Advice on SUT for Koetsu Rosewood - EAR MC-4, Slagle/EMIA, other

Indeed the 47K is there for MM inputs to properly load a MM cartridge. The 47K value is not a liability for an active MC stage but when the turns ratio of a SUT is considered things change. The 47K value reflects to the Cartridge as 47K/Turns ratio squared. The perfect example of where this becomes an issue is with the Altec 4722. This is a 30/200:40K mic transformer that found a cult-like following because it is a great match wired 1:18 (120:40K) for a denon 103. The interesting thing with this combo is you get essentially the same gain out of a 103 in both the 1:18 hookup and the 1:36 hookup. This happens because assuming 47K, the reflected load of the 1:18 is 145Ω which is fine for a 103 but moving to 1:36 nets a 36Ω load which is less than the 40Ω internal impedance of the 103 which results in a 6dB loss which roughly offsets the 6dB gain from doubling the turns ratio. If the 47K resistor were increased to say 200K suddenly the increase of gain by doubling the turns ratio would be returned however the transformer behavior would also change which may be better or worse. It is interesting to note that some people do prefer the sound of the 1:36 connection with the 103 which has two possible expanations. 1) the transformer has no ringing and is critically damped. 2) loading a cartridge with its internal impedance reduces IM distortion and increases high frequency tracking (Peter Moncrief IAR #5)

Since with the SUT connected the secondary of the transformer serves the purpose of the grid resistor, it can in theory be removed completely. The problem with this is if the SUT is external and the interconnect gets unplugged or has a fault the phono stage can immediately become unstable and wreak havoc with a system with the tweeters usually becoming a casualty. The better solution is to increase the 47K value to the largest value that still keeps the phono stage stable with an open input. This becomes a situation where it is wise to open a dialog with the manufacturer and then have a competent tech change the value retaining the original part to replace if the unit is to ever be resold. I should also point out that increasing the value from 47K serves no purpose if the SUT is relying on that value (and possibly an additional Zobel) for controlled response.

I install a 300K at the input of my LR phono and I believe shindo used 100K so no everyone adheres to that antiquated 47K value.

dave

Over a decade ago I made the below linked post and still stand by it 100% today.

https://forum.audiogon.com/discussions/thoughts-on-cartridge-loading-with-a-sut?highlight=intractaudiohttp://

Some basic takeaways I have found.
-Loading the secondary is the worst thing you can do to a transformer. With a known source  / load combo an optimal transformer can be designed.  
-47K is a silly value for the input of a phono stage when considering a SUT.  Selecting a turns ratio on the reflected load of that 47K is misguided.
-Turns ratio and cartridge load should have zero connection to each other when selecting a SUT.  Turns ratio should be set for appropriate gain (with careful attention to possible overload). Load resistor should be also as high as possible.  Any needed cartridge load should be placed directly at the cartridge.
-Zobel Networks are a Band-Aid for a source to load mismatch of a transformer.  They will make the test equipment happy but always leave the music feeling something is amiss.  
-high Q resonant behavior of a SUT at 30Khz is audible and objectionable yet Low Q resonant behavior above 100Khz seems audibly benign.

dave
 

Dave didn't believe me when I commented on AudioAsylum some years back that interwinding capacitive coupling is affected by loading, and thus to get flat frequency response loading is a pretty important aspect of transformer use.

I still don't believe you :-)

I think this was in reference to a transformer inside an ampex deck that showed different behavior than I had seen from other line level transformers.  This particular transformer showed a critically damped but rolled off top end and when loaded the midband gain was reduced without effecting the high frequency corner which in "extended" its bandwidth.   I still believe that designing for the specific situation is the best approach and only rely on the load to control things if it is the main purpose of the transformer.  Power delivery to a speaker necessitates a load, working in a 600Ω professional environment also necessitates a load.  A SUT or line level transformer driving a gate/grid of an active device is a different situation where any load or network required is a function of a compromised transformer design and not part of the transformers "job description"

I have seen the age old 'all cartridges are balanced' along with the 'all cartridges are current generators' and rather than bicker about a technical definition, I look at the primary as a two terminal generating device that inherits its mode of behavior from whatever you hook it up to.  I have found that when operating as balanced care must be taken with the cable layout to adhere to the balanced standard and simply attaching an XLR connector to an existing tonearm wire is hit or miss when it comes to hum pickup.  

dave

That's a bit of a surprise- your response on AudioAsylum lead me to believe otherwise. I can point you to a number of other examples should you need more proof.

