Apex HV opamp issues.

See if you can get Pimpom's email addy -- he's out in the sticks somewhere, but maybe you could ship it to him for the repair.

(or have them toss it in the dumpster and just send a new one?)

On a sunny day (Fri, 13 Mar 2015 13:45:11 -0700) it happened John Larkin wrote in :

Thats is a keeper.

It likes capacitive loads, which become the dominant pole in the feedback loop. With that sort of compensation, more C load just makes it more stable.

Ooooh, I wish I were still working at FLIR. Our head analog engineer was excellent, but very averse to odd topologies and to circuits that he didn't think would simulate well. Tossing this schematic in front of him in a meeting would be like tossing a cat into a cage full of antisocial dogs.

If he was adverse to odd designs, how did the E45 software wind up so weird? It's not a regular camera interface, it's a network device over USB with a bizarre cable.

If you use it one one USB port, and change ports next time, even a different slot on a hub, the PC software forgets that it exists.

The thermal part works pretty well.

Yep, most of the APEX opamps made with MOSFET input transistors are very noisy, especially at low frequencies. For example, the PA141 spec says 50uV rms, over a 10kHz bandwidth. By comparison the PA341 spec is 337uV rms. Whoa!! Their JFET-input opamps are much better, for example the PA97 specs 2uV rmw over 10kHz. So you could change to one of those types.

Or you could make your own HV amplifier, using a low-voltage JEFT opamp to control MOSFET power transistors. We show an example in Figure 3.75 in AoE 2nd edition (also see Figure 6.47). These amplifiers can be even more quiet. I updated the circuit for AoE-III, see Figure 3.111. [Note, AoE-III is being printed and will be shipping in a week or two.]

A more advanced circuit is set for the AoE x-Chapter book (for the time being it's Figure 4x.50 in section 4x.18), where there's a lengthy careful analysis of the design, and its performance into large capacitive loads, etc.

A refined version of that circuit is in development for a freebie PCB, so y'all can easily make your own. It's my Rowland EE Labs AMP-62A project, and here's a schematic:

The MP118 is JFET input, and noise is modeled in my Spice subcircuit. ...Jim Thompson

I don't understand the Q8-Q10 string. Is that safe?

dl=0

Absolutely. Avalanche is your friend, provided you keep the power dissipation under control. If we assume about 600V avalanche, at 0.5mA that's a very relaxed 300mW. The lowest-current series current-source avalanches first, then the next highest, and so on. So the current-source varies by about maybe up to 0.1mA over voltage. But hey, that's far better than using 2W carbon pullup resistors!

Another scheme I like is the MOSFETs wired in cascode, but this requires six series resistors to handle 1kV. But it does let you double or triple the pullup current.

LND150 data sheets don't mention an avalanche rating.

Looks like Microchip acquired Supertex.

Right, when you see an official "avalanche" rating, it's actually a thermal-mass parameter, for use with pulsed high-power levels. It's consistent with and derivable from the Transient Thermal Impedance plots.

But in this circuit we're not subjecting the part to any transient high-power dissipation.

Oops, we'll see what happens next. Fingers crossed.

You may have noticed (lower left corner of the schematic) my enthusiasm for Infineon BSS126, instead of "my favorite" LND150. It's a higher-current, higher-voltage part. BTW, these SOT-23 MOSFETs have different pinouts. Just saying.

I'm pretty sure that the handheld stuff is still all done in Sweden.

I worked on the high-dollar airborne imaging stuff. While we did make some profoundly stupid products from time to time (usually heralded by a company officer getting all of us into a room and telling us how smart he was to kick off the project), more often than not we made good chit.

When I started there, a mechanical drawing had to be checked by two people in addition to the engineer who designed the part. A circuit had to be checked by one, other than the engineer. Software had to be checked not at all -- you just gave them a floppy with a hex file and signed on the dotted line. I took that to mean that software was considered to be magic, not engineering (other aspects of engineering management's behavior kind of confirmed that).

Wow! thanks, That's much more HV amp than I can afford to get my head around. (at the moment) (I'm glad JL ask about Q8-10)

I'm going to try the LTC6090. My only issue for a retro fit is throwing away about 20-30 V of HV supply (@ 10 mA max.)

A silly resistor divider takes too much power. I was thinking of a series Zener (1 Watt) I then wondered if I could run a three terminal voltage regulator (LM317) With the adjust pin driven from the input side?

Which lead me to the voltage divider cap multiplier.

Vin---+---+----|-----+ R C 1 --+-- l l | | o o +-----+ a a | | d d R C | | 2 1 | | gnd--+-----+-------------+---+

About 1mA flowing through R1/R2.

George H.

Just the final update on apex opamps... (I'm still waiting on DK, but I had to leave early today.) It turns out that I'd forgotten the first apex update, where the PA141 went to the PA241... so here a noise plot (noise density is uncalibrated.. though I have the gain settings.)

Turns out the opamp got progressively worse... What's weird is that the "1/f" noise is not 1/f, but f^-2/3.

Well since I don't understand 1/f, (1/f)^2/3 is no weirder.

George H.

Oh, I should add that the x-axis "bandwidth" is the low pass filter setting, and the signal was AC coupled with a 10 Hz HP.

We discuss in detail how this can happen in our lengthy Chapter 8 in Art of Electronics 3rd edition, 40 tons of which is shipping now from the printer. It's largely due to offset shifts (i.e., popcorn-noise style) as well as large thermal offset-voltage drifts.

A good solution is a "composite" amplifier, where the poor amplifier is wrapped with a good opamp.

Excellent. As a send-off we need a re-write of Flanders and Swan's 20 Tons:

Come the ghostly Magi bearing.. 40 tons of AoE..

Best regards, Spehro Pefhany

Hi Win, I assume you are talking about 1/f noise ...(and not the 2/3rds power dependence.) I've read some stuff on 1/f. (Van-der-Ziel for one.) I guess the classic 1/f noise is from the old carbon comp. resistors. Here's a plot.... well it's a bit busy, I found a tiny hint of 1/f noise in (cheap, Xicon) metal film resistors.

George H.

IIRC most MF resistors are specified at 1E-7V/V per decade of frequency. I measured some as a voltage divider with a 9V alkaline, battery, and they were nowhere near that bad.

Cheers

Phil Hobbs