[analog97]

I was asked to report on my findings having built that "T" loading device as suggested by Amandarae.  I have a Rega P3-2000/Denon 103/Denon 1:10 SUT/Cornet2/Clarinet combo.  I soldered 2, 13K 1/4 watt resistors as detailed by Amandarae.  This, according to JH's "calculator"  results in 102 ohms. I had been listening to the 470 ohm loading for over 8 months.   Here is what I heard, of course, subjective reporting only.....no "testing equipment" used.

1.  102 ohms seems to have reduced "surface noise" significantly.  The noise floor seems to have been lowered.
2.  102 ohms seems to have produced an interesting change in the high treble.  This is hard to describe, but the highs now seem "cleaner and clearer".......e.g. cymbals are not exaggerated and somehow seem to be heard with less fatigue on the ears.  I find I definitely prefer this "new" sound, not that the "old" was bad.   
3.  it seems that my amp is now working a tad harder to produce the same volume level.  This seems pretty strange to me, but I'm convinced of this.  This is subtle and not a problem as I have plenty of power.
4.  most of all, the cumulative effect of all of these changes was to deepen my soundstage and just somehow make things sound more even and more musical (words can be pretty bad, eh?)

I am convinced of these changes, but have absolutely no scientific...i.e. electrical understanding of why these changes occurred.  My reason for doing this was that of curiosity and I'm glad I did.  This was one rewarding "tweak" and I thank Amandarae and others for their help, pictures and advice.

[amandarae]

I was asked to report on my findings having built that "T" loading device as suggested by Amandarae.  I have a Rega P3-2000/Denon 103/Denon 1:10 SUT/Cornet2/Clarinet combo.  I soldered 2, 13K 1/4 watt resistors as detailed by Amandarae.  This, according to JH's "calculator"  results in 102 ohms. I had been listening to the 470 ohm loading for over 8 months.   Here is what I heard, of course, subjective reporting only.....no "testing equipment" used.

1.  102 ohms seems to have reduced "surface noise" significantly.  The noise floor seems to have been lowered.
2.  102 ohms seems to have produced an interesting change in the high treble.  This is hard to describe, but the highs now seem "cleaner and clearer".......e.g. cymbals are not exaggerated and somehow seem to be heard with less fatigue on the ears.  I find I definitely prefer this "new" sound, not that the "old" was bad.   
3.  it seems that my amp is now working a tad harder to produce the same volume level.  This seems pretty strange to me, but I'm convinced of this.  This is subtle and not a problem as I have plenty of power.
4.  most of all, the cumulative effect of all of these changes was to deepen my soundstage and just somehow make things sound more even and more musical (words can be pretty bad, eh?)

I am convinced of these changes, but have absolutely no scientific...i.e. electrical understanding of why these changes occurred.  My reason for doing this was that of curiosity and I'm glad I did.  This was one rewarding "tweak" and I thank Amandarae and others for their help, pictures and advice.

You are welcome analog97!

Glad to hear that it worked well for you.  

3.  it seems that my amp is now working a tad harder to produce the same volume level.  This seems pretty strange to me, but I'm convinced of this.  This is subtle and not a problem as I have plenty of power

It feels that way because it should!  When you load a cartridge, the resulting signal to the preamp wil be reduced.  Well, it is not as simple as that but trust me on this.  In fact if you load the cartridge at its specified coil resistance, the output will be halve.   Do not worry, it's normal.

If you want to kill time in understanding why loading the cartridge changes its "sound", then read about cartridge damping.  In a nut shell, what it does is to reduce the unwanted resonance in the elecrtical system (cartridge).

regards

[hagtech]

In a nut shell, what it does is to reduce the unwanted resonance in the elecrtical system

Yes, to a first order, and that is what I talk about in my www.hagtech.com/loading.html paper.  It calculates a good starting point for your listening experiments.

But there is more to the story.  Except this is where I am theorizing and venturing into the not well understood aspects.  One of the major changes in loading is the fact that you are now raising the current output level of the cartrdige.  So now there is more magnetic flux involved in the motor structure.  It's no longer just presenting a voltage, but pushing power down the line.  Similarly, this is another reason why it might be better to put the loading resistance on the secondary of a step-up versus the primary; that it shifts the operating point of the tranny to higher current.  The power transfer is through the device, not shared with the primary.  A lower loading resistance value will also increase the back EMF of the motor (I think). 

