Moving magnet phase shift

0 Members and 1 Guest are viewing this topic. Read 37432 times.

dlaloum

  • Full Member
  • Posts: 710
Re: Moving magnet phase shift
« Reply #60 on: 2 Apr 2011, 10:34 am »
Hi Roger

If I am using the wrong terms, please excuse me, as I am not an EE....

No question about the RIAA EQ and its correcting the recorded phase EQ (spent a lot of time making sure my digital RIAA was phase correct....).

The issue is really with cartridge behaviour...

First the is the reactivity between inductance and capacitance leading to an electrical resonance.

Then there is the cantilever which also has a mechanical resonance.

So far every cartridge I have measured shows signs of the designers using the M-Resonance (mechanical) to balance out the F/R droop generated by  inductance/capacitance, and sometimes the use of the E-Resonance (electrical resonance) to fill out and flatten the curve as well.

So taken individually we most likely have phase issues at a minimum of 2 resonance points, plus there may also be some form of phase error as the electrical F/R drops off due to high capacitance/inductance. (not sure of that last one....)

If the cantilever resonance (M-resonance) is beyond the audible range (say 25kHz +) and the e-resonance is also pushed way out through the use of either or both of low inductance and low capacitance - then there would be little or no phase error.

So far I have measured all the following  cartridges / styli and all have had signs of M-Resonances within the audio band....

Shure 1000e with M97xE-SAS
AT440MLa
ADC SuperXLM
ADC Digital Series II
Ortofon OM20
Ortofon/Digitrack 300SE

As a result, achieving a flat F/R with all of these requires balancing an electrical HF droop with a mechanical peak.... as a minimum, and depending on the plot, sometimes using a slight E-Resonant peak to fill in a mid-high droop.

I hope I am making sense.... if not I may need to post some actual plots...

It seems that since the mid 70's cartridge design has been going backwards....

In the heyday of CD4 quad, they had to design for minimal phase issues (as the Quad FM decoding was/is sensitive to phase) and the system required adequate response out toe 38kHz...

To achieve this, the classic cartridges tended to be low inductance, required a low capacitance environment (100pf was spec) and a high impendance load (100k). Cantilevers tended towards more exotic designs and we saw the release and use of Ruby/Saphire/Diamond/Boron/Beryllium and other exotic designs - really intended to push the resonance out far enough so it is no longer a problem... (this last is my own assumption).

Current MM/MI practice tends towards high inductance (500 to 700mH seems standard), in a high capacitance environment (phono stage load of 220pf, + standard cables rather than Low C cables = Total of circa 350pf, and frequently much higher).


I would be interested in finding a simple way to actually measure the phase response of the system.... for that it would require obviously some form of test signal (test record), and some form of measurement software to analyse it.
I have software to do this function for room/speaker analysis, but it generates (and assumes the presence of) its own tones....

As an aside - this appears to be less of an issue with MC's due to their vanishingly low inductance.... making capacitance relatively irrelevant, and therefore requiring the designer to control or push out the M-Resonance outside the audio band - probably the reason we see so many exotic styli in the genre. Not to mention that with a megabuck budget it is easier to fit these sort of technologies, doing the same for a $200 cartridge is a lot more difficult!

Bye for now

David

P.S. I am also aware of the VR term - but generically Moving Iron seems to be the term that is best understood for this type of system.... Ortofon, Nagaoka, and Grado all continue to use this, but there are still plenty of true MM designs.... eg: Shure, Audio Technica
Apparently some of the MI/VR type designs also use magnets on the stylus mounting to boost the non magnetic cantilever stub - some of the latest OM and 2M styli seem to do this, and some of the TOTL vintage Empires also had magnets mounted on the stylus shank (but NOT on the cantilever.... so they are still not moving magnet designs.)

bauzace50

  • Jr. Member
  • Posts: 21
Re: Moving magnet phase shift
« Reply #61 on: 2 Apr 2011, 10:58 am »
Gentlemen,

the following is not part of "phase shift", but the most recent two posts bring up an old (but still unfulfilled) curiosity...and certainly is the theme for a separate thread.

     Are the design characteristics of MM, VR, MI, "Induced Magnet", sufficiently different to make each a   unique principle?

     Or,  are these design differences really a rehash of topology with no performance advantages?  Or restating in another manner:  do any of these principles have a performance advantage over the others?

@ neobop: not wishing to twist this thread! :oops:  This is best posed in its own separate thread.

