There's a problem I'm trying to solve where I need to limit the incoming voltage on a signal. This could be AC or DC. I found a schematic that handles a similar problem by amplifying the signal with a TL074 so that the upper limit of expected incoming voltage, +-5V, saturates the opamp. Would this be a proper way to accomplish voltage limiting? Are they any downsides?
Some amps don't like saturation at all, causing all sort of issues.
But if the amp you're using is fine with this, just take care to not exceed the input voltage limit.
Another way would be to clip the signal with some clamping circuit, a TL431 should do it with decent accuracy (but at least 2.5V).
TI makes a couple of limiting opamps especially for this kind of application -
OPA698 and OPA699.
Cheers
Alex
TI makes a couple of limiting opamps especially for this kind of application - OPA698 and OPA699.
Cheers
Alex
Interesting chips, but expensive.
Try to avoid saturating the output of "normal" op amps: it can take a long time for the amplifier to recover.
In the old days of op amps running from split power rails (+/- 15V), in the inverting configuration, there were a lot of "bounding" circuits to limit the amplifier output to less than saturation.
A simple method was back-to-back Zeners across the feedback resistor, but the diode capacitance was often too high.
Other circuits used fast diodes along with Zeners.
In a quick search, I couldn't find anything good online: I learned this stuff almost 50 years ago from Burr-Brown books on op amps.
I'll try to post a possible circuit diagram later.
I can see the point of using TL074 for this because IIRC its inputs remain high impedance even when taken many volts above the positive rail.
Beware that it has a nasty surprise at the opposite side: the output suddenly jumps positive if the noninverting input is very close to the negative rail, or below.
With all that being said, maybe it would be faster to just solve your problem than think about a bizarre old trick.
What signal are you talking about? How large, how fast?
What's the source, can it tolerate simple clamping with zeners to ground, for example?
Where is the signal going, what's the input impedance that must be driven?
How much (in)accuracy can you tolerate?
I can see the point of using TL074 for this because IIRC its inputs remain high impedance even when taken many volts above the positive rail.
Beware that it has a nasty surprise at the opposite side: the output suddenly jumps positive if the noninverting input is very close to the negative rail, or below.
With all that being said, maybe it would be faster to just solve your problem than think about a bizarre old trick.
What signal are you talking about? How large, how fast?
What's the source, can it tolerate simple clamping with zeners to ground, for example?
Where is the signal going, what's the input impedance that must be driven?
How much (in)accuracy can you tolerate?
Signal could be any here from DC -12V to +12V, so the input would be DC coupled. The goal is to clip any coming voltage that is coming in where it is > -+5V. It's possible that the signal input could also be low frequency under audio range. The VCC and VEE for any circuits would be _+12V and -12V respectively. The signal is going into specialized IC, the AS3360, a voltage controlled amplifier, that accepts a voltage on a control pin from -2V to +2.5V. Here's a link to the datasheet for the IC.
https://alfarzpp.lv/eng/sc/AS3360.pdfThe signal of interest would be going into pins 5 and 10 of the chip, the linear control voltage.
If this helps, this is a circuit from the analog computer days that I often used to limit the output of an opamp (here shown inverting) without saturating its output.
The two Zeners are always conducting, so their capacitance doesn't matter.
The extra resistor and back-to-back diodes require a large current through one of the other two diodes before passing feedback current to the virtual ground (- input of op amp).
The feedback current through the diode network keeps the output within limits, but without saturating the amplifier.
so far grey beards with out of date methods, and beginners with no idea....
clipping to the rails in 2023? 6c rail-to-rail input and output dual opamp. Inverting amplifier, gain set to ensure the input will never take the - input outside the rails.
And those who reject previous methods out-of-hand, yet cannot punctuate.
And those who reject previous methods out-of-hand, yet cannot punctuate.
I believe; the only correct answer to that, would be, says you.
Blankly criticising non-specific and off topic aspects, without any examples of how to correct that? Asshat.
Your punctuation actually makes your statements hard to read.
clipping to the rails in 2023? 6c rail-to-rail input and output dual opamp. Inverting amplifier, gain set to ensure the input will never take the - input outside the rails.
How would that solve my problem of clipping anything > +5V or < -5V?
clipping to the rails in 2023? 6c rail-to-rail input and output dual opamp. Inverting amplifier, gain set to ensure the input will never take the - input outside the rails.
How would that solve my problem of clipping anything > +5V or < -5V?
It would help if you shared the schematic of what you are thinking of using. As mentioned above by DavidAlfa, saturating an opamp not designed for operation in that mode can cause undesirable effects (including inversion). Which is why there are rail to rail amplifiers which are specified for operation up to (sometimes a little beyond) the supply voltages without any unexpected side effects. Operating parts outside the specifications can lead to unexpected results, so best to avoid that.
Operational amplifiers specified for saturation are cheap, so why not take advantage of them? What is special about the TL074 and circuit you are considering that you could change to something specified for saturation?
Here's the proposed input stage. The signal would be amplified by around -2, so that anything above 5 volts coming in would clip at 10V saturation. The next opamp would scale the signal down so that the max output would be 2V. Capacitors are there to filter out anything above 22 kHz.
What about a signal diode bridge with a Zener across the bridge in the feedback loop of a simple op-amp in inverting configuration. This would limit the +- outputs to equal magnitude levels if the signal diodes are somewhat matched.
Edit: Here's the concept using a 1N750 Zener, 1N4148 & OP07, don't have a model for a ~4 Volt Zener (1N750 is ~4.7V) which should limit around ~ +-5V. Simple, but not fast.
Best,
The limiting scheme depends on the end use and precision required. Hard or soft limiting? Unipolar or Bipolar?
Idle speculation belongs on TikTok.
Opamp output headroom is not a well defined or stable parameter.
A railed out amplifier willl have recovery time issues.
As will the simple clamp shown here
Interesting chips, but expensive.
If you want it cheap, here it is.
Cheers
Alex
Not sure if your circuit will still be stable with high signal source impedance or capacitive output loading.
I think I would put Miller around Q3/Q6, perhaps including the diodes as I think it would make reaction time essentially nothing.
Not sure if your circuit will still be stable with high signal source impedance or capacitive output loading.
I think I would put Miller around Q3/Q6, perhaps including the diodes as I think it would make reaction time essentially nothing.
According to LT Spice the circuit is stable with 100K on the input and 100nF on the output
. Caps around Q3/Q6 don't work that well (I've tried in the sim).
Cheers
Alex
Thanks. Interesting chip, but too pricy.
Is it a hobby project, building one piece, or a commercial endeavor?
Both TI and Analog have free sample programs.
Thanks. Interesting chip, but too pricy.
Is it a hobby project, building one piece, or a commercial endeavor?
Both TI and Analog have free sample programs.
Looking over the specialty ICs that have been suggested, the TI chip has a very low voltage it will control, and the Analog chip uses more current than I can spare. I'm building a device that could have as many as eight channels. With the Analog chip, being single channel and consuming ~20mA, that would be way more current than I can spare. Plus, costing $13USD a piece is beyond my budget.