I’m building a new VCO and making some improvements compared to my first one. I noticed that the square output misses a bit of sharpness that my first VCO has. Also the breadboard prototype of the second version sounds pretty good to me.
I’ve implemented the square wave by using a comparator, which gets a triangle wave and a reference voltage. By tuning that reference voltage, I can do PWM. I used a 220nf capacitor and a 470K resistor to do AC coupling (1.5hz cutoff) so that I have a nicely centered signal before I feed it to my comparator. I take this output signal and use a 100k/22k voltage divider to get it down to a lower level before sending it to the output opamp buffer.
I’m expecting the opamp output to switch very quickly between -12V and 12V, but apparently, its not that quick. I’ve managed to capture this using my cheap scope:
So as you can see, the rising edge takes more than 0.2ms to complete, which is obviously audible. I’ve compared this with the output of my breadboard prototype and that one rises much faster, but also not instantly. I thought this happened in a matter of micro-seconds
I’m curious what kind of things can be done to fix this. I can live with the fact that my oscillators are not identical, but I’m curious to learn how you would approach improving this.
200 µs seems too slow even for fake TL074s but I dunno, could be at least part of the problem if they came from AliExpress or eBay or the like.
To be clear, your AC coupling is before the comparator, right? And nothing after the comparator but a voltage divider and an op amp buffer? No capacitors in any of that?
Added: Scope pictures from an article linked in the above topic:
The genuine TL074 comparator is turning on in about 2 µs. The fake is much slower — but it’s a whole lot faster than 200 µs.
I did get my TL074’s from AliExpress. I wasn’t aware that that could be a problem. The one in my breadboard prototype seems to respond 2-3x as fast, but if I understand you correctly, even this is pretty slow.
Do note that this scope of mine is definitely not the best equipment, so there might be some error there
Makes me feel that all my modules I built so far could be upgraded by replacing all the opamps with better ones. Perhaps I should order a whole load from tayda instead.
Hard to believe that’s the only problem here (if indeed it’s a problem at all) — but comparing a more reliably sourced TL074 to your AliExpress ones (before swapping them all out!) is probably a good idea nonetheless. Some say Tayda isn’t a reliable source but I haven’t noticed any problems.
But it seems like 200 µs would have to arise mainly from something different. If there isn’t a capacitance you haven’t mentioned, it’s puzzling.
For reference, this is the output of my oscillator which I consider “good”. It’s probably still shit haha, but at least it has a lot more high-end. The edges now take ~30 us which is much better.
I just added coupling capacitors (single 10uf on both +12V and -12V), but that didn’t do anything. There has to be a capacitance problem. I think so too. Should I cut the strips on my stripboard shorter? Seems like its probably something more serious than that. I will do some conductivity tests.
Any recommendations on where I can order reliable opamps?
Although my opamps are far from ideal, they are not the cause of the problem. I’ve swapped the one on my problematic stripboard with one I used on my breadboard prototype and the problem was just as bad.
The only thing I can come up with is parasitic capacitance on the stripboard. Does this sound plausible? I did not cut the strips for the power-rails so there’s a 12V and -12V strip running right next to the strips carrying the signal. I bet someone has experience with this if this is the cause of the problem.
Stripboard capacitances are typically a few picofarads, so at audible frequencies I’d blame the circuit (as implemented, not necessarily the schematics, but we’ve seen neither) before I’d blame the stripboard.
NOTE: I see that there is a missing ‘break’ (what do you call this) under the TL074. In my soldered version, I did break up the strip between all opposing sides of the TL074
The top-right opamp is the comparator. The output it sent into a 100k/22k voltage divider before being sent to the buffer on the top-left.
I do realize that the voltage divider spans a big portion of the circuitboard (mostly the 22k resistor). I did disconnect the 100k resistor at some point and soldered a wire directly to the opamp output. The problem was still there unfortunately
Pins 4 and 5 are the output and CV input for the on-chip pulse wave. Since you’re making your own pulse wave externally they’re not used. It’s not obvious to me whether it’s best practice to leave them disconnected or tie them to ground or what. But they presumably have nothing to do with your op amp problem.
I’m all in favour of innocense until proven guilty. Making remarks like this may give people a bad reputation that they do not deserve. Until you can present proof that all of the hundreds of traders on ebay and aliexpress (or the like) sell fake stuff, I think this is rather misleading.
This post reads (to me) simply that we have seen many people complain about sellers from both of these sites for exactly this reason.
There is being charitable, and there is defending things that have been shown to have a pattern of risk.
I hardly think anyone is saying every seller on those sites are peddlers of problematic products, but to ignore the actual real risk isnt doing anyone any favors either.
To settle it, cross over distortion is clearly visible when the wave passes zero volts. This is a key indicator that you have lm324s (or similar) marked as tl074s.
Unfortunately getting remarked IC’s from these sites is now all too common. I’ve seen a bunch of people waste a lot of time on remarked chips sold as a relatively uncommon audio codec ic, because it was on back order with the usual trusted vendors.
There are good parts to be brought from these sites though, you just need to be wary!
Interesting, learned something new today and it’s only 7 am!
The output stage of the LM324 has low output impedance when it sources output current (when the ‘push’ half of the push-pull stage is active), and when it sinks output current (when the ‘pull’ half of the push-pull stage is active). It turns off (goes high impedance) at zero output current, and that causes a ‘dead spot’ in the transfer characteristic.
You don’t have to defend the Aliexpress and eBay sellers, they’re not your friends, and their supply chains don’t provide much traceability (usually zero traceability once it reaches the store).
I hope you understand that most of the points you raise do not necessarily hold in the case at hand.
You are assuming that what is shown in the video is correct and not fabricated by the person making the video. And you extrapolate from that to the current situation.
This is at best supposition / a hypothesis but that does not necessarily make this a reality.
Hmm, have you checked this or do you just believe their statements are correct? In a court of law this kind of reasoning is called “hearsay” and not admissible.
I did not say this because I thought I needed to defend anyone. Ever heard of slander and being sued for it? Aliexpress and Ebay could maybe feel that the suggestion made could be slanderous.
[edit] my original response was (in part) meant as a warning having this last point in mind.