Regards,
bauzace50

dlaloum

  • Full Member
  • Posts: 710
Re: Moving magnet phase shift
« Reply #62 on: 2 Apr 2011, 12:42 pm »
A different but interesting Pandora's box?  :icon_twisted:

neobop

  • Full Member
  • Posts: 3448
  • BIRD LIVES
Re: Moving magnet phase shift
« Reply #63 on: 2 Apr 2011, 01:36 pm »
I think these cart types are grouped together because the have high output. After all this is a consumer oriented biz, like all biz, and it's confusing enough.

I think I was speculating near the beginning of this thread about the preamp or RIAA network contributing to phase shift, and the possibility of phase consequences like a crossover. Thanks for clearing that up, Roger.

The Ortofon article on cantilever damping vs phase was a real eye opener. It's the only time I ever saw phase and amplitude plotted together. I based all my conclusions on those plots and text. Ortofon used an accelerometer, mounted on the headshell? Obviously I'm not an EE either. There is a AES paper that corresponds to the article, That should be interesting (1983).

David, I agree with you about HO designs mostly going backward. I suspect biz considerations are mostly to blame, and the near extinction of records. So any company considering the market and long term possibilities might be reluctant to jump into the mass market of phono carts. There seems to be a proliferation of pricey MC designs. Using the same market considerations, this makes sense to me. They might figure that long term, it's fanatical hobbyists sales that will generate the profits, and not the mass market. Many of these carts are like buying an expensive piece of jewelry. It's a hand made item and profits are realized much sooner, if there are sales.
neo



Roger A. Modjeski

Re: Moving magnet phase shift
« Reply #64 on: 2 Apr 2011, 05:01 pm »
Thanks to everyone for their replies. My main interest is very low noise preamp design. I see the group here had done a lot of research and thinking about the cartridges which I deeply appreciate.

I would like to read the Ortofon article, see some articles on how mechanical and electrical resonances are used to counteract each other and any pictures of how the real MM cartridges look inside.

I am also happy to answer brief (i hope) questions about noise and other performance issues in preamps.

neobop

  • Full Member
  • Posts: 3448
  • BIRD LIVES
Re: Moving magnet phase shift
« Reply #65 on: 5 Apr 2011, 10:19 am »
The issue is really with cartridge behaviour...

First the is the reactivity between inductance and capacitance leading to an electrical resonance.

Then there is the cantilever which also has a mechanical resonance.

So far every cartridge I have measured shows signs of the designers using the M-Resonance (mechanical) to balance out the F/R droop generated by  inductance/capacitance, and sometimes the use of the E-Resonance (electrical resonance) to fill out and flatten the curve as well.

So taken individually we most likely have phase issues at a minimum of 2 resonance points, plus there may also be some form of phase error as the electrical F/R drops off due to high capacitance/inductance. (not sure of that last one....)

If the cantilever resonance (M-resonance) is beyond the audible range (say 25kHz +) and the e-resonance is also pushed way out through the use of either or both of low inductance and low capacitance - then there would be little or no phase error.

So far I have measured all the following  cartridges / styli and all have had signs of M-Resonances within the audio band....

Shure 1000e with M97xE-SAS
AT440MLa
ADC SuperXLM
ADC Digital Series II
Ortofon OM20
Ortofon/Digitrack 300SE

As a result, achieving a flat F/R with all of these requires balancing an electrical HF droop with a mechanical peak.... as a minimum, and depending on the plot, sometimes using a slight E-Resonant peak to fill in a mid-high droop.

I hope I am making sense.... if not I may need to post some actual plots...

I've been thinking about this post, David. As another non-EE, I'm still trying to figure out the role of electrical resonance, specifically. I'm convinced that looking at the output of a HO cart as an inductor and a voltage generator at the same time, is not exactly correct. Electrical resonance does not show up on the measured response of carts AFAIK. If you can show me otherwise, I'd be very interested. In examining the Ortofon phase vs amplitude plots, there seemed to be no hint of electrical resonance per se, as having any presence. The role of capacitance/inductance seems to be the curving and extending the phase shift plot. If you look at the plots again you'll see that all primary phase reversal came from the high frequency resonance beyond the limits of the graph. The mechanical HF resonance defined the frequency of the MC phase reversal and I assume the MM also. The undamped MC had phase shift over a very narrow range that didn't go into the audible band. It is mechanical damping that flattens and extends the phase shift. The text implied that MMs require more mechanical damping. If moving mass is higher, this would make sense. I suspect that electrical resonance might define the frequency where this curving/extending of phase shift occurs.
neo


dlaloum

  • Full Member
  • Posts: 710
Re: Moving magnet phase shift
« Reply #66 on: 5 Apr 2011, 11:04 am »
Hmm well I'm working on identifying the weird bits - ie that part of the response which is not pure inductor/capacitor behaviour
The first chart is a series of measurements of an AT440MLa with system total Capacitance at 60pf and various resistances....

