This is where I depart from the papers I really worked on and therefore know pretty well, but in summary, if you look at Walt Jung's SID series, independently developed by 3 engineers, you will see how we can use THD to double check for TIM, even if the numbers might be slightly different. By 1980, we came to an agreement that 5V/us for preamps and 50V/us for power amps, given that there is no input stage 'dead zone' pretty much met the TIM measurement requirement. However, it didn't meant that op amps with more than 5V/us sounded as good as other designs. That brought up another challenge: WHY?
Now, that we had cracked the slew rate barrier with the 5534 and the TLO-72, why didn't these devices supplant all previous audio designs for studio boards, phono stages, and consumer preamps? Well, people tried to use these devices to make world beatingpreamps, but they still fell short, somehow. And back to the drawing board.
John,
glad to hear your computer has recovered.
I would try these items:
- reasonably low distortion with high order harmonics as low as possible
- high slew rate
- very low noise
- input devices imune to EMI/RFI, thus JFETs
- hi idle current of output stage (and VAS)
- hi OLG -3dB corner
- low or moderate FB
?
Regards,
glad to hear your computer has recovered.
I would try these items:
- reasonably low distortion with high order harmonics as low as possible
- high slew rate
- very low noise
- input devices imune to EMI/RFI, thus JFETs
- hi idle current of output stage (and VAS)
- hi OLG -3dB corner
- low or moderate FB
?
Regards,
Back in 1978, and into the mid-1980's, most serious designers stayed with discrete op amps or vacuum tubes, but we went in another direction for audio improvement. Even Dean Jensen, founder of Jensen transformers, made his own discrete op amp module. We just could not get the IC's to do it as good, not that we didn't try. In that time, a few papers came out that gave us further insight. Most are lost in obscurity at the moment for a number of reasons. In any case, we stayed with discrete designs, and looked especially at better passive parts and materials. Circuit board material came under criticism, and the term 'circuit board hook' had been invented by Tektronix, for DA in circuit board material. Dick Marsh, then working at LBL, tested caps and found them full of DA, for the most part. Walt Jung and Scott Wurcer found a NEW way to test for cap DA. We started extended measurements of every cap that we could find, and got results. More later.
Considering the given DUT, only look at transient information under complex harmonic loading as compared to a specifically designed complex input signal. the ear really only cares about the complex harmonic transient information in a real world signal, and really doesn't hear much of anything else, so simply test for what the ear is actually listening to. This would be a simple case of fitting the correct question to a fundamentally workable answer. I've already said this exact same thing in this thread and the earlier incarnation twice before this.
Why this simple and obvious point never seems to be heard is beyond me.
The point is that under the mathematical weighting of standardize testing regimes..this mentioned information amounts to something notably less than 10% of the 'entire signal' Thus, distortions tend to be in the 0.1% or less.
However, the ear only using those signal components..may see that 0.1% as seemingly like 10% distortion. Thus the need for proper weighting and proper analysis signals being in the chain. Test in the way that the ear listens, not to the way that engineers want to press buttons and/or use math...and see the world. See the world, or in this case, the ear -- for what it is.
These are vital and valid points.
Why this simple and obvious point never seems to be heard is beyond me.
The point is that under the mathematical weighting of standardize testing regimes..this mentioned information amounts to something notably less than 10% of the 'entire signal' Thus, distortions tend to be in the 0.1% or less.
However, the ear only using those signal components..may see that 0.1% as seemingly like 10% distortion. Thus the need for proper weighting and proper analysis signals being in the chain. Test in the way that the ear listens, not to the way that engineers want to press buttons and/or use math...and see the world. See the world, or in this case, the ear -- for what it is.
These are vital and valid points.
Last edited:
KBK, it is not that you are incorrect, but you are ahead of the topic at this time. Also, you have not specified HOW we could measure and weight 'transient information'. That is one serious reason that design engineers muddle on with conventional test equipment, and yet some reviewers of audio products don't even publish measurements.
In theory we would design audio products by 'ear' but we would be laughed at by most people, EVEN IF we were truly successful, like some Japanese tube equipment that I have heard over the years. Yes, true audio success, is NOT just TIM or low harmonic distortion, but it helps to eliminate these distortions if we can, and conventional testing helps us with this.
In theory we would design audio products by 'ear' but we would be laughed at by most people, EVEN IF we were truly successful, like some Japanese tube equipment that I have heard over the years. Yes, true audio success, is NOT just TIM or low harmonic distortion, but it helps to eliminate these distortions if we can, and conventional testing helps us with this.
Last edited:
That'd be a 1:3. Not aware of any 1:3's with a 10k input impedance.
Could take a pair of 1:1's and wire the primaries in parallel and the secondaries in series and quadruple the normal load resistance and that'd give you approximately 1:2 a 10k input impedance.
se
Ever onward, there were a few technical papers that were virtually ignored at this time, due mostly to the complexity of understanding the theoretical and practical foundation of two separate measurements not made by conventional test equipment. Both of these papers were submitted by professors of engineering, but didn't go anywhere. Yet, these papers and the deviations they implied, seemed to come from high feedback, low open loop bandwidth amps, while loudspeakers, tube amps, etc, lacked these distortion mechanisms. Could these be the hidden problems the ear is sensitive to? more later
Someone who has you beat on the low noise stakes, but I admit the slew rate is a little low..., but very good at reproducing the low bass notes.
EM D.C. PICOVOLTMETER MODEL P12 Specification.
Wrinkle
EM D.C. PICOVOLTMETER MODEL P12 Specification.
Wrinkle
The instrument is an evolutionary descendant of the Fluke 845, Keithley 155 and HP 419a null meters. It uses a chopper amp on the input with some balancing tricks to get a high input impedance. The noise number is a little misleading since bandwidth is not mentioned so its hard to relate it to our much wider bandwidth applications in audio. It really only looks at DC. I would wager there is a transformer in it as well, possibly a chopper relay or optical chopper given what they are doing. To understand some of what is going on inside, its input impedance increases from 30K when first connected to 10M over a period of time. Its an impressive achievement to get the effective noise so low.
- Status
- Not open for further replies.