[*Scotty*]

Unfortunately the phase shift problem is real and it is one of the primary reasons that MCs exist.   
It cannot be effectively addressed in the preamp because each phono cartridge has a unique quantity of inductance that can only be nulled by exactly the same value.
  If you can download the pdf doc at this link on the vinylengine website it contains an article from Audio magazine circa 1983 by Kevin Byrne product manager for Ortofon. This article contains a very through discussion of the causes of phase shift in fixed coil cartridges and how audibile the phase shift is. Listening tests were conducted and the sonic effects of phase shift are described.
http://www.vinylengine.com/phpBB2/viewtopic.php?t=33679
  How tolerant one is of the sonic aberrations that are caused by the phase shift a fixed coil exhibits is probably a function of what you are used to hearing. Without a direct A/B/A comparison between a MM and a MC it wouldn't be hard to see how the sound of a reasonably flat response MM cartridge could sound acceptable in the absence of a MC reference with low phase shift. This is especially believable when an Audio Technica or Grado cartridge is involved. ATs and Grados have as stable a stereo image as a MC because the cantilever cannot be pulled in and out longitudinally by the frictional forces encountered when playing a record. When the cantilever can be longitudinally displaced when playing a record the sound-stage sounds as though you are listening through a veil of moving water.
  The image wavers and and instruments are not stably fixed in space and there is a general feeling of uncertainty about where anything is located in the sound-stage. Except for the aforementioned brands virtually all MM designs suffer from this problem because their suspension consists of a simple donut of rubber or elastomer on the cantilever at the pivot point or yoke location.
 At present I am unaware of any alternative to a custom approach to solving the MM phase-shift problem using a second cartridge body in order to have a precision matching set of inductors.
  This can be an expensive proposition if you want to use an AT 150 or a upper line Grado and get everything out of it you can. You can easily spend as much as a MC cartridge. When I first went this route back in the 90s you could buy an AT 440 ML for $99, a pair cost me a little over $200 with shipping. Now days a pair of Grado Greens or AT 120Es may be the best way to get a good phono cartridge and a body with a spare stylus for the future.
Scotty
 

[Otis]

*Scotty* 

ATs and Grados have as stable a stereo image as a MC because the cantilever cannot be pulled in and out longitudinally by the frictional forces encountered when playing a record. When the cantilever can be longitudinally displaced when playing a record the sound-stage sounds as though you are listening through a veil of moving water.
   Except for the aforementioned brands virtually all MM designs suffer from this problem because their suspension consists of a simple donut of rubber or elastomer on the cantilever at the pivot point or yoke location.

The above quote has me a tad puzzled. If you're talking about the internal tie wire attached to the back of the cantilever tube, then there are other brands that have it -- I'm thinking specifically of Ortofon, and Stanton/Pickering. I think some Era-IV Shures did too.

In any case are you ascribing the MC sound to a lack of distortion caused by the stylus wobbling in/out on a rubber donut?

Is there any chance you can post amplitude-matched sound samples of your AT440 setup?

Thanks, Otis

[neobop]

Unfortunately the phase shift problem is real and it is one of the primary reasons that MCs exist.   
It cannot be effectively addressed in the preamp because each phono cartridge has a unique quantity of inductance that can only be nulled by exactly the same value.
 

That brings up an interesting question, I've been meaning to look into. Does the resistance of a cartridge remain the same throughout the frequency range? I think it's usually specified at 1K. I'm not sure how that works with a generator. With a loudspeaker it changes with frequency and is plotted. They use programs to calculate phase between drivers and they use those plots to get more exact results. Driver parameters are loaded into the program.

Series resistance and inductance are used to calculate cart electrical resonance. The cable, shunt capacitance and load resistor are the rest of that circuit. It is a little hard for me to see this in isolation. With all the electronics downstream, I'd just think there would be more to it, like a real filter. A 24dB/octave low pass has an inductor and a cap going to ground. Then it has another set of the same configuration. That changes the phase from 180 to 360.

Think I'll play a record before I get too tired. I found an old Henry Threadgill LP somewhere and it's already cleaned. That should get my mind somewhere else. I better use the Virtuoso though, the AT-95SA has too much resistance.
 
neo

BTW, Thanks Otis for that file.

