Calculating This Circuits Frequency

SgtWookie

Joined Jul 17, 2007
22,230
Bill,
It's really a matched precision amp for really low frequency stuff, like instrumentation amplifiers.

I just got done testing the real thing in operation. I adjusted the spice simulation to what my real-world measurements were, using my laboratory-quality ultraprecise gee-whiz $2.99-on-sale Harbor Freight meter :rolleyes: but the results on my Hitachi 'scope were so close to the LTSpice model, it's hardly worth taking photos.

Caps were measured at 99nF each. I didn't bother changing them for the 1% difference.
716.3 Hz LTSpice-predicted frequency
684.9 Hz Measured frequency (my 'scope is not calibrated)
Still, that's well within 5% of what the simulation said.Waveforms were nearly identical, except there was a slight clipping of the bottom of the sinewave. I'm blaming that on the 0.31v difference in the absolute values of the rails.
And no, I didn't put 0.1uF caps across the power pins like I should've (bad Wook! Bad, bad...) but they weren't in the proposed schematic, either. So there. ;)
 

Attachments

JoeJester

Joined Apr 26, 2005
4,390
The best advice Texas Instruments can give is to not do this. The very best a designer can expect is very poor performance; the worst is a circuit that does not work or even burns out.
--SLOA067, Op Amp and Comparators, Application Report by Bruce Carter
Interesting Application report concerning this topic.

Bill, my simulations also showed a lower frequency, but to get mine to establish the operating point, I needed to put some positive feedback on the op amp acting as a comparator.

Then there was the hystersis issue. Using the LF353 NS spice models, the positive zero crossing didn't start transistioning the output till the signal crossed approximately 22 mV, while the negative didn't start transistioning till about 50 mV.
 

Thread Starter

Wendy

Joined Mar 24, 2008
23,415
At this point I think we're fighting slew rate. Until I run my experiments I'm keeping my mouth shut though. If you had asked me this weekend I would have said I knew where my capacitor bag was, looks like a trip to Radio shack. I'm sure the parts will turn up, after I need them.
 

SgtWookie

Joined Jul 17, 2007
22,230
I'd set it up with a T.I. LM324 last week, and was so completely disgusted with the results that I put it on hold for awhile. :rolleyes:

I may have simply miswired something. But where I was expecting close to 700Hz out, I was getting 400Hz and a very asymetrical series of waveforms that didn't look much like the simulation. It's also possible that the IC is damaged; I've had a bunch of them kicking around for years.

Something has been bugging me; that's the triangle wave output getting lopped off. It seems to work fine if run at very low frequencies, but with the component values I'm currently using, it's clip city.
 

Thread Starter

Wendy

Joined Mar 24, 2008
23,415
I'm almost done with mine, and plan on using some 1458s too. More to follow.

Everything is taking longer than it should, situation normal in other words.
 

Caveman

Joined Apr 15, 2008
471
I've seen the triangle lopped off in a couple of the simulations, but like I've said, change the op amp and the simulation will completely change.

I'm still 100% convinced that my analysis is correct, but I've seen no one else that agrees. Nor has anyone given any reasons as to which it is not correct.
 

JoeJester

Joined Apr 26, 2005
4,390
I've seen similiar results as Caveman with the simulations. I've used the circuits outlined in TI's and NS's notes as well as a very old design of Forrest Mimms in the 1979 Engineer's notebook.

I agree his analysis is sound.

To date, I've spent most of the time reading the different application reports, notes, and looking at the bias currents and slew rates of the various devices.

It's been an interesting read.
 

Thread Starter

Wendy

Joined Mar 24, 2008
23,415
Like I said, I'm waiting for the breadboard results. The main problem I have is the square wave isn't square. This introduces a significant delay throughout. If the comparitor were swapped with CMOS (you'd probably have to parallel them) to generate a decent square wave I'd be more of a believer. 0.1v/us slew rate is slow, extremely so. As is the square waves look like extremely clipped triangle waves, and the zero crossing isn't.
 

SgtWookie

Joined Jul 17, 2007
22,230
has anybody tested the circuit with the triangle wave let to reach the peak?
:)
Take a look at my LTSpice results in the first post on the prior page; my actual results were nearly identical.

I'm also getting clipping on the triangle wave with an LM324, but am having other problems with that opamp.
 

JoeJester

Joined Apr 26, 2005
4,390
Well, I used a LM324 for the oscillator.

I didn't have any 2.4k resistors, so I used 2.2k. I measured each passive element as I constructed the circuit and changed the simulation to whatever was measured. I also recalculated the frequency.

Anyways, the comparison went like this:

Calculated:
Triangle Wave: 1135.8 Hz
Sine Wave: 1129.2 Hz

Simulated:
Square Wave: 819.7 Hz
Triangle Wave: 819.7 Hz
Sine Wave: 819.8 Hz

Measured with a freq counter:
Square Wave: 750 Hz
Triangle Wave: 750 Hz
Sine Wave: 750 Hz



Anyways, attached is the simultion and the Triangle output from the scope.

On edit:

The simulation shows a delay of 27.9 uS [slew] in the negative transistion, while the actual circuit shows about 8 uS at my best guess on the 100 uS scale using x10. The positive transistion shows about 12 uS on the scope, while the simulation is 29.3 uS. I really need to get some new TE. LOL
 

Attachments

rwmoekoe

Joined Mar 1, 2007
172
Take a look at my LTSpice results in the first post on the prior page; my actual results were nearly identical.

I'm also getting clipping on the triangle wave with an LM324, but am having other problems with that opamp.
i mean, what if we let the triangle wave form a complete triangle, not clipped. because, i can't simulate it in my microcap it's always general protection fault on xp. i was just limiting the square wave at a lower pos and neg peaks to let the triangle reach the peak. please post the result if you could get it done, sarge :), curious here..
 

JoeJester

Joined Apr 26, 2005
4,390
i mean, what if we let the triangle wave form a complete triangle, not clipped. because, i can't simulate it in my microcap it's always general protection fault on xp. i was just limiting the square wave at a lower pos and neg peaks to let the triangle reach the peak. please post the result if you could get it done, sarge :), curious here..
I have, but not with the values specified. I'll be posting the results shortly.
 
Top