In the summer of 1976, we made 100's of measurements, as I said before. In making one measurement, I realized that just flipping a switch on a function generator, could separate IM from TIM in a test. I just turned the switch to a triangle wave, instead of a square wave, which gave a similar harmonic series, BUT dropped the slew rate drastically.
Of course, quibbles from the critics, but it worked really effectively.
We also found that many IC's that were popular AT THE TIME were sort of lousy with the test, especially the UA741, used in studio boards and other pro audio electronics around the world, at the time. I concentrated on a phono stage and what happened. More later.
Of course, quibbles from the critics, but it worked really effectively.
We also found that many IC's that were popular AT THE TIME were sort of lousy with the test, especially the UA741, used in studio boards and other pro audio electronics around the world, at the time. I concentrated on a phono stage and what happened. More later.
What was IMPORTANT about the square-triangle comparison test is that we could virtually directly compare SMPTE IM with what the TIM IM test measured, and we could separate the STATIC from the DYNAMIC (TIM or SID) distortions. This gave us great insight to what was wrong with op amps at the time.
When it came to phono preamps, the situation got even more interesting. I modeled a phono preamp similar to a JC-2 phono stage, except that it used an IC for the gain element, rather than a discrete class A design that was in the JC-2.
To make it right, we had to put an RIAA network around the IC, AND we had to put a ANTI-RIAA network between the test oscillator(s) and the input of the phono stage. This is important to make the final output look like the generator output, and to emulate what REAL phono cartridges actually put out because they 'differentiate' the stylus movement, that make them rise in output at 6 dB/octave until they go into mechanical resonance. This got some really interesting results. More later.
When it came to phono preamps, the situation got even more interesting. I modeled a phono preamp similar to a JC-2 phono stage, except that it used an IC for the gain element, rather than a discrete class A design that was in the JC-2.
To make it right, we had to put an RIAA network around the IC, AND we had to put a ANTI-RIAA network between the test oscillator(s) and the input of the phono stage. This is important to make the final output look like the generator output, and to emulate what REAL phono cartridges actually put out because they 'differentiate' the stylus movement, that make them rise in output at 6 dB/octave until they go into mechanical resonance. This got some really interesting results. More later.
When we had done our tests and went our separate ways, 2 papers, at least that I was directly associated with, were presented with a summary of the data we had gathered.
The first was for the AES, and first presented in 1976 at the NY AES Convention by Matti Otala. I recommend that you get a copy if you are interested in TIM, and its formal measurement.
It was published in the JAES, April 1977, Vol. 25, #4 pp. 170-177. The title is:
'A Method for Measuring Transient Intermodulation Distortion (TIM)' Erno Leinonen, Matti Otala, and John Curl
The second article is: 'OMITTED FACTORS IN AUDIO DESIGN' John J. Curl
IEEE 1978. Tulsa Conference. order # CH1285-6/78/0000-0263$00.75@1978IEEE
I highly recommend that you get these two articles through the representative organizations, if you don't have them.
Unfortunately, I cannot at this time send them out, as my computer is still giving me trouble, and I don't know when it will be fixed, so that I can again e-mail info.
However, I feel that attempting to discuss the graphs and charts in these papers will be pointless, as it is technical and my word alone is contested here by the moderators.
So, if anyone is really interested, please get the articles and read them.
The first was for the AES, and first presented in 1976 at the NY AES Convention by Matti Otala. I recommend that you get a copy if you are interested in TIM, and its formal measurement.
It was published in the JAES, April 1977, Vol. 25, #4 pp. 170-177. The title is:
'A Method for Measuring Transient Intermodulation Distortion (TIM)' Erno Leinonen, Matti Otala, and John Curl
The second article is: 'OMITTED FACTORS IN AUDIO DESIGN' John J. Curl
IEEE 1978. Tulsa Conference. order # CH1285-6/78/0000-0263$00.75@1978IEEE
I highly recommend that you get these two articles through the representative organizations, if you don't have them.
Unfortunately, I cannot at this time send them out, as my computer is still giving me trouble, and I don't know when it will be fixed, so that I can again e-mail info.
However, I feel that attempting to discuss the graphs and charts in these papers will be pointless, as it is technical and my word alone is contested here by the moderators.
So, if anyone is really interested, please get the articles and read them.
John please stop referring in general to moderators. Your word has been contested by some of the moderators as much as it was contested by any regular member.
Especially ironic considering that I'm hosting the "omitted factors" paper.
It's easier to play martyr than to engage in serious discussion and review.
Well everyone, now that we all have access to these two papers, both published by their respective organizations, perhaps we can go through some of the main points that are part of the 'history of audio testing' which is the subject at hand. I suggest that you read pp. 170-173 of the JAES paper and then we can discuss the TIM test in more detail than casual conversation.
PMA, at least from Silicon Valley's point of view, TIM was not really recognized in an orderly form that could be analyzed. YES, slew rate was noted, especially in IC op amps, BUT it was NOT specified, except as an end effect. The same criticisms here, brought up by Scott Wurcer and yourself were put forth in 1977, by others, so I did my own independent research of slew rate, its definitions, etc, looking at every textbook, IC catalog, and every research paper that I could find. Only the END STATE of slew rate limiting was defined and measured, not the IM or HD that is associated with it BELOW actual slew rate limiting. That is why we had to make some extra measurements and we created the TIM (30) or TIM (100) test standard , because the Crown IMA wasn't giving us this distortion, in fact, it ignored it.
At this point, I would like to bring in another set of papers that were proven valuable in the middle-late, 1970's, and that is the SID papers from Walt Jung that should be on his website. For the record, SID=TIM, for all intents and purposes, but Walt Jung led the way to make TIM (SID) measurement easier and faster with extended harmonic distortion testing. More later.
Hi John,
For a complete and balanced discussion of measurement techniques, results and history, we should not omit the papers about TIM, MIM, IIM, and PIM on my site at Cordell Audio: Home Page.
Cheers,
Bob
- Status
- Not open for further replies.