John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
I can no lower myself to the same level as someone here
Making such statement not knowing what is in the box or who the designer was is enough to give you an idea...

Pictures look same to me apart that one is on a 5V division scale and shov the input

I was hoping that other members would have spotted the fact that a cartridge is only a coil of wire moving against a magnet or a magnet moving against a coil of wire.

(Not shure about the strain gauge type but I will let you know if I try one)

It wont do square waves

So what does it proves?
 
This is what PMA means.
To fix it, only needs a simple high frequency roll-off.
Maybe there is one somewhere in your system anyway, so it's OK.

edit: This is just a simple paint drawing - not very accurate.:eek:
 

Attachments

  • phono.GIF
    phono.GIF
    2.1 KB · Views: 333
Last edited:
Well, back to business then. I hope that many here have come to realize that the mistracking MC phono cartridge signal is a differentiated square wave, or a pulse with a very short duration of maybe 5us or so. This was measured and shown in Fig. 7 in my IEEE paper. The resulting output of the mistracking pulse passing through the phono stage should look like a square wave (not triangle) with a rise-time approximating 10us. This is the worst case that I could find 32 years ago. Whether there are even faster transients possible from a mistracking phono cartridge, I do not know.
 
Last edited:
OH!? Show me more! Please!

I agree, John. I'd like to see that as well.

If there is such a thing as an equivalent diagram for a cartridge, that must be for it's supposed "linear" behavior, not when it's mistracking or playing back clicks & plops.

So i'd like to know what kind of "abusive" signal from a cartridge a phono preamp be must able to handle. And worst case, please.
 
I tried a lot of records to test for cartridge mistracking, AND I always used the recommended tracking force for the phono cartridge. I had trouble capturing transients that made much sense, from normal musical selections, so I looked at the SPECTRUM of the mistracking, which typically extended to 200KHz, sometimes to 500KHz, when the cartridge mistracked the record. It happened on many records, usually on musical climaxes. The mistracking artifact is really generated separate from the music as the stylus is forced to leave the record, because of the difficult musical passage, and then crashes down on the vinyl, making a fast full output transient, mechanically more like a step, but when differentiated by the phono cartridge pickup, into a pulse. One of the 'secrets' of making good sounding phono reproduction is to handle these 'mistracking transients' with ease, not allowing for clipping, slew rate limiting, or momentary latching. That is the main reason my phono stages are more complex than the minimum possible.
 
Please look at Fig.7. It is a calibrated 5us/div in the horizontal plane. You are now showing 20us/div. Why not tighten up your scale so we can more easily compare?
By definition, you should be able to reproduce a good 10us rise-time square wave with the input pulse as shown by either the TIM test waveform as shown on Fig.8 which is similar to Fig.7, when you adjust for horizontal scale differences. After all, what goes in, before passive pre-emphasis or differentiation, should come out, is my first guess.
 
Last edited:
@John

interesting info. Can you describe with which cartridges & tonearms - in combination maybe - you realized this?
For better understanding, how do you handle the mistracked signal in the phono stage? Is it not that the mistracked signal results in distortion or do I think not correct?


@PMA
nice images from your phono stage. Are you willing to share your circuit?
 
I tried a lot of records to test for cartridge mistracking, AND I always used the recommended tracking force for the phono cartridge. I had trouble capturing transients that made much sense, from normal musical selections, so I looked at the SPECTRUM of the mistracking, which typically extended to 200KHz, sometimes to 500KHz, when the cartridge mistracked the record. It happened on many records, usually on musical climaxes. The mistracking artifact is really generated separate from the music as the stylus is forced to leave the record, because of the difficult musical passage, and then crashes down on the vinyl, making a fast full output transient, mechanically more like a step, but when differentiated by the phono cartridge pickup, into a pulse. One of the 'secrets' of making good sounding phono reproduction is to handle these 'mistracking transients' with ease, not allowing for clipping, slew rate limiting, or momentary latching. That is the main reason my phono stages are more complex than the minimum possible.

Should we talk about power bandwidth vs small signal bandwidth?
 
Please look at Fig.7. It is a calibrated 5us/div in the horizontal plane. You are now showing 20us/div. Why not tighten up your scale so we can more easily compare?
By definition, you should be able to reproduce a good 10us rise-time square wave with the input pulse as shown by either the TIM test waveform as shown on Fig.8 which is similar to Fig.7, when you adjust for horizontal scale differences. After all, what goes in, before passive pre-emphasis or differentiation, should come out, is my first guess.

I have reviewed your paper. The plots are for small signal (Fig. 8, in 10mV/div, e.g.). My measurement was done with 675mV of input signal amplitude, 1.35Vp-p, and with faster rise (shorter rise time) of input derived square. I have verified the result of measurements by math simulation, and results computed by SW are same as those measured. So I have no need to rearrange the measurement for conditions that are less severe, just to make them same as in your paper.
 
You DO have a fast input pulse, but I have difficulty knowing how fast. Those of you following this and asking questions, PLEASE look at the IEEE paper first for details. They are in the paper, often with the explanation just under the Fig.
For example: "Fig.7 'The rise time of the Ortofon MC20 moving coil mis-tracking a 1 KHz sine wave on the B&K 2010 test record at 24 cm/sec RMS."
As far as Fig.8 is concerned: It is an EXAMPLE of the ENVELOPE of the TIM(30,30) test waveform. For clarity, I OMITTED the extra 15kHz sine wave, only necessary for actually doing the test. The amplitude is not relevant , in this case, just the envelope. I just used a level high enough to get a clean picture. My only MISTAKE was to NOT use 5us/div. Instead, I used 10us/div. and it makes comparing Fig's 7 and 8 slightly more difficult for those not aware of the horizontal sweep change. I have always regretted this, but this is better than trying to compare to a 20us/div graph that PMA has put up. The 5us/div or 10us/div, come from calibration data printed on the screen by the Tektronix 7633 series storage scope that did the test. I have one on my bench, do you have something that can do the same job? If so, why not duplicate my results, or improve on them? I'm sure that the last 32 years has made faster mistracking artifacts, perhaps with faster MC phono cartridges.
 
Simon, I am measuring high output RISETIME, not power bandwidth. You know: dV/dT.
For example, if you take a 3kHz SQUARE WAVE, and do a spectrum analysis, you will find that it gives series of harmonics that drop 6 dB/octave or 20/dB per decade. I just tried to explain what I was up to in the paper, but I missed you by telephone. It would seem less 'elementary', if we spoke in person about this. However, for everyone else, please try to understand this. A perfect square wave, has an INFINITE bandwidth. This is impossible, so we must always LIMIT the bandwidth of the square wave. The least invasive way is to use a 6 dB/octave RC filter, to bandwidth limit the square wave somewhat. It still goes on 'forever' but it starts to drop at 12dB/octave, rather than 6dB/octave. Our TIM test in this case is based on a TIM(30) test signal that is defined as an additional RC time constant rolling off at 6dB/octave at 30KHz, to even make the test waveform then roll off the harmonic series, at a total of 12dB/octave. Still, this is a 3KHz square wave that has a 10us risetime. Yet there is no full level high frequency information that you might imply power bandwidth with. That is for single sine waves that have to extend at least to 20KHz, full power, in order to get similar dV/dT that we are trying to measure. That is WHY we used a square wave in the first place, rather than two tone IM or a single high frequency. We wanted to use a more REALISTIC signal, than a full power sine wave, or even 2 tones at high frequency. (Test on Friday)
 
Last edited:
Status
Not open for further replies.