[*Scotty*]

Quoted from Otis at 11:22 PM

In any case are you ascribing the MC sound to a lack of distortion caused by the stylus wobbling in/out on a rubber donut?

  No,there are certainly a large number of mechanical differences between MCs and MMs other than the lack of of a tieback wire on most MM designs which contribute to a sonic advantage for MCs.
  If there IS control of the cantilevers longitudinal freedom of motion, than you have a much better chance of an impressive outcome from canceling out the inductance the cartridge has.

Is there any chance you can post amplitude-matched sound samples of your AT440 setup?

I am not sure what information you are requesting in the above question.
  Scotty

[dlaloum]

It is not so easy...

Due to having picked up a couple of shop lots of stylus less cartridges, there are some cartridges which I have up to 6 or 7 examples of.

The inductive variance between the channels is SUBSTANTIAL.

I have gone through and measured both DC resistance and Inductance for about 90 cartridges...

Many of them have substantial variances between left and right channel, let alone between different cartridges!!

For example - a series of Technics EPC P24 cartridges:

Inductance (mH)   Resistance (Ohm
  338     344          571     577
  333     342          578     573
  334     346          602     574
  357     327          574     589

Note that there is up to an 8.8% variance in Inductance between channels let across a series of supposedly identical cartridges.

In trying to find the best inductive match for my Shure 1000e - I settled on an AT SLT96e....

1000e    520/521mH
SLT96e  522/521mH

I am not sure that having a perfect inductive match is quite so critical  - Phase remains in alignment as long as you don't hit a resonance - after that all bets are off and phase change varies based on cartridge and electrical design (as shown in the ortofon article)

You can use an inductor (2nd cartridge) to shift or remove the resonant peaks - I suggest you take a look at the model developed by LuckyDog in the thread I linked earlier )You plug in cartridge, and system electrical values - capacitance, resistive loading etc... and it plots a theoretical F/R graph... very handy). Using this technique does NOT require an identical cartridge - it does however require an inductor, and used cartridges are among the cheapest available relatively matched inductors....

By the way - for those in the US the AT440MLa is still an incredible bargain at $99 (from Amazon) - the rest of the world pays 50% extra.

There are tieback wires in a number of styli designs - specifically: The original Shure V15V & V15IV apparently have tieback wires. The Jico SAS styli all have tieback wires - I believe this technique is more widespread than believed among the better Styli.

Once I have finished working through the testing of the current electrical model (and we have adjusted the model accordingly) - I intend to post samples with different cartridges showing the effect of different loadings/inductors etc....

I am doing this as much to test it for my own satisfaction as for anyone else, but everyone might as well enjoy the fruits of that labour.

But one thing that is clear already - you can achieve similar results without a second inductor - all that is required is VERY LOW CAPACITANCE  (50pf is good!) and custom adjustment of Resistive loading.
With that you can maintain Linear frequency response with almost any MM/MI cartridge right through the audio range (frequently out to 30kHz or more) - and linear F/R in an "unmessed with" circuit will also provide linear phase response.
As the inductance of your cartridge drops you will find yourself needing less dramatically low capacitance to achieve the result.
A Cartridge with inductance under 400mH should be fine with capacitance of 100pf

bye for now

David

[neobop]

But one thing that is clear already - you can achieve similar results without a second inductor - all that is required is VERY LOW CAPACITANCE  (50pf is good!) and custom adjustment of Resistive loading.
With that you can maintain Linear frequency response with almost any MM/MI cartridge right through the audio range (frequently out to 30kHz or more) - and linear F/R in an "unmessed with" circuit will also provide linear phase response.
As the inductance of your cartridge drops you will find yourself needing less dramatically low capacitance to achieve the result.
A Cartridge with inductance under 400mH should be fine with capacitance of 100pf

Thanks David,
That was exactly the gist of the Cartridge Loading Explained thread at VE. Keep capacitance as low as possible (arm cables + preamp) and adjust resistance loading for flat response. The effect of shunt capacitance is said to lower the actual measured high frequency resonance as well as the theoretical value. If the actual resonance is in the audible band it will usually make the cart sound brighter, while rolling off the extreme high end. If there is an associated phase problem with the electrical res peak, then it would only be made worse with higher capacitance, going further down into the midrange. Some people removed their shunt capacitors with no neg preamp  effects, but YMMV.

