Really? Try telling that to the thousands of pedal steel guitarists who have changed out the RC4558s in their Peavey amps with OPA2134s (Ken Fox mod) and noticed a HUGE difference. And the OPA1642 is even a step better than the 2134.
Assuming your anecdotal evidence is true, are you so sure the reason is because of the noise performance of the 2134, and not because it would raise the input impedance significantly to swap the bipolar input 4558 for the FET input 2134? Have you even forgotten what this thread is asking?
How about the thousands seeking out and paying top dollar for vintage 4558's for their Tube Screamers? Didn't they get the memo about noise performance?
Actually, I'm not all that interested in top-shelf. I'd gladly take a 5532 if it were FET input. I just want something better than a TLO72, with enough output power to not stop working if I plug a cable into it. And I am not crazy about the sound of the 2134. If top-shelf is the alternative, then it's the alternative.
Really, this is getting as pissy as diyaudio.
Oh yeah, it is diyaudio.
.
Really, this is getting as pissy as diyaudio.
Oh yeah, it is diyaudio.
.
yes 1-10 kOhm DCR isn't the (only) issue, mag pickup's inductance can give Z to 100s of kOhm that resonates with pF cable and input device C causing audible frequency response variation
Pickup Inductance | Harmony Central
https://courses.physics.illinois.ed...Undergrad_Research_Symposium_Talk_1-29-10.pdf
Pickup Inductance | Harmony Central
https://courses.physics.illinois.ed...Undergrad_Research_Symposium_Talk_1-29-10.pdf
The impedance of regular magnetic pickups in the critical audio range of resonance, typically between 2~5kHz, is indeed several 100KOhm, often limited by the potentiometers.
From Wikipedia:
"The Tube Screamer has a drive knob, a tone knob, and a level knob. The drive knob adjusts gain, the tone knob adjusts treble and the level knob adjusts the output volume of the pedal. The pedal is used to try to mimic the sound of a vintage tube amplifier. The classic Tube Screamer sound includes a "mid-hump", which means that the circuit accentuates frequencies between the bass and treble ranges (mid-frequencies). Some guitarists prefer this sort of equalization, as it helps to keep their sound from getting lost in the overall mix of the band. Much has been made of the operational amplifier chips used in the various versions of the Tube Screamer pedal, and several "fairy tales" about the merits of these devices have been written on the subject. The JRC4558D chip is well-regarded by some. The RC4558 is a low priced, general purpose dual operational amplifier, introduced mid 70's by Texas Instruments as an "improved" version of the early 741, and used in thousands of consumer and industrial designs. In fact, JRC4558D is nothing else than the licensed product manufactured by Japan Radio Company, and identical to any other 4558 chip. In reality, the type of op-amp has little to do with the sound of the pedal, which is dominated by the diodes in the op-amp's feedback path."
Last edited:
A little more about the TLO72: I got samples from TI when the TLO "Bi-FET" product line was introduced. I was horsing around with them, and noticed that they sounded kind of rough but distinctly better when I bandlimited them with a capacitor across the feedback resistor. I had inadvertently discovered something that was explained some time later in an Audio Amateur Magazine article.
An externally hosted image should be here but it was not working when we last tested it.
So now those guitarists who prefer vintage 4558 in Tube Screamers are psycho acoustically deluded but those who like the 2134 swap for amps are "correct"? And it's because of the improvement in noise performance buried deep under the noise floor of their rig, and not the big tonal change from increasing input impedance greatly?
C'mon leadbelly, there's plenty of room to just answer the question the guy is asking, which seems to boil down to "is there a higher performance option than my 2134 for this application". Especially since we're talking musical instruments, maybe he just wants it that way? Or it makes no difference in the end? We're not talking something sacred.
OP: if the 164x isn't doing it, then you're starting to look at more complex solutions than a simple opamp swap, e.g. grafting a jfet input on to a modern bjt opamp (or heck a 5534, there are circuits around to bypass its input transistors) or some sort of buffer in loop with the 164x to get the drive needed for your cables.
