find the best value experimentally.

It may be a good idea to run a simulation (e.g. LT spiece) of the amplifer. The Chopper part should be reasonable separate (lower frequencies only), so it should be OK to leave that out. The result of the simulation depends on the models and would thus not be enough to get a final value. However the simulation would help to get an idea how critical the values is and get an idea how the resistor would effect the response, e.g. ringing with a square wave input.