Another parameter used for the calculations is cart resistance or impedance, which are not the same thing. Although the terms are often used interchangeably, there can be a dramatic difference between the 2 values. For example, I have a spec sheet for the AT-95 series. They all have a DC resistance of 410 ohms. The coil impedance is 2800 ohms @ 1KHz. Wow, that's a biggie. I guess that answers my previous question about the frequency of resistance/impedance. A meter measures DC resistance. Computing Impedance is a more complex. 

That brings up an interesting aspect of the calculations for the electrical model. Is it correct to use DC resistance for computing the electrical peak? When you compute the value of an inductor or capacitor in a passive crossover, with its associated frequency/phase implications, impedance is used, not DC resistance. Maybe it's different with a generator, but I really can't see why. If you're looking at shunt R/C in isolation and not part of the circuit topology, then it would seem to be passive. I haven't read the Ortofon article yet, maybe that's covered, and maybe using impedance would only make things seem worse. Impedance is usually a higher value. Maybe the whole thing is a gross oversimplification. The Hagerman model proves to be of limited value when actually trying to find the right load values.

BTW David, The AT-95:
Static cu = 20
Dynamic @100Hz = 6.5

neo

[dlaloum]

Yes resistance vs impedance gets tricky - and having the tools to measure this stuff easily is another issue...

Take a look at the Near Zero Capacitance thread on VE - LD is working through some of the issues involved in non-linearity of transformers at low signal levels (eg cartridge with lower leve signals)... it may or may not end up being incorporate into the model

The model as it stands is quite usefull in selecting an appropriate set of loading values.

Impedance vs resistance is not a major issue for standard MM / MC's simply because the load resistance is a whole order of magnitude larger....

And I would suggest that at this stage of development - the model can only give you an indication of an appropriate loading setup - it is after all an electrical model - and does not take into account some of the electro magnetic non linearities, but more importantly it cannot model the mechanical behaviour (cantilever resonance, stylus radius limitations etc...)

So you use the model to get an indication of your desired resistive load - say 65k ohm - then play (measure) with 62k  and 68k ohm (and perhaps further afield)

The variation between impedance and resistance becomes largely irrelevant.
I find resistance most useful in trying to identify unlabelled cartridges...

I rekon that a good MM /MI (exotic cantilever, Micro Ridge style stylus) run at 50pf capacitance and properly loaded resistively gives any MC a real run for its money, and does it at half the price or less....
In that kind of environment, some of the other advantages of MM (eg: high compliance) can start to make a difference....
I do not have TOTL MC's to compare against mind you. (A future project...)

bye for now

David

[BaMorin]

Another parameter used for the calculations is cart resistance or impedance, which are not the same thing. Although the terms are often used interchangeably, there can be a dramatic difference between the 2 values. For example, I have a spec sheet for the AT-95 series. They all have a DC resistance of 410 ohms. The coil impedance is 2800 ohms @ 1KHz. Wow, that's a biggie. I guess that answers my previous question about the frequency of resistance/impedance. A meter measures DC resistance. Computing Impedance is a more complex. 

BTW David, The AT-95:
Static cu = 20
Dynamic @100Hz = 6.5

neo

Impedance rises as frequency rises as a function of inductance. The higher the inductance of the coils the more rapid the rise in impedance vs frequency.  Which BTW is the only electrical advantage of "most" MC carts vs MM/MI iron carts.

[BaMorin]

http://lgbtlineage.net/ems/MUSIC%20167/impedance.pdf


This might explain some of what you might be looking for Neo on phase shift as a function of inductance and capacitance.

[neobop]

http://lgbtlineage.net/ems/MUSIC%20167/impedance.pdf


This might explain some of what you might be looking for Neo on phase shift as a function of inductance and capacitance.

Thanks Marc, it's next on my list.
I just looked over the Ortofon article. Plots are from measured phase shift vs amplitude response. If anything, I'd have to say that it validates my suspicion about MM vs MC phase shift. No data was given about cart specifics, apparently that's in an AES article they published, but the MC phase performance didn't look all that much better. The only sound quality attributed to reduced phase shift was imaging. Nothing was said about muddy sound or lost detail that I recall. I'll have to read it again, but it seems that the MCs with the flattest frequency response had the worst imaging. Mechanical damping was used so electrical resonance didn't come into play.

