Plenty, one three star chef watched as an oil-man's wife put ice into Romanee-Conti because she was hot. (hot not HOT)
I agree this pairing gives me pause, Jerry was a classmate and one of the most brilliant people you could ever meet, but he was totally invested in the audio needs 100k response thing. I took my daughter to inspect Cal Tech and was very dissapointed to find that Jim had lost his position there.
It was very nerd oriented, kids too smart for their age in the math/science direction. I can't fault the quality/intensity of the scholarship. The upside is the culture was pretty clear on a casual visit so nothing ventured nothing gained on checking it out for yourself.
Last edited:
My experience with folks who got their undergrad degrees at Caltech is that they were well-schooled indeed, but also steeped in a not-invented-here culture and an attitude of extreme skepticism. Of course this may well have changed in more recent times.
I don't know how much the current atmosphere preserves the playfulness depicted in one of my favorite movies, Real Genius (1985 I think). But that fun-loving atmosphere I believe was one of the distinguishing differences between Caltech and the other preeminent tech-oriented school, MIT.
I don't know how much the current atmosphere preserves the playfulness depicted in one of my favorite movies, Real Genius (1985 I think). But that fun-loving atmosphere I believe was one of the distinguishing differences between Caltech and the other preeminent tech-oriented school, MIT.
As usual Pavel is right.
Here is a set of measurements on the noise from a power supply.
The lowest level trace is with the power supply unplugged and the probe shorted to the output ground of the power supply.
All the rest are virtually the same. They show plugging in the power supply even to a shorted probe introduces noise. The noise doesn't change even when looking at the DC rail output. All the noise is what was passed from the AC line.
Any questions ask Dick he's an expert on AC line noise.
ES
P.S. the dB reference level is 1 volt.
Attachments
Nice how you didn't even pause to contradict yourself.
So, let's go back to Pavel's comment (which echoes the same comment I've made a few times in the past couple days) and look at this as a system instead of one piece in isolation. The noise of that supply is pretty low (how you determined that it's from the AC line and not from diode switching, inductive kickback, or whatever is left unsaid)- the worst component is 60Hz at -92dBV, or about 25uV. We hook up the supply do a decently designed amp, which is likely to have better than 90dB of rejection at 60Hz. Now the noise at the output due to this supply is at -182dBV, or 0.8nV. Remember, this is at the output, so input-referred noise from that source is that number divided by gain. I think we're ridiculously below where anyone would care about it.
One may then argue that the PSR of the amp is likely to be lower at higher frequencies. That's a valid point, BUT... if we take a common amp like an AD797, the PSR is still 80dB at 20kHz and the power supply noise is -140dBV. So the noise is even lower than before, down around the thermal noise of the Universe.
You've demonstrated that this is not something to worry overly much about as a system question. The weak links lie elsewhere.
You could hook this power supply up to an amplifier and voila there would be nothing to come out!
This is a power supply for a wideband low noise preamp. The AC to DC conversion is almost as quiet as possible. There is still some regulator noise creeping in!
The next issue on this power supply is to get rid of the common mode AC line noise.
Power supply rejection ratio decreases as you increase the gain of an amplifier and as its bandwidth decreases.
You will be hard pressed to get your hand waving 90 dB if you just put two empty connectors in a metal case.
You can keep making up all the numbers you want. I'll stick to actual measurements.
This is a power supply for a wideband low noise preamp. The AC to DC conversion is almost as quiet as possible. There is still some regulator noise creeping in!
The next issue on this power supply is to get rid of the common mode AC line noise.
Power supply rejection ratio decreases as you increase the gain of an amplifier and as its bandwidth decreases.
You will be hard pressed to get your hand waving 90 dB if you just put two empty connectors in a metal case.
You can keep making up all the numbers you want. I'll stick to actual measurements.
Last edited:
I have a strange problem. I nearly always can tell whether I'm listening to 'real' sound, or Just A Hifi. Must be my super duper ears, I guess ...
For some bizarre reason I prefer the former variety, so I work towards having a system perform at that level - and find it satisfying when it does so. This unusual abberation of mine is fortunately fairly rare, so it probably won't contaminate the normal world of audio, if everyone crosses their fingers, and their hearts ...
For some bizarre reason I prefer the former variety, so I work towards having a system perform at that level - and find it satisfying when it does so. This unusual abberation of mine is fortunately fairly rare, so it probably won't contaminate the normal world of audio, if everyone crosses their fingers, and their hearts ...
Thus my use of the 797 as the example amp. The performance at the output is quite predictable, and I'm sure your measurements of the complete system (that's what you agreed was important!) will show that this supply noise is far from the weak link as far as system noise performance goes. You can make a chain out of titanium links and worry about alloying and annealing, but if one of the links is aluminum, you're looking in the wrong place.
Fas42, I LOVE IT when for just a moment, I think that I am listening to a real human voice or acoustical instrument, through a hi fi system. It is RARE, even with my own set up, but it has happened. That is our goal!
For example, with the STAX electrostatic phones, I once had a flashback of working on stage with the GD, based on their live recording and a proper playback set-up.
I often 'fall in love' with human female singers, who, when well recorded, can get their 'emotions' across on hi fi playback. Without these events, I might as well stick to TV. '-)
For example, with the STAX electrostatic phones, I once had a flashback of working on stage with the GD, based on their live recording and a proper playback set-up.
I often 'fall in love' with human female singers, who, when well recorded, can get their 'emotions' across on hi fi playback. Without these events, I might as well stick to TV. '-)
Having built preamps it is the AC line noise that is one of the limits. A typical power transformer, diode bridge and filter capacitor into a single IC regulator produces such poor results it is laughable.
Using the AD797 it has around 120 dB of PSSR in the AD test circuit with a gain of one. The open loop gain drops by 60 dB by 10,000 Hz. So with a preamp gain of 60 dB you got nothing.
Now for your chain just make the aluminum link bigger.
Using the AD797 it has around 120 dB of PSSR in the AD test circuit with a gain of one. The open loop gain drops by 60 dB by 10,000 Hz. So with a preamp gain of 60 dB you got nothing.
Now for your chain just make the aluminum link bigger.
Its not always the case that the open loop PSRR plots in the data sheets can be translated in this manner to the closed loop case - it depends on where in the internals of the IC the supply noise creeps in. You've assumed the best-case condition here where its the input stage which is most sensitive and the other ingress points contribute nothing at all. However this might not always be the case.
I appreciate the sweet, tender advances ...
Yes, the ones I keep mentioning: take a realistic mains supply, pass it through the conventional power supply with the average, everyday level of parasitic behaviours, feed it into a quite nicely designed power amp, and drive that amp hard into a less than ideal load. Of course, I don't need to take into account that the ideal grounding points that litter the schematic are not that - I don't want to completely humiliate the poor model ...
Indeed, if someone is a bad enough designer to structure the gain that way, they will get non-optimum results (distortion and noise as well). The idea that the designer was not an idiot was assumed in my reply, but perhaps I should have been explicit. It is unclear that the AD797's PSR scales the way you claim but never having done that poor of a designing job, I don't have first-hand knowledge.
- Status
- Not open for further replies.