The nice smooth curves are the theoretical modelled behaviour assuming a perfect inductor/capacitor - the jagged plots are the actual measured F/R


Here is a plot where the model has been deducted from the measured


The difference between the two is that part of the behaviour which is either "electrically imperfect" (or more correctly, which a simple model does not reflect!) - and the mechanical resonance.


Take a look at the graph for 21k resistance...

The mechanical resonance (which is most likely 80%+ of the delta plot) is being balanced out by the inductor/capacitance drop resulting in almost perfectly flat response from 7kHz onwards.

This shows up the mechanical resonance which exists on this AT stylus - somewhere out around 21kHz perhaps (I didn't measure high enough to see the peak and drop off, so I don't know the exact resonant frequency)

It also shows that by selecting a Resistive loading of somewhere around 26 to 30k is most likely optimum for the AT440MLa - that is where the balance between the electrical and mechanical behaviour will result in something as close as possible to a flat F/R.

BUT - the question remains as to what happens to phase?

I still don't have a way of measuring phase - electrically the capacitance and inductance are low enough to avoid a resonant peak.... but the mechanical resonance is clearly there, and will be affective phase.

Here is the same type of test using a Digitrac 300SE (basically an Ortofon engine with a stylus that under the microscope looks the same as an OM30... but half the price)





In this case you need to raise the capacitance to bring the electrical resonance back, so that you can fill the dip in between 7-10k... again the resonance is clearly there at the HF end....


Here you can see that the 45k resistance plot is better....

These are both examples of styli with mechanical resonance at the HF top end.

My SAS stylus has a resonance around 14kHz, Ed Saunders V15 styli have a resonance around 11kHz.

So the optimum cartridge loading is determined by the cantilever design rather than the generator design.

The Ed Saunders stylus has a soft resonance quite low down, to fill out the mid-high dip - as a result it sacrifices some high end (especially in a higher capacitance setup)

The AT & Ortofon styli accept a midrange dip in exchange for more extension in the high end.

There is also the likelihood that moving the m-resonance further up also moves the phase anomalies further up to where they are less audible (and possibly inaudible?).

Choose your tradeoff!

bye for now

David

P.S. the AT440 measurement used a CBS test record, spun at 45rpm with appropriately adjusted RIAA EQ - giving accurate measurement out to around 26kHz

The 300SE measurements were taken some weeks back with HFN test record Pink Noise track - which unfortunately has an anomalous peak in the extreme high end - so the measurements above 16kHz cannot be trusted.... there is in fact a mechanical peak there, but it is exaggerated by a peak on the pink noise track as well... so I need to redo my 300SE measurements. :(

I now do my measurements with either the HFN Pink Noise at 45rpm (which gives me trusted results to 21kHz) or as above with the CBS record at 45rpm



neobop

  • Full Member
  • Posts: 3448
  • BIRD LIVES
Re: Moving magnet phase shift
« Reply #67 on: 5 Apr 2011, 12:27 pm »
Jeez David, I was trying to decipher the 1st set of plots. You've got 10,  superimposed. Maybe I'll try to magnify it later. It is a really cool thing to do, and I commend you for that. You have 83K and 97K, but you go down to 10K and 21K, then jump to 46K? I'm sure you've read that many prefer 32K. 60pF total capacitance is unrealistic for most. I suspect that if you take it up to 100 to 200pF (as recommended by AT) it would fill in the dip in the lower treble. Then load with 32K and you'll have near flat response.

