Stairstep generator using 555 and op amps

[lordvader88]

I made this Stairstep generator using 555 and op amps, according to w2aew's version



A good schematic shot is at 1:12s

I used a LM358, and it didn't work any amount. Those never seem to work well in any type of op-amp oscillator I've tried, so time to order some modern op-amps.

So I used a TL072 and it basically works. However the step-level plateau's are not staying level, so the 100nF cap at pin5 of the op-amp is discharging when it's meant to be fixed.

Looking at the base of Q2 it only turns fully on (~0.7V) for under 10% of a cycle, to discharge the cap. And then it seems to gently slope from 0.6V to 0.45V or so iirc. And so the 1st 2-3 stair steps decay more like a stepped sawtooth, and the last 2 or 3 are still not very level, as Q2 turns off a bit more.

The same base signal drives Q1 for the 555-reset, it turns on/off just fine., making a nice square wave.

I made this on perf board, with 2n3904 and I haven't tried any other transistors.


I've started trying the DC analysis for the oscillator, I'm really new at that on the rigorous level. So IDK what to expect the base voltage to be yet, or how to tweak it.

Any idea's ? I want to use this to make the curve tracer Allan makes by adding more sections.

Also I only had about 6-7 steps I think, there's 14 in the video. I used the same value caps/resistors, and used a good plastic potted type 10nF cap for the 7555 (the modern CMOS 555). That's quite a difference . I should have made this on a breadboard for easy parts changes, but I love soldering stuff.

[t1d]

I have played around with building a function generator... Mostly, I have learned how the circuit works and traditional approaches and ICs.

I do not know how those circuits compare, to what you are doing, with a 555. But, you mentioned considering other Op Amps, so I just wanted to say that a lot of slightly better, basic, Function Generators are built around the XR2206, or the AD8038. FGs built on LM324s seem to perform less well.

[spec]

Hi lordvader88,

Looks like the opamp is the problem. You need an opamp capable of working with an input down to 0V, like the OPA191.

But, can I just advise that if you are making a curve tracer, better to go for a more accurate and flexible circuit.

I would recommend a 74HC4060 counter oscillator chip and an opamp. If you are interested in this approach just say, and we can hammer out your requirements and then point at/design a circuit for you.

[lordvader88]

Yeah at 1st it worked a little bit w/LM358, but then i did something wrong with the 7555 part of the board, and it also wrecked the desired B-side of LM358.

Everything on the 555-circuit seemed to measure ~ right/expected with a DMM to me. But something was very wrong, some soldering mistake, or maybe it's the chip, I have fried 2-3 of them and I can't say that isn't 1 of them, I must check that chip.

I re-built the 7555 circuit on the proto-board, checked it worked, then fully soldered it.

And it at least works, but only as a saw-step generator, not a nice level stair step


IDK really know much DC or AC analysis of op-amps when caps and diodes are included, I need to refresh Thevenin Eq/node voltage/mesh current, because I still make silly mistakes with them.




I varied Vcc, (not over 11V yet tho), and that made little effect. I tried a cheap SMPS and a 60Hz linear LM317 regulator, and that didn't seem to change much on my oscilloscope.



Also when trying to 'figure out' these circuits on paper, without having a formal education on it, a lot of situations egg up like the 'chicken and the egg' type situations, and IDK if it's the Rooster. I can do most the calculus/diff.EQNs/linear algebra, but I don't have any EE textbooks. So I'm missing a lot

[lordvader88]

And the TL072 seems to be 0.5V to Vcc-1.5V, and that was running off 5V to 10.8V , so that can't help. IDK my op-amps, yet.

w2aew mentioned about rail-to-rail op-amps, so he means op-amps that actually get very close to R2R.


I have to order a bunch of newer and more expensive op-amps

[David Hess]

It could certainly be done with a 324/358 type of operational amplifier but the rail-to-rail input and output and low input bias current of the LMC6482 make circuit design much easier.  If you are looking for a general purpose rail-to-rail input and output operational amplifier with low input bias current, then the LMC6482 (dual) and LMC6484 (quad) are one of the better choices but do not rely on them for precision.

The step generator designs I am more familiar with like in the various old Tektronix curve tracers use a capacitor-diode charge pump to drive an integrator and would be more suitable for a single supply design.

[lordvader88]

Ok so does " rail-to-rail " mean that they actually go V(neg supply) and V(pos supply) on the output ? Where as a LM358 or TL072 are about 0.5V and 1.2-1.5V off on either end

And are u the guy that made the video, or do I just recog or name anyways

thanks anyways


So the base of Q2 is stuck at 0.45V when the TL072 goes low. I have made a LM317/LM337 split rail PSU, could I use it to help somehow offset the 0.45V low output ?

And what about if I just set up a higher base voltage...there's an idea....I do the basic transistor eq's 1/2 often.  But IDK the op-amp oscillator side well enough yet to mess with it...I shouldn't be so cheap and lazy, I should just try more stuff like engineer's/experimentalist's

I know where to find the basic op-amp eq's for this, but that cheating because I can't figure it out on my own

[rstofer]

You aren't going to get there without using a rail to rail op amp.  And, remember, 'rail to rail' is a marketing term, it doesn't really work that way.

Look at the datasheet for the LMC6482 - it gets within 20 mV of the rail ONLY if the load is on the order of 100k Ohms.  And that's just in the upper left corner "Features", look at page 10 for a LOT of gotcha's.

http://www.ti.com/lit/ds/symlink/lmc6482.pdf

Nevertheless, these op amps get a lot closer to the rails than some of the older variants.  As long as the transistor cuts off, getting down to the absolute rail is not important.

[David Hess]

Ok so does " rail-to-rail " mean that they actually go V(neg supply) and V(pos supply) on the output ? Where as a LM358 or TL072 are about 0.5V and 1.2-1.5V off on either end

Rail-to-rail means the outputs closely approach the supply rails.

And are u the guy that made the video, or do I just recog or name anyways

No, that is not me.

So the base of Q2 is stuck at 0.45V when the TL072 goes low. I have made a LM317/LM337 split rail PSU, could I use it to help somehow offset the 0.45V low output ?

Using a split power supply can considerably relax the circuit design and part requirements.

[lordvader88]

So the cap is draining some amount when Q2 should be off but is gently sloping from 0.60V to 0.45V, that's while the op-amp is low. The 1st 1/2 the op-amp is doing it's thing, just it only drops to logic low of 0.45V. IDK how bad the other side is doing, or what problems the reduced high output causes on the op-amp or oscillator equations. I haven't tried any this time

How much is C2 draining from any problem with the 7555 and the other half of the 8dip IDK, I could measure some currents, with my Keithley 195, the datasheets of these are scary, I never know what leakage currents I should use , so I don't bother

[rstofer]

As the revised circuit shows, it can operate from a 9V battery or anything higher than about 5V.  Redesigning it to use split supplies is counterproductive.  You would also need to rethink that reset timing circuit because it would be pulling the timing capacitor to well below 0V.  Timing would change, and so on.

There is a user 'w2aew' (alan) and he posts here quite often.  His videos are excellent.