This article was dummied down for general consumption, but the graphs tell a lot about actual phase shift. I'll have to look over the plots again and compare to the plots made from the Hagerman model.

Thanks again Otis,
neo

[neobop]

Assuming this Ortofon paper is based on actual testing, like it says, there's no doubt that mechanical resonance effects phase behavior. The same MC cartridge exhibited different phase shift when nothing but cantilever damping was changed. We don't have enough information to draw any conclusions about electrical resonance vs mechanical resonance. I suspect, just like in cartridge loading considerations, the electrical resonant frequency is of limited value. The actual high frequency resonance, as opposed to the calculated electrical HFR, is what determines the overall FR. Too much shunt capacitance with a high inductance cart will lower the mechanical resonance. When that mechanical resonance goes into the audible band, is when problems arise. In some cases it could even be seen as a solution if used to augment a deficiency in the treble region. Why else were caps supplied with some carts?

Ironically, the MC cart used for this test imaged best with no damping, however it was said to be unlistenable due to ear bleed type FR. The more the cart was damped, the flatter the response but the imaging suffered more.

I think Scotty is right about eliminating body resonances and securing the cantilever assembly on a MM/MI cart w/removable stylus. The 440 can also be problematic with a high value of shunt capacitance. I have no experience with cancelling the inductance. Maybe you or David, could tell us a little more about your impressions of sound quality with vs without inductance cancellation or reduction.

At this point, I have to say that to a large extent, I agree with Wayner. When I get everything set up just right with my preferred MM carts, it sounds pretty damn good. Too good actually for some of my recordings. I'm not saying better than my MCs. I have a Charlie Parker record, recorded in the '40s. This is a live date. It was recorded on a portable tape machine off a bathroom or hallway monitor. Sometimes a little tone control phase shift comes in handy. 
neo

[*Scotty*]

 What can't be seen in the photos I posted is the bead of super glue I ran at the junction between the plastic mounting plate and the metal cartridge body. While I didn't do any before and after testing for this specific mod I was reasonably sure it couldn't hurt. What I did notice is that the stock AT 440 was a little bright in the upper midrange and it didn't have the correctness of timbre or imaging that my CD playback had.
  After the mods to the cartridge and nulling out the inductance problem,the playback from vinyl sounded very similar to CD replay in the parameters of timbre,image size and high frequency clarity. It seemed to me at the time that it was unreasonable that the sound of vinyl should depart that far from a another medium which doesn't have frequency response problems or for the most part high frequency phase shift problems.
  I figured if I tried to optimize the system for the nonlinearities present in the vinyl replay then the CD playback would be very screwed up. When the phase problem was eliminated from vinyl playback it was like having your cake and eating it too. I had the same freedom of coloration that I have from CDs and no digitius which is present at times in some digital sources.
  I am of the opinion that when vinyl and digital replay are both free of gross nonlinearities they should sound very similar in many ways. I don't think that you should have to optimize your system for one medium to the detriment of the other.
Scotty

[neobop]

I'm sure the body and stylus mods made a significant difference. One of the biggest criticism of MMs is that the practical aspect of having a replaceable stylus, is also a compromise in performance. I never used superglue, but I secure the plastic stylus holder with tack usually. In the case of an AT, the holder can removed altogether. That just leaves the plug, which is the rectangular part that fits into the body. That seems to sound better. Body Vibrations aren't transmitted through the holder. Your body is damped with the 3M stuff so this is accomplished in a different way, perhaps better. There's no chance of the cantilever moving when it's glued in. Although it could be glued in without the plastic wing. There's a thread a few pages back about the AT-95 and the Clearaudio MMs. AT is the OEM for these. The plugs are the same and styli can be swapped if the holder is cut away on the 95 stylus. I even potted, and mounted an aluminum top plate on the 95.