I think the effect of capacitance/inductance is to lower the HF resonance to augment the treble but to roll off the extreme high end. As you say, the mechanical response is the major factor in determining FR. Loading should be used to compliment the mechanical response IMO. It seems like you approach from the other direction. You only really have mechanical response to work with. Electrical response is so theoretical, it's near worthless. If you have a high inductance cart you know the implications. IMO you should start with capacitance below or at the bottom of recommended range, then find flattest FR with resistive loading. Once you're in the ballpark you can tune both. Most of these big cart companies like AT and Ortofon actually know what they're doing. They might have no choice but to recommend 47K, but I doubt if capacitance recommendations are arbitrary.
Ortofon 2M - 150 to 300pF
AT 150, 440 - 100 to 200pF
Goldring 1042 - 150 to 200pF

Tonearm/cart resonances can also effect your plots and show up as mechanical resonance. I think it's the HF resonance that is of primary concern.
http://www.tnt-audio.com/sorgenti/load_the_magnets_e.html

neo

dlaloum

  • Full Member
  • Posts: 710
Re: Moving magnet phase shift
« Reply #68 on: 5 Apr 2011, 01:13 pm »
I started out having a quick look, and on my first set of measurements it looked like the delta would vary by the same amount, as I doubled the R loading... so I initially selected 10/20/40/80K (roughly) as my values.... then I added 100k due to Raul Iruegas' plugs and the old CD4 quad standard...

My Capacitances are limited by the C loading plugs I've made up so far.

I figured my base capacitance would cover the "low" territory (including 100pf) and then I wanted something that would simulate "standard" loading - most phono stages have 220pf then you add the table and cables... so 300 was about right... and then I wanted a notch higher...

I have it in mind to maybe make up a couple more C plugs for fine tuning. (I need to be able to get very close to the 100pf spec... and then maybe something around 175pf....)

But I've found that this gives me enough data to be able to plug into my model - I use an interpolation of the Delta values added to the electrical model - then I can vary C and R in the model and predict the end result with some degree of accuracy...

I only got this far the last couple of days.... paint is not dry yet!

Now I need to look at some of my prediction/interpolations and try some of them out to see whether the real results match the prediction. (I must say the underlying model maths is LuckyDog's work... the interpolation and delta stuff is mine)


I have completed a test run with the  AT440MLa, Signet TK6-440MLa, AT150-440MLa, Shure MT105p-SAS.

And I have earlier data (not as good....I was not being quite as rigorous at the time) which gives me quite a bit of info on Digitrak 300SE, Shure 1000eSAS.

Interestingly, I am finding that the best (flattest, dip is best filled while maintaining F/R extension) is with the TK6.
I expected better from the AT150, but the 440MLa stylus becomes the constraint.
I wonder where the resonance sits an an ATN150 stylus??

In any case it becomes possible with these models and spreadsheets to set a cartridge up for +/- 1.5db worst case.... so far best case seems +/- 1 db primarily due to the low frequency hump. (it is of course only a hump in comparison to the mids/highs)

I have not had a chance to focus in on the low frequency resonance(s) yet....

I am running these tests on a JVC QL-Y5F with servo damping, so they should be relatively controlled.
Also the Shure cartridges have the damper brush - but it makes no difference to that bass hump (100Hz to 500Hz roughly +1db)

And results measured on my Revox TT (linear 4g effective mass, no damping) are showing the same Low/mid hump. - So I suspect these are an electrical rather than mechanical artifact.

At some point soon I will map out some MC's as well.

In any case getting back to the topic at hand - either type of resonance (mechanical or electrical) throws phase off... and it is not clear whether when you cancel out the resonant peak, it cancels out the phase variation - my intuition is that it doesn't - which would mean you might have flat F/R but phase all over the place due to both electrical and mechanical resonances being present.

lots to think about....

bye for now

David

neobop

  • Full Member
  • Posts: 3448
  • BIRD LIVES
Re: Moving magnet phase shift
« Reply #69 on: 5 Apr 2011, 02:44 pm »
I beg to differ. Electrical resonance has no perceivable effect on phase response, IMO. I just looked at the Ortofon plots again. You can't argue with measured response unless their methodology was faulty. I doubt that it was.

Although all of the peaks went beyond the top of the chart, they all were at least around 17K as minimum and 1 was obviously well past 20K. If these were centered around electrical resonance, some would probably be much lower. I think CarlosFM had it right in the first place on the VE thread. Inductance combined with capacitance, has the effect of lowering the high frequency resonance. Capacitance also will compensate for treble response dips while lowering high frequency extension. I was hoping someone could come up with electrical model for exactly how electrical resonance fits in, but at this point I wonder if it does.

