Tl084 drops in volume with too much feedback resistance

Started by DCF, March 04, 2022, 01:10:10 PM

Previous topic - Next topic

DCF

Hi, I'm trying to design my first distortion pedal, but alas, I can't make a simple inverting amplifier work. Here's a link to my current breadboard. https://youtu.be/I6gaBNwDvtA

In text, it's an inverting tl084 amplifier with 100ohm in resistance and it works until about 200k of feedback resistance. However, after that, it will either have a large pulse of silence after a note or make no noise at all. This op amp is supposed to be latch-up free, and can clip in a klon without problems. What am I missing?

PRR

Welcome.

No DC voltages?

No schematic?

Movies are not as useful as YouTube wants you to believe.
  • SUPPORTER

Fancy Lime

Dito. Not much we can do without a schematic. But from what you describe it sounds like it could be a DC gain issue. Could the opamp be amplifying a small DC offset to the point of being stuck at one rail?
My dry, sweaty foot had become the source of one of the most disturbing cases of chemical-based crime within my home country.

A cider a day keeps the lobster away, bucko!

antonis

Quote from: Fancy Lime on March 04, 2022, 01:46:47 PM
Could the opamp be amplifying a small DC offset to the point of being stuck at one rail?

With only a resistor (100R) and not a cap, I'm afraid there isn't offset matter only..

P.S.
What about the rest 3 amps..??
"I'm getting older while being taught all the time" Solon the Athenian..
"I don't mind  being taught all the time but I do mind a lot getting old" Antonis the Thessalonian..

ElectricDruid

Quote from: DCF on March 04, 2022, 01:10:10 PM
I can't make a simple inverting amplifier work. Here's a link to my current breadboard. https://youtu.be/I6gaBNwDvtA
What am I missing?
Too many wires!! It's supposed to be a "simple" inverting amplifier, remember?!? What are all those wires doing?!?

I'm pulling your leg, really. We do need a schematic. And you need a schematic, so you can see what you're doing. The video doesn't help anyone see what's going on, unfortunately.

antonis

Quote from: ElectricDruid on March 04, 2022, 04:56:28 PM
It's supposed to be a "simple" inverting amplifier, remember?!?

I should call it "humble" due to 100R input impedance.. :icon_wink:
"I'm getting older while being taught all the time" Solon the Athenian..
"I don't mind  being taught all the time but I do mind a lot getting old" Antonis the Thessalonian..

DCF

After a LOT of feedback, I have made a schematic. I know it's cruddy and hand drawn, but the schematic is simple enough for hand drawn to be sufficient. The wires are a system of quick prototyping, you can blame Brian Wampler for that. And the dc offset bias is a likely story, when I add clipping diodes one at a time, one clips a lot more noise than the other, implying asymmetry. Thanks for the tip!

To the people saying videos are not helpful, I am 100% in agreement when it comes to schematics. However, I figured you might want to hear the sound for yourself rather than me trying to describe it.


ElectricDruid

Quote from: DCF on March 04, 2022, 06:51:34 PM


Ok! That helps a lot, thanks.

How does the negative input get any DC bias if it's blocked from DC by capacitors at both the input and the output? Clue: it doesn't! Take out that 4.7uF next to the 500K and then you've got an op-amp that stands a chance of working. The op-amp will try and keep its two inputs at the same voltage level, but in this case, it can't because it can't set *any* DC level on the negative input (which would usually be held at a "virtual ground"). You need DC feedback as well as AC. This amp can only do AC.

Also the input cap is probably the wrong way around. The negative input should be at Vref, 4.5V, and the input will probably be referenced to ground, so the op-amp end of the cap is more positive than the input end - the reverse of how you've shown it.

After that, you've got a gain of 500K/100R = x5000 = +74dB. That's enough to drive the op-amp completely to the rails with even a small input, and there's no guarantee that the op-amp clips symmetrically or neatly. That's not generally part of the spec. so the "bottoming out" your schematic mentions might well be totally legitimate behaviour for the amp in this situation.

HTH,
Tom

PRR

  • SUPPORTER

PRR

Quote from: DCF on March 04, 2022, 06:51:34 PM...the schematic is simple enough for hand drawn to be sufficient.....

Until the 1970s, EVERYTHING was hand drawn. By the mile! The Golden Gate Bridge. The B-29 bomber. There's real history behind the Merlin engine being re-drawn in the US. The Atlas missile and all the moon-rockets.

Here's a scrap of that technology.


  • SUPPORTER

duck_arse

Quote from: DCF on March 04, 2022, 06:51:34 PM
After a LOT of feedback, I have made a schematic. I know it's cruddy and hand drawn ....

as PRR says, not after feedback, it's the first thing you do - draw the circuit. and it's as good as we need, only missing your IC pin numbers. and a damm sight neater than my hand-drawns, let me tell you.

also, welcome.
Froggy for the cup.

DCF

I can't get the quotes to work to save the life of me, but just know I'm referencing the Electric Druid post above.

Ok a few things. 1.) the cap was a frustration-fueled attempt to make things better, which actually did extend the range of the potentiometer, just by a little. Not saying that to be arrogant, just to state the effects to diagnose the problem.

2.) pulled the feedback cap anyway and used a trim pot to adjust Vref and saw that when Vref was 4.5V, the output was the positive rail. Making Vref 3V or so made the output pin 4.5V. However, throwing clipping diodes on the output (with a cap in between) had a gated, almost fuzz-like effect. I've made a fuzz before that sounded less fuzzy than this. Is changing Vref the best way to solve the problem? I was actually hoping for about 74dB of gain, since I'm hoping to make a dual-rectifier-ish distortion, so I don't want to increase the first resistor drastically. However, the sound isn't right and I can't add a different reference voltage to the input without significantly lowering input impedance or lowering the gain. Is changing Vref the best way to go, should I add a reference voltage at the inverting input and have a buffer at the input, or is there a better solution?