All this does make it harder to sort out exactly what's doing what. The 95 and 440 have a different character, frequency response, etc, so different sounding improvements could be realized with the same type of mod.









neo

[neobop]

Assuming this Ortofon paper is based on actual testing, like it says, there's no doubt that mechanical resonance effects phase behavior. The same MC cartridge exhibited different phase shift when nothing but cantilever damping was changed. We don't have enough information to draw any conclusions about electrical resonance vs mechanical resonance.

I take that back. Further examination of the actual FR response vs phase reveals some interesting conclusions, that I think are a safe bet. The situations where electrical parameters are a factor appear to be with high inductance (MM) carts. The 5 unnamed carts that were tested had phase anomalies that extend into the audible band. Without specific info on the carts in question, it would be imprudent to guess about much more.

There were also 4 plots of the same MC-200 LOMC cart, with varying amounts of mechanical damping. The phase reversal peak coincides with the response peak around 29KHz with the undamped version. The electrical model would put that somewhere, orders of magnitude higher. It appears that the electrical model just doesn't come into play. This would make sense because the inductance is typically so low that it's just not a factor. What is a factor however, is the mechanics. The undamped version had an extreme phase angle associated with the peak at 29K. The phase shift did not appear to go down below 20K. The cart was said to be unlistenable though, in the undamped version, due to the FR. As damping was added to the suspension, the FR flattened out. The phase peak went higher in frequency, and the phase angle also flattened out. Because the phase angle also was flatter, it now extended into the audible band. When the cart had FR of + 2 to 4db @ 20K, the phase shift extended down to 7 or 8K.
neo

[neobop]

Some of us here at VC participated in a VE thread called Cartridge Loading Explained. Although our posts are probably no longer there, most of us gained insight about loading a cart. The OP, CarlosFM, used a computer program, Spice I believe, to model electrical parameters and predict FR. This was more accurate than using Hagerman's electrical resonance calculator. Whatever, it turns out that it's actual FR that's important for loading, and electrical resonance is meaningless. Values of inductance and capacitance give an idea of what's going on and how to deal with it. It's the mechanical or actual resonance that is affected by inductance/capacitance. That's what is important. The way it can be affected is by lowering the frequency of the mechanical HF res.

I know all this sounds unnecessarily complicated for many. Bear with me, cause the practical aspect is a whole lot simpler. As a general rule, keep the capacitance load on HO carts as low as possible. If you're using a HOMC or a HO Grado, it isn't as critical. So, long cables from your table to your phono pre are not recommended. This will add to the shunt capacitance. If there is a switch or user selectable capacitance on the pre, use the lowest setting. If you're using a very high inductance cart, you might have to change the preamp load resistance to a higher value, for better performance.

Previously, Felix said that he uses 62K to load a stock M97. Why? That load will address the dip in the FR in the lower treble and yield a flatter response. Using a higher load than 62K will make the response curve jagged and the upper treble too bright. This was confirmed by actual measurements by a VE poster named Werner. He's a reviewer at TNT an an EE or something. BTW, he also said that electrical resonance is worthless for loading calculations.

The M97 is a perfect example of a high inductance cart with somewhat dull response. It's not only current, but an upgraded stylus/cantilever, the Jico SAS, is available for it. Users of the M97/SAS stylus report a change in the character of the sound. Loaded at 47K it no longer has the treble dip. Why?
It's the mechanical aspects that are the overwhelmingly biggest factor in determining frequency response.
neo

[neobop]

Based on the Ortofon evidence, I think it's indisputable that the mechanical HF resonance is what determines phase shift with the MC200. In the commentary it says that typical undamped MCs have amplitude response that begins to rise around 5KHz, goes to +8dB @ 20K, and peaks at +15 to 18dB at 25 to 28K (mech Res F). I suspect, that is the case with virtually all MCs. Resonant frequency can vary depending on moving mass and stiffness, but there are real world limits. A mechanical res F @ 28K is vibrating 28,000 times per sec, as Wayner reminded us previously. The amount of phase shift at res F is severe, 180 degrees. It's only has 10 to 15 degrees at 20K (undamped). 