There's no question that the mechanical aspects overwhelm all other considerations regarding both phase and amplitude response. Stick a beryllium/ML on that 440 and the frequency response sounds nearly flat at 47K, low capacitance. Put a SAS on a M97 and Felix goes back to 47K from 62K with stock stylus. BTW, the 150MLX has response to at least 30K.
http://www.vinylengine.com/phpBB2/viewtopic.php?t=33679
neo

neobop

  • Full Member
  • Posts: 3448
  • BIRD LIVES
Re: Moving magnet phase shift
« Reply #70 on: 7 Apr 2011, 12:36 pm »
David,
If I seem outspoken on this subject of phase response, I guess it's in reaction to that Asylum thread. Sorry, moving on....

Your plots of the 440 seem to confirm the "rightness" of user experimentation almost unanimously winding up at 32K. I think it's a safe bet that no one had 60pF total capacitance, so the depression should be somewhat less. This seems similar to Werner's measurements (TNT) of the M97 confirming Felix's use of 62K for better response.  The original 440ML OCC was spec'd for response to 32K. The generators of the ML and MLa seem identical. Resistance, inductance are the same. It seems that the OCC had stronger (neo?) magnets and output was slightly higher. The inductance is 490mH @ 1K and impedance (not resistance) is 3.2K.

That JLTI you got looks great. Congratulations!! It should be easy now to load capacitance right along with resistance. Just use a Y adaptor or stack-able plugs. I'm sure we'll get some very cool info in future.

It seems to me that the primary high frequency mechanical resonance is the major determining factor in both response extension and phase behavior. Inductance/capacitance will effect this, but it seems that electrical models are somewhat overrated in this respect except in the case of high inductance carts. I believe the Ortofon has even more inductance than a 440? Actually, 490mH isn't as high as many modern offerings. I think the 2M are around 600. We can see by your plots with the 300SE, what happens to extension. The same thing happens to phase linearity. The lower the HF extension, the lower the phase shift goes. I think you can also see with the 494pF plots just how flawed the electrical model is. Maybe flawed isn't the right word - inconsequential maybe. AFAIK, amplitude response drops off like a rock at HF resonance.
neo

dlaloum

  • Full Member
  • Posts: 710
Re: Moving magnet phase shift
« Reply #71 on: 7 Apr 2011, 02:16 pm »
Not quite...

It appears, that resonances (whether mechanical or electrical) cause substantial phase anomalies...

So wherever the setup is such that the base modelled (pure electrical) response has no "peak" - phase response will nicely follow amplitude response....

BUT: if an electrical resonance (the peaks shown in the model plots) is present and used to balance out the frequency response, then although the amplitude frequency response is (relatively) flat, the phase response won't be...

On the 300SE plot, all the R values above 45K have a resonant peak.... so there would be phase anomalies at all the higher resistances in a 149pf capacitance setup.

Now this is all based on theory - I still have no way of measuring phase.

for situations where the mechanical resonance rise begins above 15kHz and the peak itself is above 20kHz, I think phase will most likely be a non issue in terms of audibility. (at least in situations/setups where there is no electrical resonance)

But based on what I have been learning about real life response - yes the stylus and the mechanical resonance are absolutely critical.

With a stylus with a resonance high up (higher is better), and a cartridge with low inductance, combined with low C cables and load, it is theoretically possible to extend response out well beyond 20kHz.
In actual fact the objective would be:
1) Avoid electrical resonance (limit R load to just below resonance point as max)
2) Extend sufficiently far and tailor the rolloff curve so it provides the inverse of the mechanical resonance - thereby resulting in extended HF response
3) Select a cantilever which places the resonance at 21kHz+, with a relatively steep rise to the resonant peak

To make things more interesting (!?) there is also the midrange droop and some frequency related anomalies that happen when the voltage (signal) drops below a certain point - these are most likely related to inductor core permeability - and are likely similar to the hysteresis curves reflecting magnetic behaviour in tape heads.

Once you start digging into vinyl - the fact that it works at all is pretty miraculous - that it works so well is astounding!

In any case I am starting to meander away from the phase shift topic.

I have made an effort to obtain low inductance MM/MI cartridges (as it extends the electrical F/R.... and makes things easier in combination with adjustable capacitance)- which is not as easy as it may seem.

ADC SuperXLM and ZLM were low inductance, the rest of the XLM/QLM/VLM series (which look identical!) were high inductance.
The ZLM's are usually labelled (and usually expensive) - the SuperXLM's were often NOT labelled other than on the stylus.

