What if any practices are done to make the 4558 perform better?
Using it in inverting mode would be obvious, but what other things are there to do? is bootstrapping it worth the trouble?
I admit I haven't thought of them sounding "bad", but I hadn't thought of 2$ headphones sounding bad either until I got better ones. So I'd be more than interested in hearing what there is to say about improving the 4558.
I noticed the other day that the Revox B77 uses a 4558 in non-inverting mode in the playback amp. Maybe that's what people refer to as the "revox sound?" ;D
Using it in inverting mode would be obvious, but what other things are there to do? is bootstrapping it worth the trouble?
I admit I haven't thought of them sounding "bad", but I hadn't thought of 2$ headphones sounding bad either until I got better ones. So I'd be more than interested in hearing what there is to say about improving the 4558.
I noticed the other day that the Revox B77 uses a 4558 in non-inverting mode in the playback amp. Maybe that's what people refer to as the "revox sound?" ;D
Do you mean this type of 4558:
https://www.ti.com/lit/ds/symlink/rc4558.pdf
https://www.rcscomponents.kiev.ua/datasheets/jrc45584i743ncft874nfdt34ufguygf43.pdf
If so, depends what circuit its in.
https://www.ti.com/lit/ds/symlink/rc4558.pdf
https://www.rcscomponents.kiev.ua/datasheets/jrc45584i743ncft874nfdt34ufguygf43.pdf
If so, depends what circuit its in.
The old Fairchild 4558s sound better than any other variant produced. That being said, its mediocre compared to the 5532 in just about every way. I would use signetics 5532s in place of 4558s. You'll need some decoupling and possibly some 22pf HF feedback caps.
The RC4558 has lower power consumption, slower, has no back-to-back diode protection on the inputs, is somewhat noisier, so unless you want extended battery life the NE5532 would get my vote too, or one of the new higher performance audio opamps.
Supply decoupling is an issue to be wary of, faster opamps often need high-speed decoupling between the rails (i.e. ceramic caps close to the chip) or they under-perform (oscillate internally).
Its a general rule of thumb that a good fast opamp for audio will have about 4mA current requirement per amp, ie 8mA for a dual device.
Supply decoupling is an issue to be wary of, faster opamps often need high-speed decoupling between the rails (i.e. ceramic caps close to the chip) or they under-perform (oscillate internally).
Its a general rule of thumb that a good fast opamp for audio will have about 4mA current requirement per amp, ie 8mA for a dual device.
4558 is perfectly adequate for gain 1 to 5 designs. Thousands of them were used in graphic equalizers with no bad effects.
What they will not do is high gain circuits. I had a product that used them at 50x gain for a RIAA LP circuit. Very hissy. I changed them out for ST33078. Hiss gone. As noted above, I had to add 33 pf around the feedback resistor, and a .1 uf cap between +- rails (one for two IC's) to kill ~1 mhz oscillation. Faster parts can oscillate. These additional parts would have run the cost of the product up $1, a terrible loss of sales to the OEM.
What they will not do is high gain circuits. I had a product that used them at 50x gain for a RIAA LP circuit. Very hissy. I changed them out for ST33078. Hiss gone. As noted above, I had to add 33 pf around the feedback resistor, and a .1 uf cap between +- rails (one for two IC's) to kill ~1 mhz oscillation. Faster parts can oscillate. These additional parts would have run the cost of the product up $1, a terrible loss of sales to the OEM.
Agreed. Anyway it is a nice allrounder for an affordable price. You find these inside Motu4 for instance.
There are two ways to approach good performance:
First, inherent linearity, including, but not limited to, wide class A operation, means high idle current.
Second, a crapton of GNFB wrapped around whatever nonlinear circuitry.
High idle current may not be the first choice for battery powered equipment however.
"The RC4558 device is a dual general-purpose operational amplifier, with each half electrically similar to the. μA741"
Not very impressive .....
Not very impressive .....
Even making them run in class A does not work well as their distortion increases with loads below
10KΩ. The last time I used one was designing a voltage regulator for my 1974 Saab. Much improved
over the two stage relay regulator.
10KΩ. The last time I used one was designing a voltage regulator for my 1974 Saab. Much improved
over the two stage relay regulator.