3.) why is there a DC problem in the first place? I see no place for it to come from. I did leave the output without a capacitor, but it was running into a fender amp which certainly has a cap on the input, and it was behaving the same when the amp was on or off. I will add a capacitor in the future, however. Is the breadboard acting as a capacitor or something? Where is the DC coming from?

4.) I am OK with the op amp clipping. It's going to be a hard clipper, no one will hear it after the clipping diodes. The thing is, the klon makes use of a similar op amp in a similar configuration, almost assuredly clipping the rails according to electro smash, but has no problems. Am I doing something differently?

anotherjim

Need to correct the polarity of the input cap. When a polarized one is used, and the input is from the outside world which is nearly always connecting to ground via a guitar pickup or whatever, the cap - lead faces the input and the + lead faces the amp input. As drawn with + facing the world, the cap will leak current and, to the opamp, look like a connection to ground.

ElectricDruid

I'm going to start with this question, because a lot of the other problems stem from this:

Quote from: DCF on March 05, 2022, 09:32:48 AM
3.) why is there a DC problem in the first place? I see no place for it to come from. I did leave the output without a capacitor, but it was running into a fender amp which certainly has a cap on the input, and it was behaving the same when the amp was on or off. I will add a capacitor in the future, however. Is the breadboard acting as a capacitor or something? Where is the DC coming from?
There's a DC problem not because "there's no place for it to come from" but exactly *because* there's no place for it to come from! The op-amp *needs* a DC bias if it's to work correctly. We're running on a 0-9V single supply, so we need the op-amp inputs to be at the 4.5V Vref level. You've got the bias level taken to the op-amp's +ve input, so that's good. The op-amp will try and make the voltage at its two inputs equal (that's one of the basic "rules" we're taught about op-amps), so it will try and bring the output to a level where the -ve input is also at 4.5V. However, in your circuit, that cap in the feedback path stops that from being possible, so the poor op-amp is left struggling to do the impossible. Which won't sound good.

Once the feedback cap is removed, you should get 4.5V on the -ve input as well. At that point, you'll need to flip that input capacitor around, because you've got the positive end connected to a ground-referenced guitar signal, and the negative end connected to 4.5V. You want the positive end to the higher voltage, and since 4.5V is higher than ground, it needs to go the other way around. Jim just mentioned this too.

There's another couple of problems with the input. The first is that the 4u7 and 100R make a highpass filter and since 100R is such a tiny value, the filter's cutoff is relatively high even with a big cp like 4u7. How high? Well, this high:

https://electricdruid.net/rc-filter-calc/?f=&r=100r&r_series=3&r_errors=1&c=4.7u&c_series=1&c_error=10

338Hz. That's going to be audible. Some fuzzes do this deliberately to avoid muddiness by letting too much bass in, but that may no be what you're after.

The second problem is that the input impedance of this pedal is extremely low because of that 100R. The -ve input of the op-amp is what's known as a "virtual ground" so as far as the guitar is concerned that 100R is between the signal and ground. That's not a dead short, but it's pretty close.

I would suggest that you rebuild this amplifier stage as a non-inverting amp so that you can keep the 100R/500K feedback values you want, but also have a high input impedance.

HTH

Fancy Lime

Sorry, I didn't catch that, do you run a guitar straight into this stage? In that case the resistance of your pickup is forming a voltage divider with the (as Tom pointed out) tiny input impedance of the stage. Assuming, say 10k resistance of the pickup means you loose 99% of your signal. The RL filter formed by the pickup inductance and the input resistance will also kill a lot ot treble. Do as tom said, use a non inverting stage for this.

Where does the DC offset come from? Electrolytic caps have leakage currents. In reality they act like a cap with a vert large resistor in parallel, so some little bit of DC can creep through. This depends a lot on the specific model of the cap and also it's age and use history. Normally this is not a problem because the effect is very small. But if you amplify something very small by 5000, it may become a problem.

Andy
My dry, sweaty foot had become the source of one of the most disturbing cases of chemical-based crime within my home country.

A cider a day keeps the lobster away, bucko!

antonis

Quote from: ElectricDruid on March 05, 2022, 11:35:08 AM
you can keep the 100R/500K feedback values you want,

Of course it's matter of taste but 74dB at +/- 4V headroom look more like square wave generator..
"I'm getting older while being taught all the time" Solon the Athenian..
"I don't mind  being taught all the time but I do mind a lot getting old" Antonis the Thessalonian..

ElectricDruid

Quote from: antonis on March 05, 2022, 04:39:43 PM
Quote from: ElectricDruid on March 05, 2022, 11:35:08 AM
you can keep the 100R/500K feedback values you want,

Of course it's matter of taste but 74dB at +/- 4V headroom look more like square wave generator..

True. But if we get to having a working square-wave fuzz with a solid input impedance and sensible biasing, then things have improved from where we started. 8)

Fancy Lime

I finally watched the video. I see at least one un-used opamp unit that is not stabilized. That may very well also contribute to your issues. Pleas connect all unused non-inverting inputs to 4.5V and all unused outputs to their respective inverting inputs. It you have unused opamps flailing about wildly, that can cause the used opamp units to become unstable via their shared power supply lines. Also, for the same reason you might want to add a 100n or so cap right from the Vcc pin of the TL084 to its ground pin.
My dry, sweaty foot had become the source of one of the most disturbing cases of chemical-based crime within my home country.

A cider a day keeps the lobster away, bucko!

DCF

Sorry that I forgot to reply, but thanks to all of you for your responses! Been a busy few weeks, but I have gotten around to fixing my design. It was the leakage of the electro caps that were killing the sound. Appreciate the help!