2
\$\begingroup\$

I need 3 opams for a circuit. One for amplification, one for level shifting and one for sharpening badly shaped pulses, For the last part(pulse sharpening) I want to use a comparator/schmitt trigger. Can I just use LM324 for all since it has 4 opams inside or I need another IC?

\$\endgroup\$
  • 1
    \$\begingroup\$ That would depend entirely on your application, but it is definitely an option for consideration: analog.com/media/en/technical-documentation/application-notes/… \$\endgroup\$ – EwokNightmares Mar 27 '15 at 14:44
  • \$\begingroup\$ what i do not understand is that some IC is sold with a name Schmitt Trigger. If any opamp can be used as schmitt trigger, what is so special about the ones sold with that name? \$\endgroup\$ – user16307 Mar 27 '15 at 14:47
  • 1
    \$\begingroup\$ Schmitt trigger comparators have resistors inside them and usually can switch very fast, unlike an op-amp configured as a comparator. \$\endgroup\$ – Andy aka Mar 27 '15 at 15:12
  • \$\begingroup\$ so do you mean it matters when we want a very sharp rising edges? \$\endgroup\$ – user16307 Mar 27 '15 at 15:41
3
\$\begingroup\$

A Schmitt trigger is simply a comparator with some positive feedback.

You can usually use an op-amp as a kind of comparator (some are better than others) and add resistors to get positive feedback. Op-amps generally don't make very good comparators in terms of speed especially (they can be much better than purpose-built comparators in terms of precision). The reason for this is mostly that op-amps are deliberately slowed down (compensated) with an internal capacitor to make them stable with feedback. So, they behave quite sluggishly to a fast-changing input. Also, many op-amps, of which the LM324 is not one, do not behave well with much differential input voltage applied (they may have a relatively low input impedance if the differential input voltage exceeds a diode drop or two). This is not normally much of an issue for op-amps, but for comparators it can be. Also, some op-amps will drift slightly in offset voltage when a relatively large differential voltage is continuously applied (with a time constant in the months or years).

For a ST, if you're simply feeding back the op-amp output for the positive feedback, then the output voltages of the op-amp when "railed" form your reference voltages. This is imprecise, especially if the op-amp is not rail-to-rail output (the 'high' from the LM324 for example) so you have to evaluate if that's good enough. Switching more precise voltages for the feedback is not that much more difficult, if necessary.

Bottom line though, if you need a rise and fall time of <100ns to clock a counter, and to be able to follow a 1MHz input, a normal op-amp is probably not going to cut it.

A Schmitt trigger gate effectively has the positive feedback built in (usually different sizes of MOSFETs rather than resistors for CMOS logic), saving components and allowing quite high speed, however the specifications as to trigger points and overall hysteresis are usually quite loose. If you want something that will trigger on at 3.00V+/-5mV on the rising input and off at 2.00V+/-5mV on a falling input, you'll need something more precise than a ST gate.

\$\endgroup\$
  • \$\begingroup\$ what if the signal is maximum 1000Hz and Im interested in pulse frequencies? For those frequencies is LM324 reliable instead of Schmitt Trigger? \$\endgroup\$ – user16307 Mar 27 '15 at 15:45
  • \$\begingroup\$ Well, it should switch but the slew rate is such that it might take a few hundred microseconds for the output to change from one rail to the other (depending on overdrive and supply voltages). That's a significant part of a 500usec half-cycle. \$\endgroup\$ – Spehro Pefhany Mar 27 '15 at 17:21
2
\$\begingroup\$

You can make most op-amps act like it has Schmitt trigger inputs by adding hysteresis.

enter image description here

The advantage here is you can set the trip points to be anything you like. They are controlled by the three resistors R1, R2, and R3. Here is the formulae used:

$$Vh =\frac{Vs*R2}{\frac{R1*R3}{R1+R3}+R2}$$

and

$$Vl = Vs * \frac{\frac{R2*R3}{R2+R3}}{R1+\frac{R2*R3}{R2+R3}}$$

Vs is the reference supply. Vh is the upper threshold, and Vl is the lower threshold.

Here are the same formulas suitable for input to a spreadsheet if you want to experiment with the values:

= ((A10*A2) / (((A1*A3) / (A1+A3)) + A2))

= A10 * ((A2 * A3) / (A2+A3)) / (A1 + ((A2*A3) / (A2+A3)))

where A1..A3 are the resistor values R1..R3, and A10 is the reference supply.

\$\endgroup\$
  • \$\begingroup\$ So you cannot set adjustable trigger levels in Schmit Trigger? \$\endgroup\$ – user16307 Mar 27 '15 at 15:46
  • \$\begingroup\$ Well I was wrong. I though you could get op-amps with preset hysteresis. I'm unable to find any. You can get logic gates with Schmitt trigger inputs -- e.g. 74HC132 but not apparently op-amps unless someone else knows of one. So you have to use external resistors as shown above, which do allow you to precisely set both high and low trigger points. \$\endgroup\$ – tcrosley Mar 27 '15 at 16:28
  • \$\begingroup\$ Not any op-amp. For instance, TL084 can (depending on the input) exhibit phase inversion in this very configuration. \$\endgroup\$ – Peter Mortensen Apr 3 '15 at 15:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.