I think we are missing each other on a few of these things.  I have no doubt that assuming enough inductance and a reasonably competent transformer design that for a given cartridge an acceptable network can be found to give flat response.  My contention is those networks put in place to account for a mismatch also do some sincere damage to the sound quality.

We agree that SUT's can never be "universal plug and play" and while you say issues can be fixed by loading, I contend while that is technically true, this imparts a rather large sonic compromise.  

While on this topic I will add that in a perfect world I would install the "ideally matched" SUT into the front end of the phono stage but then what happens when you want to compare a Lyra Atlas to a Van Den Hul Colibri?  In my world you would need a new SUT and in Ralph's world you would need to hope that there is enough inductance to support the VDH and then adjust the load accordingly. In either case with an internal SUT this becomes a difficult task which is why i advocate for using external devices as the best option for this situation.

Hey Tom,

Transformer "misbehavior" (ringing etc) is a function of source and load values.  With a known source (cartridge impedance) and load (47K+||100pf) the transformer can be wound such that it gives the desired behavior.  This is a completely different concept than loading the cartridge and much of the confusion in the topic comes from lumping everything together into an "ideal black box"

For your 5 ohm cartridge and a need for 20X gain, a 1:20 can be done to have a low Q resonance of 3-5dB above 100Khz loaded only by the test rig.  Adding the actual input impedance of 47K+ and 100pf for cable capacitance will damp this behavior down to an extended well behaved top end.    Lets assume you want the cart to se 50Ω and the input resistor is 300KΩ.  The turns ratio of 20 will give an impedance ratio of 400 so that 300K becomes 750Ω.  Placing a 54Ω resistor at the cartridge in parallel with the reflected 750Ω will net a 50Ω load with minimal effect on the transformer behavior.  From the transformer primary viewpoint, the 54Ω resistor will lower the 5Ω source impedance to 4.6Ω.   The change in source impedance from 4.6 to 5Ω  is well within the range of source impedances that will work for a given design.  It is when you start changing the source or load impedances by a factor of 2 or more from the "design center" that transformer behavior starts to dominate the sonic signature.  Take the peerless 4722 that works nicely with a denon 103 and drive it with a 2Ω SPU and that controlled top end suddenly turns into a 10+dB peak @ 22Khz.  Fix that 10dB peak with a 10K secondary load and you end up with syrup.

The important thing to understand here is that the load the cartridge sees and the turns ratio of the SUT do not need to have any relationship to each other and simply mathematically attaching them to each other as a ROT generally leads to inconsistent sonic results.  

The K&K link does not "back" what Ralph says.  Ralph states (correctly) that for any given source there is one load that will make the transformer act in the manner for which it was designed.  The K&K link assumes the transformer to be a perfect device that simply reflects an impedance and has no other impact on the sound.  It is my belief that SUT's should be designed to work optimally in an unloaded situation with 47K input resistor and ~100pf of cable capacitance being the worst case situation.  If the cartridge needs additional loading that should be placed at the primary of the SUT to load the cartridge directly.  I find secondary loading to make a cartridge behave under various situations to be a sonic band-aid.

Years ago I found it interesting that people who use active gain stages tend to find their MC cartridges insensitive to loading and those who use SUT's can get fanatical about it.  I then went through the process of playing with loads on both the primary and the secondary of various SUT's and found that loading the secondary had a much more profound effect on the sonics.  If the transformer were ideal then a 100Ω load should be a 100Ω load independent of whether it is at the cartridge directly or reflected through the transformer.  Since the differences were not subtle I came to the conclusion that the load on the secondary was altering the sound of the transformer as much or more than  it was the cartridge.  This goes a long way to explain the "black art" label attached to loading SUT's since loading the secondary changes both the sound of the SUT and the sound of the cartridge at the same time.  

dave