Basically, there are three modes to operate.  Voltage, current, and power.  Some preamps run in total current mode, where the input impedance of the phonostage is zero (Norton amp).  With a typically good sounding MC load, we're in power mode, with both voltage and current, but amplifying only the resulting voltage. 

With load current, the magnetic circuit of the cartridge becomes more significant.  Perhaps we are also damping it?  Maybe we are just shifting it to a better sounding operating region.  Just for a reference point, optimal power transfer occurs when load and source impedances match.  That is, load the cartridge equal to its internal coil resistance.  Or maybe it's more complex than this, maybe the match has to occur not at dc, but at the broadband impedance seen in the midrange.  Anyway, this is another good place to start with listening experiments.

jh

[analog97]

"Just for a reference point, optimal power transfer occurs when load and source impedances match.  That is, load the cartridge equal to its internal coil resistance.".....per JH.....Can someone tell me what the "internal coil resistance" value of a regular Denon 103 is?  I might try that value as well.  I had thought that was the 100 ohm value.

[amandarae]

"Just for a reference point, optimal power transfer occurs when load and source impedances match.  That is, load the cartridge equal to its internal coil resistance.".....per JH.....Can someone tell me what the "internal coil resistance" value of a regular Denon 103 is?  I might try that value as well.  I had thought that was the 100 ohm value.

40 ohms! 

[amandarae]

Now, a caution........


If you load the 103 at 40 ohms with 1:10 gain, your output voltage after the transformer is 1.5 mV (0.5 x 10 x 0.3mV).   The Cornet 2 has 44 dB gain.  So that is 158 times the input which comes up to be 237 mV (158 x 1.5 mV)!   If your amp is rated at 1V input for rated power, then your minimum linestage gain (from main preamp) should be at least 13 dB for full power.

cheers 

[robertwstephens]

I had the same effect.  Thanks for the report!  Peace.  Robert

[hagtech]

then your minimum linestage gain (from main preamp) should be at least 13 dB for full power

True, for a 1kHz sine signal carved into the vinyl at 4cm/s.  This is just a reference tone upon which to specify components.  I believe it has been found that LPs often produce peak levels 16dB or so higher than this.  So it all depends on how far into clipping you want to drive your amplifier.

Seriously, most systems have too much gain.  On the other hand, it's psychologically comforting to know you can crank it up past what is needed.  Most 'philes like to listen with the volume at less than 12 o'clock.  That's fine.  As long as you have enough gain in the first place.  Do what gives you the most enjoyment. 

jh

[andrew]

Hi,

  How does this 'T' device work? I'd be interested in a link if you have one.

  For my part I found that loading the primary with an equivalent resistance was profoundly better than loading the secondary, especially with high Zout carts like the 103 family, less so with something like a Kontrapunkt.

  I suspect, in  a perfect world with a perfect SUT then there shouldn't be a difference, of course, we can't build a perfect SUT and we live in an real world that has imperfections, core losses, parasitics and lots of other nasties.

  Seriously, try loading the primary with 100R, its a simple change and easily reversed if you dislike it.

    cheers,

-- Andrew

[analog97]

Andrew,

To my way of thinking "primary" loading means the resistors are between the turntable leads and the SUT.  "secondary" loading is after the SUT output and infront of the Pre-amp.  If this is correct, I will try it.  Also, you will see the "T" loading device suggested by Amandarae on an earlier thread.  That's what I built.  I was impressed with 100 ohms this way, but have not yet tried the 40 ohm "ideal" load.

[amandarae]

Hi,

  How does this 'T' device work? I'd be interested in a link if you have one.

  For my part I found that loading the primary with an equivalent resistance was profoundly better than loading the secondary, especially with high Zout carts like the 103 family, less so with something like a Kontrapunkt.

  I suspect, in  a perfect world with a perfect SUT then there shouldn't be a difference, of course, we can't build a perfect SUT and we live in an real world that has imperfections, core losses, parasitics and lots of other nasties.

  Seriously, try loading the primary with 100R, its a simple change and easily reversed if you dislike it.

    cheers,

-- Andrew

Hmmmm, IMHO it is totally different!

