News:

Forum may be experiencing issues.

Main Menu

Op amp substitution

Started by Jay, September 23, 2020, 01:46:42 AM

Previous topic - Next topic

Jay

I'd like to sub a TL072 for a 4558. Don't have any 4558 on hand. The pinouts look compatible. One question I do have is that in the build docs it shows what Voltage I should expect at each pin for the 4558. Would the voltage be similar for the TL072? Or completely different?


Sent from my iPhone using Tapatalk Pro

Willybomb

TL072, TL082, plus a whole slew of other 8 pin opamps will work fine.  The sound difference is fairly negligible in my experience.

mauman

Voltages at the pins should be the same and they're pin-compatible as Willybomb mentioned.  The TL072 might be a touch noisier, but is better at driving capacitive loads than the 4558 and has a higher input impedance, so a substitution should be fine. 

Scruffie

Quote from: mauman on September 23, 2020, 03:08:33 AM
Voltages at the pins should be the same and they're pin-compatible as Willybomb mentioned.  The TL072 might be a touch noisier, but is better at driving capacitive loads than the 4558 and has a higher input impedance, so a substitution should be fine.
TL072 noisier than a 4558?  ???
Works at Lectric-FX

mauman

Quote from: Scruffie on September 23, 2020, 10:02:13 AM
Quote from: mauman on September 23, 2020, 03:08:33 AM
Voltages at the pins should be the same and they're pin-compatible as Willybomb mentioned.  The TL072 might be a touch noisier, but is better at driving capacitive loads than the 4558 and has a higher input impedance, so a substitution should be fine.
TL072 noisier than a 4558?  ???
For voltage noise, yep, according to the TI datasheets for both, typical voltage noise @ 1 kHz for a TL072 is 18 nV/square root Hz, while RC4558 is about half of that at 8 nV/square root Hz. 

Scruffie

Quote from: mauman on September 23, 2020, 07:58:23 PM
Quote from: Scruffie on September 23, 2020, 10:02:13 AM
Quote from: mauman on September 23, 2020, 03:08:33 AM
Voltages at the pins should be the same and they're pin-compatible as Willybomb mentioned.  The TL072 might be a touch noisier, but is better at driving capacitive loads than the 4558 and has a higher input impedance, so a substitution should be fine.
TL072 noisier than a 4558?  ???
For voltage noise, yep, according to the TI datasheets for both, typical voltage noise @ 1 kHz for a TL072 is 18 nV/square root Hz, while RC4558 is about half of that at 8 nV/square root Hz.
Ah, for that one spec under test conditions perhaps, but its input current noise is much smaller and in practice, I think most would agree the TL072 is the quieter chip for pedal use.
Works at Lectric-FX

TFZ

#6
That's not just "test conditions". Overall noise is the result of voltage and current noise. Depending on the impedances involved, one or the other will dominate. Hence there will be circuits where the 4558 can be quieter of course.

The fact of the matter is, many guitar pedals are terribly designed in that regard. High impedance environments with old bipolar opamps put in haphazardly. Many EHX circuits come to mind. And the result of that is people jumping to simple conclusions without any understanding and getting wrong impressions about parts, but the problem is they are only used improperly.

Suffice it to say, the 4558 isn't a good part by today's standards, far from it. For the same price you can get a NE5532, which beats it in every way. But we're building guitar pedals here, so at least for distortion the ear might have to decide what you put in. Still, many circuits can be adapted to lower impedances in order to reduce noise, without changing the tone.

Scruffie

Quote from: TFZ on September 24, 2020, 11:45:08 AM
That's not just "test conditions". Overall noise is the result of voltage and current noise. Depending on the impedances involved, one or the other will dominate. Hence there will be circuits where the 4558 can be quieter of course.
I refer to the 1Hz bandwidth at 1kHz that rating is taken at which isn't very realistic in practice.

The NE5532 has its own quirks, very low input impedance and not to mention massive current draw. There's no one size fits all if you get to the nitty gritty which is why there's so many op amps to chose from. But rule of thumb, TL072 quiet, 4558 less so but usually fine, NE5532 very quiet but not suitable for all applications.
Works at Lectric-FX

TFZ

> I refer to the 1Hz bandwidth at 1kHz that rating is taken at which isn't very realistic in practice.
I still don't understand. That number is a way to compare different opamps, and gives you the information you need to calculate the actual noise in your circuit.

> The NE5532 has its own quirks, very low input impedance
You've stated that before, and when I asked you to explain how you came to that conclusion you didn't answer unfortunately. Please do  ;).

> not to mention massive current draw
For the dual the supply current is 16mA, I certainly wouldn't call that massive. A PT2399 is about the same, a Spin FV-1 takes 55mA. Some people drive their bypass LEDs harder than that ;D, a bypass relay takes way more current as well. I really don't think there is any problem with that.

> There's no one size fits all if you get to the nitty gritty which is why there's so many op amps to chose from
Well yes of course. Engineering is the art of finding a workable compromise from conflicting requirements, same goes for choosing an opamp. But it's usually not too hard to find a solution that doesn't suck.

Scruffie

It's an extremely specific parameter, not a real world one, hence, test condition. A handy quick reference, but there's more comprehensive noise specs elsewhere.

Yeah a PT2399 might be the same, but that's an entire digital delay circuit including a dual op amp on a chip, not an audio op amp... If your design used 4 duals you'd be at 60mA, I'd call that high for an guitar pedal audio path... if you had a charge pump in your circuit you could easily run in to issues, in a high current circuit that might push you over the 100mA of a lot of standard 9V taps.

The 5532 has about 2-300k impedance at its non-inverting pin IIRC, read the datasheet.
Works at Lectric-FX

TFZ

Ok then let's agree to disagree on the current draw. Regarding the input impedance: erm... no that's wrong. The value in the datasheet is the input impedance running the opamp open loop, which means no feedback applied. So that is not the input impedance of any practical circuit. Once the loop is closed, which of course is the case in pretty much all applications, the input impedance is a lot higher than that. It is not possible to blankly state the input impedance in the datasheet, because it depends on the circuit the opamp is put in.

It is roughly the open loop gain times the applied gain times the open loop input impedance. So for the NE5532, that would be 100,000 x 300,000kOhm x gain. As you can see, its input impedance is far from "very low". Other opamps will produce similar numbers, even ones that are older and worse. That's why we can always safely assume that the input impedance of a non-inverting amplifier stage is the same as the bias resistor (in parallel to maybe other stuff).

Scruffie

I'll take you at your word on it, I'm not fussed enough to go digging as given the option I'd pick the FET input op amp for any high impedance jobs if I wanted low noise. But even if it's not the input impedance, I distinctly remember under certain conditions an NE5532 isn't suitable... high bias current?
Works at Lectric-FX

alanp

Quote from: TFZ on September 24, 2020, 11:45:08 AM
The fact of the matter is, many guitar pedals are terribly designed in that regard. High impedance environments with old bipolar opamps put in haphazardly. Many EHX circuits come to mind. And the result of that is people jumping to simple conclusions without any understanding and getting wrong impressions about parts, but the problem is they are only used improperly.

This reminds me of a documentary I saw about Vox Amplifiers. Back in the early days (the AC30 was an interesting story in itself), one of the guys runs in with a little box (the first fuzz) and shows it off... the other guy's reaction is, we're trying to make our amps sound as good as possible, and you're deliberately adding this in?!
"A man is not dead while his name is still spoken."
- Terry Pratchett
My OSHpark shared projects
My website