So with this 40 year old cartridge, it is pot luck - you buy a cartridge on fleabay... get it in, then measure it - and try again until you get one of the low inductance ones!

The AT12 series in the early days were low inductance - later they moved to higher inductance - but right through both stages they were called AT12Sa... same colour, same look body - only way to tell is to get one and measure it! (AT13/14/15/20's also had early series with low inductance and later with high)

There are other examples out there...

Once you have a low inductance generator, then you need a stylus with the right type of resonance curve and placement....
I have tried with the AT102/120 family (440MLa) - the resonance is too low down the frequency range (as can be seen in the delta plots - which are 95% mechanical resonance... by my guess) - I don't have a 150MLx stylus to test - but that may be an option at some point

The 300SE curve is far closer to my target. (Ortofon has been banging on about phase for a long time... so their design approach may have some synergy with what I am trying to do!)

I did some initial testing with two ADC styli, a Digital Series II eliptical and a SuperXLM - both had very similar profiles to the 300SE.... so the ADC cartridges may be a good match for what I am trying.

I had been hoping that the SAS stylus would be a good fit as well - but at least in its N97xE Version, the resonance is relatively broad, not too extreme (sever db) and located in the high centered around 14kHz (with impact going down to 10k and up to 18kHz.

So although it is a great tracking stylus, it is unlikely to meet my requirements (although it may make a great "stalking horse" in that theoretically the "ideal" setup should sound better....)

And yes I am very happy with the JLTI - it has loading plug RCA's on the back into which I have then plugged metal RCA double adapters - C loading goes into one, R loading into the other.

And yes you are right, once an electrical resonance occurs, the amplitude response curve becomes far steeper after the resonance....

I believe the 2M series are over 700mH, I have an OMP body at around 400mH and most OM's are 460mH (ish)

The Digitrak 300SE is around 520mH (might be the same as an OM Super?)

The Technics EPC-P22/23/24/30/33 series are interesting as they are cheap as chips, readily available and have inductance around 350mH. (I have a HE stylus on the way to try out)

The end result is the sum of the electrical and mechanical response - so both have to be right to achieve the desired outcome.

Unfortunately, there is absolutely no way of knowing the mechanical behaviour of a cantilever from its specifications. - You have to buy it and test it.... so I am on a fishing expedition for appropriate styli - and preferably ones that can be fitted to a low inductance cartridge.

I may of course be tilting at windmills ....

This would be a lot easier if phase measurement tools were as readily available as level measurement tools.

oh well

I'll be off now, I see another windmill in the distance

David






neobop

  • Full Member
  • Posts: 3448
  • BIRD LIVES
Re: Moving magnet phase shift
« Reply #72 on: 7 Apr 2011, 05:09 pm »
Not quite...

It appears, that resonances (whether mechanical or electrical) cause substantial phase anomalies...

So wherever the setup is such that the base modelled (pure electrical) response has no "peak" - phase response will nicely follow amplitude response....

BUT: if an electrical resonance (the peaks shown in the model plots) is present and used to balance out the frequency response, then although the amplitude frequency response is (relatively) flat, the phase response won't be...

On the 300SE plot, all the R values above 45K have a resonant peak.... so there would be phase anomalies at all the higher resistances in a 149pf capacitance setup.

Now this is all based on theory - I still have no way of measuring phase.

for situations where the mechanical resonance rise begins above 15kHz and the peak itself is above 20kHz, I think phase will most likely be a non issue in terms of audibility. (at least in situations/setups where there is no electrical resonance)
This would be a lot easier if phase measurement tools were as readily available as level measurement tools.

David

Maybe because you're in your PC making this analysis, phase linearity will be easier to plot somehow. Other than that, AES articles can be purchased, and Ortofon probably describes the use of an accelerometer in their corresponding article. Hopefully there will be a much more detailed description of both methodology and cart specifics.

I fail to see where electrical resonances have any effect what so ever on phase. That's my point. If somehow you get phase plotting capability, I'll be very interested in results. In the mean time, my conclusions are based on the Ortofon measured results, not conjecture. There was 1 primary frequency that was cause of all phase non linearity. That was the high frequency resonance. It appeared as if the HF resonance was lowered by cap/inductance in MM  cases. That would be interesting if in fact the primary frequency was the electrical resonance, but it didn't look like that was a possibility. 