On our application  with the SUT,  increasing the current through it makes it work better because it increases the electromagnetic induction(higher current produces a stronger magnetic field).  So, for us loading the secondary, we want to increase the current in the primary which improves the way a transformer functions.   Loading the primary defeats this purpose because it reduces current into the primary.  When you shunt a resistor to the primary,  much of the energy generated by the cartridge coil is dissipated as heat in the loading resistor before it reach the transformer primary.  This reduces the magnetisation current(AC) available and forcing the transformer to operate at a lower point in its curve.

I doubt it if there's a big difference in "sound" if one elects to load the cartridge in any of the two method.  One thing for sure, if you load the primary, the output voltage of the cart will be significantly less compared to loading the secondary.

I hope someone can correct me here if I am wrong......

[andrew]

Hi,

  Yes the 100R resistor is in parallel with the primary coil, across the SUT phono inputs (or the TT cable if you wire direct).

   I can't fault your logic about lost current in the primary to the shunted load; that makes total sense. However, I don't think in my case that was my problem, or put another way, while I agree with you, I still find I prefer loading the primary.

   I have a theory, and I'm  not convinced by this either, that the combination of a large resistor on the primary and losses in the core, to reactance, parasitic capacitance and winding resistance etc were causing a low pass filter. The filters 3db point was quite high up the audio band but was enough to cause loss of high freq and more importantly slew rate limiting; it made my system sound slow and ponderous.

  Interestingly, my conversion was a result of playing with a $10 head amp using 1N4001/1N4403 by a chap called Marshall Leach which, despite its faults, showed up what the SUTs were doing wrong.

  I am simply offering this as an alternative to the received wisdom of loading the secondary, it's a two resitsor change and easily reversed if you find you don't like it. After all, changing load resistors has to be one of the easiest tweaks in audio.

  I'd be very interested in your findings on what you actually prefer, say between a 100R on the primary and 10K on the secondary, assuming 1:10 SUT. Especially, if having tried both you still prefer secondary loading. There might be other system variables, your cart, phono etc to consider that change the parameters and might provide a clue to what's really going on. I remain, in my heart of hearts to be convinced by SUTs as a solution for MC, are they just the best set of compromises?

  Jim, there's perhaps a 1/2 kit in this for a solid state headamp.....just a thought?

    cheers,

-- Andrew

[analog97]

The gent Marshall Leach is an engineer and below is the link to what appears to be a project almost within my reach.  Has anyone a picture of this completed circuit?  That would be a great help to a near novice.  Also, has anyone compared this "active" device to a SUT?  I also wonder if JH could do a "super-active step-up", instead of $10......maybe $40-100???  I wonder what that could buy in the hands of JH?

http://users.ece.gatech.edu/~mleach/headamp/cbamp2.gif

[hagtech]

That's a very clever little circuit.  I've seen it around before.  Thinking there is a very similar one from 30 years ago.

I'd consider removing C2 and C3.  I wouldn't use C8 either.  R1 doesn't do anything.  Theoretically, it should work, but may not provide an optimum loading on cartridge.  And the 9V battery floats with the signal, so any body capacitance to chassis (same as C8) will roll off bandwidth.  It can also easily inject noise.  But hey, there are ways to overcome this.  Engineering is the art of compromise.  I just mention it so you are aware before you start to build it.  Similarly, the switch must be shielded.

This also isn't really an amplifier.  It is an impedance transformer (transimpedance).  Input is very low resistance, so it looks almost like a short to the cartridge, hence it works on input current, not voltage.  Basically, it takes that current and dumps it into R5 (47k?).  Output is then V=I*R.  If the cartridge is not very low impedance, then you won't end up with much gain from this circuit. 

Curious to hear from anyone who has actually built it.

jh

[andrew]

Hi,

I built the common emitter version, works very well for the money. Showed me what me SUTs were getting wrong, made them sound slow and bandwidth limited by primary loading cured that. As a breadboarded circuit, no box, it picks up EMF and isn't too good on overload for some reason, any clicks and pops on the surface are more audible than the SUTs, otherwise pretty good for $10. I have a layout somewhere that might help you build one on stripboard/veroboard. I'll post it when I get a moment.

http://users.ece.gatech.edu/~mleach/headamp/

Here's the full URL. I'd be interested what Jim makes of the common emitter circuit. Again C8 isn't probably needed.

   cheers,

-- Andrew

[andrew]

If you go here

http://www.world-designs.co.uk/forum/showthread.php?t=2113

You'll find a layout for the verobaord, get the second one as the first had a mistake. Just follow the thread and you'll be OK.

   cheers,

-- Andrew

[analog97]

Seems to me there is a big potential market here, if someone can improve the Marshall Leach approach.  If $10 in parts can achieve the specs reported by Dr Leach, why can't BETTER parts, BETTER design, MORE parts, real-world engineering and testing take $100 and make something that could compete with a $1,000 SUT? 