OP: if the 164x isn't doing it, then you're starting to look at more complex solutions than a simple opamp swap, e.g. grafting a jfet input on to a modern bjt opamp (or heck a 5534, there are circuits around to bypass its input transistors) or some sort of buffer in loop with the 164x to get the drive needed for your cables.
Haha, yeah, I guess I wasn't getting that level of onerous (that seems so pervasive her and everywhere else on the net). I generally just try to pick fights with the consistent loudmouths. 🙂
I mean if the TL072 is working, just grab one of the fast unity gain buffers that's around and fold it into the FB loop. Bam, cable problems solved.
I mean if the TL072 is working, just grab one of the fast unity gain buffers that's around and fold it into the FB loop. Bam, cable problems solved.
Actually I was thinking about that today. A 5534 would probably give me the mellower sound that I want, and I remember seeing circuits for adding FET inputs to standard opamps....unless of course it's the FETs that are the culprit.
Or just put in a 5534 and outboard a stinkin' FET on a perfboard...
.
> 390kOhms ...a good old 5532 op amp here, would that be too high a resistor for its noninverting input?
Read data. '5532 can have as much as 1,000nA (1uA) input current. In 390K this is 0.4V (not mV!) of DC offset. Typical is 300nA which is near 0.1V of DC offset.
Does that matter to your circuit?
If you run single 30V supply, with a V/2 DC reference, centering the action at 14.9V or 14.6V is same-as 15.0V for all practical purpose.
If you run +/-15V and hope to avoid an output cap, the 0.1V-0.4V of output DC may offset the next stage in the chain, if it expects dead-Zero DC on its input.
At single 9V supply, 0.4V offset is non-negligible.
If your stage runs a gain of 100, flat to DC, that 0.1V-0.4V becomes 10V to 40V, which sure will cut your headroom. (So force DC gain to unity!)
There are other tricks. A complementing 390K in the other input nulls the gross input current offset error. The offset current error is 10nA typ but over 100nA on a bad day in the factory or at un-playable temperature. Now error is 4mV to 40mV, so not-bad unless you do other wrong things. This compensating resistor is a large hiss source so must be bypassed. Which puts the parts-count as high as other less-clever approaches.
You can bootstrap the 390K. The inverting input is essentially unity-gain to the NI input. If the NFB parts are low impedance, an added R and C can push the effective audio impedance much higher. 39K will be much more than 390K. >3Meg over most of the audio band. So put a 410K in front to set the actual impedance you feel you need to simulate.
> fail to drive a cable
The '072 is dandy for many things but not a brute. First thing is, going off-board (out of box) you "must" put >300 Ohms series (outside NFB) so cable capacitance does not suck-off the '072's supersonic gain (which can cripple its audio performance). You also do not drive hundreds-feet of cable with '072 because it does go sour below ~~1K loading. This is true more/less of any amplifier. The '5532 is the go-to for driving long audio lines because it does no suck down below 600 Ohms. You still need a series resistor but 100r leaves enough gain for the amp to stay clean. Next step is Jensen 990 (not strained at 150r load, uses 10r isolator) or similar built with about any good opamp and a couple transistors.
'072 is not best-of-breed in non-inverting work because the CMRR is good not great and there's distortion mechanisms. That said, ANY music which goes from studio through mix-down to mastering to consumer will pass through some or many (even 100!) '072s along the way through most pro gear. It really is the go-to when heavy loads are not involved.
Read data. '5532 can have as much as 1,000nA (1uA) input current. In 390K this is 0.4V (not mV!) of DC offset. Typical is 300nA which is near 0.1V of DC offset.
Does that matter to your circuit?
If you run single 30V supply, with a V/2 DC reference, centering the action at 14.9V or 14.6V is same-as 15.0V for all practical purpose.
