I have two problems to solve and hopefully this can be done with a single (dual) part.
1. Supply a "clean" -15 V at about 35ma to a transimpedance amplifier from a -38V source,
2. Provide a steady-state -38V and (2) a timed sweep from 0V to around -5V. Here, nanoamperes of current are involved.
I considered an LDO for -15, but those I can find won't take the -38V presently coming from a converter and we don't want to create another noisy switching converter!
So I'm thinking a "high output current" amplifier (dual package) powered by our available -38V (with the + rail of the amplifier to system ground) would be an option. One of these would produce the -15V based on bias at the inverting input, and the other driving the -38V rail with a mosfet-selected triangle generator for the timed sweep.
Does Analog Devices have a suitable part for this idea.. or other ideas that will give us these voltages?
Look at ADA4700-1.
Those are two interesting problems. There are a few high voltage opamps to take a look at but I suspect none of them meet your needs for one reason or another. Harry's recommendation is probably the best: ADA4700-1: Single 100V precision opamp. However,35mA would be pushing the output drive capability of the ADA4700-1 close to its limits. ADA4700-1 wouldn't work for your second requirement since is not a rail-to-rail output. I'm not sure what you mean when you say provide a steady -38V and a timed sweep from 0V to -5V. Those seem mutually exclusive.
All in all, I still think you're better off with a switcher and an LDO. I might suggest the ADP5071 and ADP7182. What generates the -38V. Does the -38V have a lot of variability or just switching noise?
Harryh and Dave,
I looked at the ADA4700-1 and considered the same concerning the current limit. I'm powering a IVC102 transimpedance amplifier and am looking for an alternative to the +-15 v power module in our original design (12V in +-15 out with 45ma quiescent current - it's that standby current that kills us). Although the IVC102 uses less than 35ma, some headroom would be good. I've gone to great lengths trying to clean up both rails (LC filtering).
So decided to stay with our module, and use the ADA4700-1 as our -38V buffer (drives a static plate, so nanoamps of current here), fed with a single -38V supply. Rail to rail here would be 0 to -38. Hmmm.. might I have to split the supply a bit? I need to sweep the output from 0 (gnd ref) to -2V at a period of 60 seconds. I suppose this can be easily done by configuring the device as an integrator. But, from the same chip, we also need to turn on the -38V full bore. Maybe bypass the integration cap for that move...
Anyhow, good suggestion and one I'm planning to stick with, unless you folks have better ideas.
ADA4700-1 is not a rail-to-rail output opamp, i.e. output swing is limited to 1-1.5V from either rail. If you need to drive 0V, then the Vs+ needs to be +1.5V or higher. Pay attention to the capacitance of the load (static plate) to maintain stability.
Do take a quick look at the ADP5071. It is a dual (independent output) DC-DC converter. It can accept your 12V input and provide adjustable bipolar outputs up to +/-39V with only 4mA of quiescent current. It's a very nice part.