Hey, JH, you could retire in Hawaii!!  Then, when you are rich, you can design the 100 watt tube amp your customers want.   

[amandarae]

Hi,

  Yes the 100R resistor is in parallel with the primary coil, across the SUT phono inputs (or the TT cable if you wire direct).

   I can't fault your logic about lost current in the primary to the shunted load; that makes total sense. However, I don't think in my case that was my problem, or put another way, while I agree with you, I still find I prefer loading the primary.

   I have a theory, and I'm  not convinced by this either, that the combination of a large resistor on the primary and losses in the core, to reactance, parasitic capacitance and winding resistance etc were causing a low pass filter. The filters 3db point was quite high up the audio band but was enough to cause loss of high freq and more importantly slew rate limiting; it made my system sound slow and ponderous.

  Interestingly, my conversion was a result of playing with a $10 head amp using 1N4001/1N4403 by a chap called Marshall Leach which, despite its faults, showed up what the SUTs were doing wrong.

  I am simply offering this as an alternative to the received wisdom of loading the secondary, it's a two resitsor change and easily reversed if you find you don't like it. After all, changing load resistors has to be one of the easiest tweaks in audio.

  I'd be very interested in your findings on what you actually prefer, say between a 100R on the primary and 10K on the secondary, assuming 1:10 SUT. Especially, if having tried both you still prefer secondary loading. There might be other system variables, your cart, phono etc to consider that change the parameters and might provide a clue to what's really going on. I remain, in my heart of hearts to be convinced by SUTs as a solution for MC, are they just the best set of compromises?

  Jim, there's perhaps a 1/2 kit in this for a solid state headamp.....just a thought?

    cheers,

-- Andrew

Hey Andrew,


I will try it when I have time.  Thanks for the explanation and will report back about my findings.  I recognized your name from the DIY forums.  I am a lurker there as well and currently following the Greg Ball SKA amps.

Happy Holidays!

[hagtech]

I also wonder if JH could do a "super-active step-up"

Huh, waddaya know.  I actually did a headamp design years ago (the PICCOLO project).  It was fully differential and used the 2SK389 JFET pairs.  So I played with it a little and changed to single-ended.  Not sure about the resistor values, but I trust they're not too bad since I must have calculated them at some point.  I think the nominal gain was supposed to be 20dB. 



This circuit allows you to set loading (a few ohms up to 1M) and gain (input to JFET is very high impedance).  Output impedance should be under 500 ohms.  Enough to drive a MM phono input.  Circuit has a few tricks.  C1 reduces input capacitance, C2 forms a rumble filter.  Output dc voltage is very close to 0V.  The differential input allows you to wire up single-point ground for low hum.  Each JFET should probably have a small gate resistor.

Anyway, maybe I'll breadboard this up and see how she works.  Original design used +/-6V battery packs.  High voltage would be better.

jh

[hagtech]

After I did the TRUMPET years ago, I had thought of doing a lower cost JFET version.  But I was afraid it would reduce sales.  Hmmm, maybe it's time to go that route again.  That sort of led to the "MC" version of the TRUMPET, which had a headamp made from this dual matched JFET.  It worked.  It sounded real good, but was a bit on the noisy side.  So what if the BUGLE GOLD (new product idea) was done with JFETs and not opamps?  What if I copied the TRUMPET topology and added a headamp?  It would look like this:



With everything integrated, no need for the buffer stage on the headamp.  Just pull it all together and remove excess parts.  Simplify.  With higher supply voltages I would not need the current sources in the tails, but could just use resistors.  They make things more stable, easier to layout, and cheaper.  I worry that the output buffer does not have enough standing current, and it won't drive large output signals.  The nice part is that with 5 transistor pairs, I get enough gain for MC carts, gain and loading options, and fully balanced operation to boot.  I can add a phase switch to the output. 

Just stuff I am thinking about.

jh