Hi edgarsls,
Op amps have come a long way since the 4558. You can replace it with just about any newer bipolar op amp, or J-Fet input types and get better performance from a noise and distortion perspective. Given a half decent choice, you have no where to go but up. An NE5532 is an excellent choice. CMOS, naw. Many better choices exist.
Your load current will probably swamp the current used by the op amp. Not a major concern really. Don't worry about selecting "the best op amp" in this case. Super low noise, and super low distortion really doesn't matter driving headphones. Output drive capability will if the op amp drives the headphones directly.
-Chris
Op amps have come a long way since the 4558. You can replace it with just about any newer bipolar op amp, or J-Fet input types and get better performance from a noise and distortion perspective. Given a half decent choice, you have no where to go but up. An NE5532 is an excellent choice. CMOS, naw. Many better choices exist.
Your load current will probably swamp the current used by the op amp. Not a major concern really. Don't worry about selecting "the best op amp" in this case. Super low noise, and super low distortion really doesn't matter driving headphones. Output drive capability will if the op amp drives the headphones directly.
-Chris
I'd go with an LME49720 (NJM4562) if the requirements were higher instead of an NE5532, but in this case the NE5532 is just about perfect. I use and like the OPA134 family when the application calls for it.
Which is not defined. 5532 instead of 4558 is not a sure thing. In the circuit I upgraded, increasing Vcc current by 60% would have decreased the ouput voltage by 60% or more. Herald used series 330 ohm resistors in the Vcc to cut the hum. Cheapo designer trick. I had to totally redo the power supply in the Herald-RA88a to use low hiss 33078 and also not have gross hum.
Hi edgarsls,
Without the circuit schematic in front of us, how do you expect an intelligent comment specific to your case?
I will say this. The 4558 is internally compensated and you have no access to internal functions as you might in the single versions of some op amps. So therefore you do not have any options to improve anything. So it then follows the only way to improve it is to fork lift the part, that means replace it with better. Now, there one thing that may help. Loading an op amp always decreases performance to some degree. So if that 4558 is driving your headphones and you are fixed on keeping the 4558, installing a buffer between it and your headphones would improve its performance.
I hope this answers your question. Your options are:
Without the circuit schematic in front of us, how do you expect an intelligent comment specific to your case?
I will say this. The 4558 is internally compensated and you have no access to internal functions as you might in the single versions of some op amps. So therefore you do not have any options to improve anything. So it then follows the only way to improve it is to fork lift the part, that means replace it with better. Now, there one thing that may help. Loading an op amp always decreases performance to some degree. So if that 4558 is driving your headphones and you are fixed on keeping the 4558, installing a buffer between it and your headphones would improve its performance.
I hope this answers your question. Your options are:
- Live with it (easiest)
- Replace it (best for performance)
- Install a buffer (poor option)
If you're saying that there's nothing that can be done to improve the performance of the 4558, then why are you asking for circuit details?
I'm asking the question so I can design the circuit in the first place, to whatever is best to improve the peformance of the op amp.
And yes I am aware that there is op amp 'x' which has much better performance already.
Wouldn't bootstrapping the op amp increase the linearity of the output stage. but decrease the linearity of the input stage because of increases common mode voltage swing because of the shifting power rails? Maybe if you bootstrap the op amp to the non-inverting input, it would have the least distortion because the common mode voltage would be eliminated, as well as the output would be bootstrapped.
I'm asking the question so I can design the circuit in the first place, to whatever is best to improve the peformance of the op amp.
And yes I am aware that there is op amp 'x' which has much better performance already.
Wouldn't bootstrapping the op amp increase the linearity of the output stage. but decrease the linearity of the input stage because of increases common mode voltage swing because of the shifting power rails? Maybe if you bootstrap the op amp to the non-inverting input, it would have the least distortion because the common mode voltage would be eliminated, as well as the output would be bootstrapped.
Bipolars do not suffer from the same common mode issues as J-Fet types ... but then you obviously know this as you are designing.
Okay, you've decided to go with an op amp that performs sub-par. That's your decision. Hands off from me. I'm not about to debate anything.
Okay, you've decided to go with an op amp that performs sub-par. That's your decision. Hands off from me. I'm not about to debate anything.