I think that eventually, you'll need plotting capability to at least 30K, preferably higher. 50K would be better. That way you could really see what's going on up there. You mentioned audibility. The Ortofon conclusion was that it was clearly audible with all carts. It was only the mechanically undamped MC that had obviously superior imaging capability. However, it was unlistenable undamped because of the frequency response. With damping, MC  phase shift was extended down to 8 - 10K and imaging was not as good. All the MMs had phase shifts extending lower, some to 2K. None had imaging capabilities that competed with the MC. The MC had a HF res of 28K. My conclusion is, if response is more extended, imaging will tend to be better.
neo

BaMorin

  • Full Member
  • Posts: 407
  • AR turntable rebuilder/modifyer
Re: Moving magnet phase shift
« Reply #73 on: 7 Apr 2011, 10:46 pm »
The AT12 series in the early days were low inductance - later they moved to higher inductance - but right through both stages they were called AT12Sa... same colour, same look body - only way to tell is to get one and measure it! (AT13/14/15/20's also had early series with low inductance and later with high)

David


Must have been why I thought my AT 12Sa sounded much better than my AT440MLa.

(with "normal" 47K)



HAL

  • Industry Contributor
  • Posts: 5180
Re: Moving magnet phase shift
« Reply #74 on: 7 Apr 2011, 10:53 pm »
Has anyone researched the Kinergistics phono stage that had inductance cancellation as part of the design? 

neobop

  • Full Member
  • Posts: 3448
  • BIRD LIVES
Re: Moving magnet phase shift
« Reply #75 on: 7 Apr 2011, 11:56 pm »
Wasn't that a MC stage? Apparently it wasn't well thought of. I found this DIY forum that mentions it, among other things. John Curl joins in on the discussion.
http://www.diyaudio.com/forums/analogue-source/32318-current-amplifying-phono-stage-3.html
neo

neobop

  • Full Member
  • Posts: 3448
  • BIRD LIVES
Re: Moving magnet phase shift
« Reply #76 on: 17 Jun 2011, 01:27 pm »
I was looking over this thread - it's been a couple of months, and maybe it's time for a progress report. There seems to be a discrepancy between David's statement about mechanical resonances other than the primary HF, causing phase shift vs my contention that the HF primary are the defining cause of all phase shift in both LO and HO carts. I know David is doing some great work measuring different aspects of cart performance. I don't expect him to have phase linearity measurement capability, but a reexamination of some evidence we already have might help.

Why is it that that the measured phase performance shows no evidence of any phase anomaly other than that caused by damping the moving system and primary HF resonance? I think this question defines the problem. It would seem reasonable that any resonant peak might cause a similar or secondary phase anomaly. Maybe not. The primary peak is severe and causes a phase reversal on either side of the peak. There could be other factors as well. Is mud caused by arm/cart resonances or cart vibrations also phase?

All psychics and physicists welcome to respond.

neo

dlaloum

  • Full Member
  • Posts: 710
Re: Moving magnet phase shift
« Reply #77 on: 17 Jun 2011, 01:59 pm »
Hi Neo,

the phase question has not faded from my mind....
I found a test record apparently intended for phase measurement, it was delivered this morning - I have not yet removed it from its packaging and have no idea whether or not it will help shed any light on things!

The physics or resonance peaks is well known (apparently.... from my reading of other people's conclusions!).

In an attempt to remove resonances from the audible range of a cartridge - I have been spending most of my time working on cartridge loading...

Then I started looking at cartridge loading to move resonances "out of harms way" - so to speak - while using digital linear/phase neutral equalisation to provide flat F/R... - With very good preliminary results (how about +/- 0.3db from 80Hz to 20kHz !?! )

Until I get some tools with which to investigate phase - all I can do is focus on methods of tweaking the system that eliminate possible phase issues.

To which end:
1) select a stylus with as high a resonance as possible - so far the best are ADC RZL, Ortofon OM20/30 and AT15/20ss - all have mechanical resonance around 21kHz to 23kHz - all of them show resonant amplitude influence down to at least 15kHz and possibly lower... - so there is a likelihood of some phase anomalies above 15k with all of these - but that leaves the main audio range phase linear...
2) optimise cartridge choice and loading to completely avoid electrical resonance - I believe the EE term is keeping the loading below the critical threshold - while adjusting that same loading to provide as smooth a Frequency response as possible. (so far in every case this has meant raising the loading resistance to as high as possible without passing the threshold - while keeping capacitance as low as possible)
3) Use Digital EQ - keeping the adjustment very "smooth" in terms of the adjusting curve - using pink noise to set the parameters...