If you run +/-15V and hope to avoid an output cap, the 0.1V-0.4V of output DC may offset the next stage in the chain, if it expects dead-Zero DC on its input.
At single 9V supply, 0.4V offset is non-negligible.
If your stage runs a gain of 100, flat to DC, that 0.1V-0.4V becomes 10V to 40V, which sure will cut your headroom. (So force DC gain to unity!)
There are other tricks. A complementing 390K in the other input nulls the gross input current offset error. The offset current error is 10nA typ but over 100nA on a bad day in the factory or at un-playable temperature. Now error is 4mV to 40mV, so not-bad unless you do other wrong things. This compensating resistor is a large hiss source so must be bypassed. Which puts the parts-count as high as other less-clever approaches.
You can bootstrap the 390K. The inverting input is essentially unity-gain to the NI input. If the NFB parts are low impedance, an added R and C can push the effective audio impedance much higher. 39K will be much more than 390K. >3Meg over most of the audio band. So put a 410K in front to set the actual impedance you feel you need to simulate.
> fail to drive a cable
The '072 is dandy for many things but not a brute. First thing is, going off-board (out of box) you "must" put >300 Ohms series (outside NFB) so cable capacitance does not suck-off the '072's supersonic gain (which can cripple its audio performance). You also do not drive hundreds-feet of cable with '072 because it does go sour below ~~1K loading. This is true more/less of any amplifier. The '5532 is the go-to for driving long audio lines because it does no suck down below 600 Ohms. You still need a series resistor but 100r leaves enough gain for the amp to stay clean. Next step is Jensen 990 (not strained at 150r load, uses 10r isolator) or similar built with about any good opamp and a couple transistors.
'072 is not best-of-breed in non-inverting work because the CMRR is good not great and there's distortion mechanisms. That said, ANY music which goes from studio through mix-down to mastering to consumer will pass through some or many (even 100!) '072s along the way through most pro gear. It really is the go-to when heavy loads are not involved.
not wrong
I have experienced this in a pro ISM application - TL07x does have limited C drive ability - the pull down side of the TL07x output gets really poor as it approaches the negative rail
driving a flat cable with alternating gnd and analog signal lines the TL07x buffering our prototype's analog outputs did oscillate on 6' cables into a DAQ board despite 50 Ohm isolating series R - only on negative output
playing with composite op amp circuits, I've used the TL027 as a cheap test chip that I could afford to blow up in early stage debugging - then it performed very well as a input op amp - as long as it wasn't driving any load
I have experienced this in a pro ISM application - TL07x does have limited C drive ability - the pull down side of the TL07x output gets really poor as it approaches the negative rail
driving a flat cable with alternating gnd and analog signal lines the TL07x buffering our prototype's analog outputs did oscillate on 6' cables into a DAQ board despite 50 Ohm isolating series R - only on negative output
playing with composite op amp circuits, I've used the TL027 as a cheap test chip that I could afford to blow up in early stage debugging - then it performed very well as a input op amp - as long as it wasn't driving any load
NE5534 with a new front end ...anybody?
But honestly, I'd graft a jfet front end with appropriate compensation onto a fast BJT (heck an lme49720) rather than hijack half a ne5534. It was a nice trick years ago, but been long superseded.
Quote:
That is way too kind. He has also said the 2134 is too bright and the TL072 can't drive cables. What's the point of a technical discussion in cess pool of egomania and ignorance? Where do we go next? "My wife hates LF412"? "LM4562 killed JFK"? "797's are made of adamantium"?
__________________
The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts. Bertrand Russell
That is way too kind. He has also said the 2134 is too bright and the TL072 can't drive cables. What's the point of a technical discussion in cess pool of egomania and ignorance? Where do we go next? "My wife hates LF412"? "LM4562 killed JFK"? "797's are made of adamantium"?
__________________
The whole problem with the world is that fools and fanatics are always so certain of themselves, and wiser people so full of doubts. Bertrand Russell
- Status
- Not open for further replies.