There are plots for 5 unnamed MMs that suggest a similar scenario regarding phase shift. Without specific info on the carts, we can only guess about the effect of inductance. The commentary said that mech res F was lower due to greater moving mass, and that inductance/capacitance was a factor. The MMs typically employ more damping which increased phase deviation. But, looking at the plots, there is no 180 degree phase reversal within the audible band as suggested by plotting the electrical resonance. The peak, which would correspond to 180 degrees, is off the chart, either due to frequency and/or amplitude limitations. Note: That was the case with most MC plots as well.

The phase vs frequency of the MMs showed greater deviation than the MC200. It did extend down to 1 or 2K on some, where the MC had a worst case shift of + 20 degrees at 10K. Listening tests suggested that imaging was the only thing effected, FWIW.
neo

[dlaloum]

On the subject of the audibility of phase issues...

There is a very good article by Robert Everest Greene @ http://www.regonaudio.com/HighRomanticism.html

It is focused on classical and more specifically Romantic music  (of which I am a fan)...

an extract from that article:
Here we encounter one of the great myths of audio, that transient behavior is the best indicator of musical quality. Ironically, it is quite true mathematically that perfect reproduction of impulses requires perfect reproduction of everything else (in a zero-distortion system, free of non-linearities). But, in listening terms, the ear/brain is more sensitive to sustained sonic errors.

There is in fact evidence accumulating that this is true even of phase behavior, not just frequency response. In visual, graphical, mathematical terms, phase non-linearity has dramatic effects on the shapes of transients. But these effects have very limited audibility. What is more audibly obvious by far, according to recent evidence, is the effect of phase non-linearity on complex sustained music, e.g., choral singing (c£, the research of the audio group at Essex University).

I have been doing some searches for any information on the research at Essex University...so far without much luck.

Phase issues might effectively be inaudible for fans of rock music, and extremely audible for fans of Gregorian Chant....

Bye for now

David

[neobop]

David,
Very interesting article, thanks.

As I was looking over the article I couldn't help but think of tonality in general. Since the classical period, halls have gotten bigger and tuning has gotten sharper. Middle A at one time used to be 400Hz. Now it's 440.
neo

[neobop]

The link for the Ortonfon article, once again:
http://www.vinylengine.com/phpBB2/viewtopic.php?t=33679

If anyone is an AES member, the paper was presented at the 71st AES convention, reprint #1866.

Looking at the actual phase nonlinearities of MC carts, they are caused by damping the moving system or suspension. Electrical parameters don't come into play (for phase) because of the low inductance. Some LOMCs have as little as 11 or 12 turns of wire, as opposed to 2000 or 3000 turns for some HO carts. If the mechanical resonant frequency is high enough, there will be no phase nonlinearities in the audible band for an undamped MC. However, damping is necessary to correct the frequency response. When the MC200 is damped for FR of + 1.8dB @ 20K, the phase deviation begins at 7 or 8KHz. Most of the phase deviation will be in the overtone region. The primary frequency of a cymbal usually is around 12 or 13KHz. Anything above that is in the hearing test range, but is said to effect our perception of tonality.

It would be interesting to have specific info about the MMs. In the phase vs amplitude plots, it appears as if mechanical resonance is also the primary cause of phase anomalies. It is suggested in the commentary that inductance/capacitance is also a factor. They exhibited phase shift starting around 1K to 4K. There were 5 examples of different carts. They all had  greater phase shift than the MC. It was suggested that the primary cause was the necessity to employ a greater amount of damping.

Damping the suspension seems to be the primary cause of phase nonlinearities in all carts. If phase is linear, it will coincide with amplitude response. Using the lowest amount of shunt capacitance possible with high inductance carts, will help eliminate electrical parameters from contributing to phase shift. It would be hard to go much further with the information. If a cart has a higher mechanical resonant frequency, it should exhibit better phase linearity. The more damping used, the greater the phase deviation.
neo
 

[dlaloum]

Interesting new (old) stuff...

an AES paper presented by some of the Engineers at JVC in the 70's

http://www.vinylengine.com/phpBB2/viewtopic.php?t=35852

They tested 150 cartridges.... and found the ones designed for Quad to be very phase as well as amplitude linear....

They show some sample phase graphs of "typical" MM, MI, MC and "Condenser" cartridges.

An interesting read.... just wish we had access to the data on the 150 cartridges - I bet a few of them are our current highly desirable classics, and that info has never been published.

bye for now

David