I am in the process of making comparative recordings with at least 6 different cartridges each at different loadings/setups and each either with or without EQ... this is a lengthy process to say the least. (each EQ'd setup takes quite a while to get right.... even with automating software, the final touches are manual)

Once I'm done with the recordings I will get into the in depth comparison....

At present my preliminary work was done to ensure that the digital recordings I made are sufficiently high quality that I cannot discern a difference between digital and analogue - got that pinned down a couple of days back. (and very satisfying it is too...)

I have also compared my ADC SuperXLM at 2 different loadings (quite clearly different) to the same cartridge with the same 2 loadings with digital EQ applied - 45minutes careful listening and re-listening to a range of tracks and I cannot differentiate between the two. First setting was with 50pf/83k second was with 300pf/47k. Both met the "no resonance" critical threshold requirements. - I probably need to add a third setting to the mix, one with a distinct electrical peak, then equalised - ie intentionally generate a phase anomaly in the audible range, and see whether that is then audible. (doing that is another afternoon's work....)

Plenty going on in Laboratory Laloum..... (and I haven't even pulled out the ouija board or the crystals!)

Bye for now

David

P.S. the 50pf/80k ADC setup sounds very "airy" wide soundstage, nice!, the 300pf/47k sounds more constrained, shut in, not as nice (all relative it still is a very good sound) - this is the typical standard MM config. The Digital EQ versions sound slightly less airy than the 50pf/80k un-EQ'd version - but still have the wide soundstage, a lot more air, an open sound ... quite lovely really - almost sweet..... looking forward to comparing to my "sweetness" champion the Sony XL-MC104.

neobop

  • Full Member
  • Posts: 3448
  • BIRD LIVES
Re: Moving magnet phase shift
« Reply #78 on: 17 Jun 2011, 05:10 pm »
Hi David,
Hmm..... Some far reaching implications with your post - digital EQ and all that. I would speculate that you're not actually measuring the high frequency mechanical resonance (21-23KHz) with those carts. You must be measuring resultant resonance after being lowered by loading? I believe those carts have response well beyond that. I don't think that would be possible if primary HF res was that low. I know that in the past you speculated about response possibly being down 30dB with an AT spec. It just isn't so. A cart like  a 4-ch model or any of the older ones might not be flat to spec but will have usable response out to there, depending on load I would think.

Maybe that Sony could be used as a control? Is that a LO? I'm down to 2 MCs, one is a Sony MCXL-2. If you want to borrow it to test, you're welcome to. My Monster however isn't going anywhere. I believe the HF spec on the Sony is published. Your numbers just don't seem right. Another explanation maybe?
neo

dlaloum

  • Full Member
  • Posts: 710
Re: Moving magnet phase shift
« Reply #79 on: 17 Jun 2011, 11:02 pm »
Hi Neo

I begin my adjustments for a cartridge by measuring frequency response....
To do that I use either:
1) HFN Test record Pink Noise run at 45rpm (and adjusted EQ accordingly) - 50Hz to 27hz
2) Denon Frequency Response test 1000Hz to 50kHz
3) CBS STR151 RIAA test record sweep & Spot (40-27kHz - using 33rpm & 45rpm)

I'm also part way through developing some calibration data to correct for the HFN test records innacuracies... the Pink Noise track is inaccurate above 16kHz - as compared to CBS & Denon test records (which are in agreement with each other) and based on comparison of measured results with Miller Audio Research results.

As I recall the CD4 spec aimed at F/R being 15db down at 40kHz

The Sony is a HOMC - I also have a Benz/Empire MC1 (also HO)

I have not put the MC's on the TT in about 6 months, and since before I started measuring in detail....

Recording at 24/96 does limit my top end to 48kHz...

As an example of a cartridge with a low resonance but extended F/R consider the Grado Gold1 - there is a strong resonance at 12.5kHz - but F/R out to 50kHz... it is considered a good CD4 cartridge - limited primarily by its lack of a line contact stylus.

Bye for now

David

P.S. thanks for the offer of the MCXL-2.... but I feel I have sufficient known control points in that I have at least one cartridge that has also been measured by Miller Audio Research (AT440MLa)
Unfortunately they measure only 20-20k