Op-amps themselves are very complex circuits, with several stages, often designed for very high open loop gain, low power consumption, etc. If you look at the typical pinout for 8-pin dual op-amps, they have two power supply pins, and for each "amplifier" 3 more pins (an inverting input, a non-inverting input, and an output pin), but no ground. Parasitic capacitances and inductances between the main circuit board and the discrete "opamp" may effect circuit operation / stability. The inductance of one-half-inch of wire or circuit board trace can form a resonant LC circuit with the power supply bypass capacitors at megahertz frequencies, for instance, injecting high frequency noise onto the power supply of the discrete op amp module.
Properly selected for the application, and with careful circuit board layout, commercial high performance op-amps (better spec-wise than NE5532, and a lot more expensive) like the ones used by Mike and Jason as buffers or I-V stages on the DAC chip outputs can sound pretty nice.
Is that mostly a matching issue, i.e. opamp A in a given amp doesn't sound great, just like compatible tubes might not sound great in a given amp, or is there more to it? Are the megahertz resonances the "oscillations" I have read about? Is there a concern that your amp is becoming a radio transmitter and the FCC might knock on your door? Is there a concern about damaging the amplifier or the opamp?
It has been suggested elsewhere that nobody without access to an oscilloscope should swap opamps. Is it conceivable that an amp designed for swapping opamps could come with a self check and simply shuts down with a warning LED lighting up if there's an issue? I'm assuming that wouldn't add all that much to the overall cost.
The only difference I see is that due to the limited life of tubes swapping them is a necessity, so the amplifier will have a reasonable life before it needs to be fixed by a technician. Being able to try different brands and models of tubes is a welcomed side effect. For op-amps there is no need to, but if possible is also a welcomed feature.
Definitely. I'm thinking normally you would not want to build an amp with big holes in the top, through which vital components could be removed by the user at any point, even while running. Components that potentially run quite hot indeed. But since it's "necessary" (there are amps and DACs that just have tubes inside, without making it easy to swap them, so evidently it's not strictly necessary), you do it anyway. Well, if that is okay with tubes, then why not maybe with opamps? Though Ableza raises a good point below.
My point is that considering the little I know about schiit, rolling op-amps is not something Jason is willing to do
I'm okay with them not doing it, but I definitely would like to understand better why the risk assessment is different compared to allowing it for tubes. Tube amps with user accessible tubes are not entirely un-mad.
specially condiering someone would need to open the amplifier, and he is really against that for safety reasons.
That's where Schiit would have a chance to distinguish themselves from all products with swappable opamps that I have seen: make the opamps accessible without needing to open the amp, just like the tubes are (or the modules in the Bifrost 2). I.e. provide a removable card with a DIP8 socket or two, or an opening just like for tubes, though possibly with a raised socket because opamps tend to be rather shallow, and make either approach sufficiently big to have flexibility about the various shapes and sizes opamps come in:
Even the Burson Fun, from a company that makes several types of discrete opamps, has to be opened by removing four screws in order to swap opamps. Makes the design cheaper, of course. And because of the point Ableza raises below, there would be a bigger incentive than for tubes to at least make it possible to cover the hole again after installing an opamp, at least for the designs without a case around the opamp's PCB.
I know you were not asking me but it seems the obvious difference is tubes are generally external, protected from lethal voltages, and designed to be changed by a user. Opamps are not.
That they are usually external is probably owed more to their size, since otherwise the chassis has to be significantly bigger (at least if you want to support a variety of compatible tubes), and you'd have to dissipate the added heat somehow. Tubes generally need high voltages, but opamps operate far below 50 V from what I have seen. So it would still be iffy if you could touch an opamp while the amp is running, but less so than with the triple-digit voltages typically used in tube amps.
Accessible hot tubes are a possible danger to you, your kids and your pets (hence the guards on many tube amps), though possibly a less severe one than touching an open circuit - that is a good point. Most opamps do not come wrapped in plastic cases like the Burson opamps, fragile as they can be (three of the eight I had access to came apart without trying, two with a newer design with just one plastic part):
The case makes it difficult to touch the PCB, but the pins are still exposed - but that's the same as with tubes. Is it generally advisable to take ESD precautions when touching the pins of tubes? It's a concern for opamps, but I'm wondering whether an opamp with a case like the Bursons is still more vulnerable to ESD than a tube.
Further, having built the Whammy (likely the DIY Amp referenced earlier); while it uses 8 pin DIP OPAMP sockets there are a limited number of opamps in the 8 pin PDIP form factor that can be used just for the very simple reason that they vary widely in which pins serve which I/O function. This is before you even begin to consider power tolerances and other more detailed specs.
Think about what happens if you stick the wrong octal tube in to an octal socket in your Freya, not pointing any fingers here.
Or mis-orientate (Army-ism) your loctal tube into your dual mono triode adapter in an unnamed hybrid amp. Pointing finger at self on this one, but hey I thought the whole point of the loctal was to prevent mis-insertion. But in my case no harm was done due to quick response and robust amp design.
Exactly my point, we already accept that kind of risk with tube amps, so the precedent has been set.
Thanks for the input, everyone!