Hi all, I?ve got a circuit with that uses an opa2134. I find myself pus hing the 36V supply limit and I?m a little worried. So I went looking f or a similar opamp but with a bit more headroom. So specs might read,
Dual opamp unity gain stable
4(+) Mhz BW (8MHz or more would be nicer)
10 V/us slew (again more is better)
20mA current (positive, again more would be nicer) small input C (
Well I might be using 'headroom' the wrong way. I'm splitting a 48 V wallw art, ground is a few volts above the bottom rail and then ~35 volts down from th e top for the positive rail. But 35 volts designed and 1% resistors that's about a 4% max error... 36.4 volts. I'm feeling squeezed.
Thanks I'll look at those. (I rejected the MC33072 already.) I've got some TCA0372's in stock, but haven't played with them yet. (too slow for this job.)
Make sure there are no transients possible on the power supplies etc. if you do this. I once had a really horrible experience using 16V TI CMOS op-amps at the system supplies of +/-8V (almost 1% failure rate).
You can get a lot of soft failures due to overstress, but the part keeps "working." A soft failure in a chip is something like the shutdown current out of spec or the input impedance is low, etc. Often chips have soft failures and nobody notices. For instance, a bit of increased shutdown current can be lost if there are a lot of parts on the board. A lower impedance on an input may not matter if the rest of the components are low impedance.
But your problem sheds light on an interesting test problem, namely absmax. There is no standard on how to test parts beyond their datasheet limits. Some companies only QA absmax. They stress the part and see if it still meets spec when used at stated voltages. Others do a real absmax test, but not at strict electrical limits. But then there is the question of should you be stressing your parts at all?
But to get 99% failure when operated at the datasheet limits implies to me a part that has other problems rather than your supply being the culprit. There is a lot of margin in absmax.
That said, transorb diodes are highly recommended!
Grin, the first time I did the circuit I had big caps (10uF) from the inpu t 'grounds'* to the regulated voltages. (LM317/337's) The positive supply w as fine, but the negative supply would 'turn on' to ~1/2 the supply voltage . (Caps basically split the supply to start with) and then went down as th e regulator took over. This took out a voltage reference downstream ~50% of the time. (Bad cap!)
George H.
*'ground' for the LM337 was the positive supply input.
Another opamp that you can't push is HFA1130. It's spec'd for 12 volts, and it will die at 12 volts.
Older analog processes, like for classic 741 and LF353 types, would fail way past ratings, numbers like 60 volts. Newer parts cut things tighter. Or maybe specmanship has got more intense.
--
John Larkin Highland Technology, Inc
jlarkin at highlandtechnology dot com
Ahh, well there's nothing wrong with the supply. I just need 30V of stable bias. The LM317's had too much drift and so I've now got a 10 Volt reference gained up to 30V.
A 317, or an LM1117, makes a great power stage. It still current and thermal limits. You can put a zener in series with the adjust pin so the opamp only has to swing a few volts. Then you can power the opamp from the 1117 *output*.
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
I've often used 317s to run temperature control loops, where I could make the heater run on high enough voltage that the 3V dropout didn't kill me. LM1117s have half the dropout, which helps.
Re higher voltage supplies:
I just built one of those self-oscillating half-bridge supplies with a
1CT : 2CT ISDN transformer wired as an autotransformer. Unloaded, it saturates at about 750 V*us, and at 150 kHz it puts out a watt or two at
+-50V with + and - voltage doublers.
The half-bridge topology has one nasty feature--if the core saturates just once, it instantly discharges the coupling cap, which doubles the volt-seconds on the next half cycle, so it saturates on every half cycle thereafter until the FETs cook. That needs a bandage or else a 2:1 derating of the volt-seconds, which is a pain. Alternatively I suppose I could just reduce the size of the coupling cap, which hurts the maximum power output. (I suppose there's a Royer-style solution there somewhere as well, but I'm using both windings for other things.)
No sound effects this time, but I've officially joined Joerg's FET demolition derby club. ;)
able bias. The LM317's had too much drift and so I've now got a 10 Volt re ference gained up to 30V.
Ohh, that's a nice idea.. can I still power the opamp from the LM317 outpu t? In this case I wanted a new voltage reference after the 'ground' and V(+) w ere defined by the regulators. I gain up the 1.25V's by factors of 20-30 a nd any change in the 1.25 V gets magnified. Once I get the 30 volt bias I' d like it stable to ~1mV or so. (For maybe 10 minutes to an hour.) So anyway the current circuit works fine... (the pcb is spun and in the box .) I don't want to redo it. I guess I should think of all the gotcha's before doing the design.
Got it.. that's good to know. I've used the LM395 as an opamp power booster.. I think Jim T. said that the guts are similar to a 317. Hmm I don't need any power I could have just stuck almost any transistor on the output. OK maybe next time.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.