What's so hard about it though,
nothing that hard about it, past knowing how to verify what you have done has actually improved matters, I agree LP did present it a little too extremely, but thats what he does for effect =) however there is most certainly truth in what hes saying. there is no general guidelines of which chip will go well where, replacing X for Y, Each chip has its own set of preferred operating conditions and each circuit represents a set of parameters of its own, even when the parts are the same, the lead, or trace inductance, grounding, gain, compensation, power supply voltage and load current, capacitance, as well as parasitic and/or parallel capacitance, inductance etc; all have influence on these parameters.
bipolar or fet is a minor thing, that will tell you little and cannot be used to make decisions. the datasheet and knowing how to read a layout, as well as having some way to verify performance is the only way to do it. simulation can help to a degree too, but even when spending a great deal of time creating complex models there are still some elements that will be unknown until its done in reality and thats if you can get an accurate model, or be able to trust the datasheet enough to build one based on that. the best models come from building and measuring the real thing, then you have the change from loading....
a basic scope can be had for pretty cheap and unless you are really needing to play with high bandwidth parts, should be enough to get a rough idea. handheld ones are starting to become more useful, but you still cannot beat an old analogue scope for value.
of course as mentioned, I didnt even touch dealing with the incompetence of the designer, or otherwise compromised designs
that being said, there are ways to make sure what you are doing is safe, or likely to cause damage, many of which have been discussed in the last few pages; you can still have fun, but know when to recognize the signs something is really wrong. you WILL hear differences, because of the change in conditions, some may even be preferred, whether you have made an improvement or not is all relative, if you like the sound better and the amp isnt going to kill itself or your headphones, then hey I call that a win.
do remember though and this is important, we are talking about audio bandwidth here, the blinkered chase for high slew rate and wide bandwidth at all costs is not all that relevant for our purposes; in the same way that oven controlled oscillators are not relevant either, only the OCX do tend to be a higher grade part overall, nothing to do with being ovenised
@WiR3D: which build were you talking about? the LPUHP amps? the GB is over, there was a single run of blank PCBs, no kits were available, the PCBs sold out in less than 24hrs!! I ran the group buy for the passives, custom transformers from sumR and keratherm thermal interface material to drive the cost down, people sorted the actives out themselves. there will not be another run of PCBs
Edited by qusp - 6/12/12 at 6:39am