An update.
I built the 5:1 attenuator and connected it with custom cables I built using DH Labs BL-1 interconnect wiring. Got the capacitance of the cables and the arm wiring to 155 pF. With these cables connected to the ADC LMF arm and the AT 150 mlx cartridge mounted, everything connected to the attenuators and the PS Audio GCPH phono preamp, I can use the 66 dB gain setting with very acceptable noise levels. I probably only need 55-60 dB of gain to get a 300 mV into my line pre-amp in any case. The 100 pF input capacitance in the GCPH is no longer an issue, as its now isolated from the cartridge by the 37K series resistor in the attenuator.
The attenuator was built inside a male/female pair of locking RCA connectors similar to the DH Labs versions. So these are plugged into the GCPH and the IC BL-1 interconnects are pugged into the female RCA. |
Al:
Not to worry.
This should not be a problem because the physical layout of the system w/ attenuator is not much different than the stock system without it. The attenuator plugs are short (about twice as long as the standard locking RCA) and connected to the pre-amp RCAs. Everything then connected to attenuators is identical to the stock layout. So the only difference from a noise pick up point of view is the short length of the attenuator barrel, which is of coarse metallic and grounded to the shield.
The impedance up to the attenuator input would be the low MM cart impedance. Its only the impedance after the series divider resistor of 40K (which would be 8-10K) that would be susceptable, and that would be about 1/4" inside a shielded RCA plug, Not a significant area for noise induction IMHO.
But your concern is exactly why I would not use a "Y" adapter. Those are much larger, and even though still shielded, are more susceptible to induced noise.
The other issue I think you are implicitly raising is that the interconnect wiring from the RCA to the actual input of the IC within the preamp would also be susceptable. PS Audio has done a good job here, judging from my noise measurements with a 47K input impedance and the input RCAs open. Very slightly different than the 1K loading choice, which can be easily explained by the increased thermal noise of a 47K resistor vs a 1K resistor. |
Very nice analysis, DHL.
My only further comment is that when I raised the signal-to-noise concern I was not so much thinking of noise generated BY the input stage or by the source impedance, but rather that the increased source impedance at low and mid frequencies would increase susceptibility to pickup of EMI or hum that would be introduced TO the input stage. It could very well be that that won't be an issue, but that was the possibility I had in mind.
Best of luck, however you decide to proceed!
-- Al |
Al:
Thanks for your comments. To address your points:
The phono stage of the GCPH is bipolar. PS Audio uses and Analog Devices microphone pre-amp SSM2019 which has a bipolar input.
If it were an FET stage, it would be a non-issue, as FET input stages have extemely high inherent input impedances, so a source impedance as high as 100K would not impact them much. But bipolar stages, on the other hand, are quite different. Depending on bias conditions, and the gain of the input stage, input impedance may only be a megohm or less, and 50-100K source impedances can make a big difference. The input stage corner frequency is a function of source impedance, and may affect stability in closed loop mode, depending on the design.
Besides the sound signatures of the two types, generally designers use an FET input for MM carts and biploar for MC, because the lower noise current of an FET gives better noise performance with MM, and the noise voltage created by higher noise current of a bipolar is reduced by the lower source impedances of MCs.
Now to the attenuator. You are right about the DC resistance and that the MM impedance will be low at low frequencies. But the response pole created by the RL network in the MM will have the impedance climbing starting at about 200 Hz. It will climb at 6 dB/octave until it reaches the resonant point, and by that time it will be close to 50K. This is exactly where you don't want a high source impedance, at the highest frequencies in the audio band. If I use a 5:1 attenuator, with a total load impedance of 47 K facing the MM cart, the preamp will see a more constant 8-10 K source impedance, all the way up to the cartidge resonance and beyond. Plus, any capacitance at the input of the preamp will be isolated from the cart by the 40 K divider resistor, leaving only the native capacitance of the tonearm cable and interconnects loading the MM cart. IMHO, it is better for a bipolar input stage to see a constant impedance, independent of frequency, throughout the audio band, even if its higher than one at low frequencies.
Re the noise performance, yes there will be a bit higher noise from a 10 K source than from a 500 ohm - 1K source. And, I will need to use a higher gain as well, which could also increase the noise. But my noise measurements of the GCPH (see my post in the amplifiers section) indicate that even at 60 dB of gain in the GCPH, the noise levels are not bad. So today (as opposed to when I did this in the 70's) we are blessed with high gain low noise designs from both the FET and bipolar camps which makes using this attenuator principle more attractive.
Re input capacitance, I would build the attenuator in a RCA loading plug, with a male RCA on one end and a female RCA on the other. The loading plug would be connected directly to the pre-amp input RCAs. The two resistors would be wired in the plug itself. I doubt if the stray capacitance would exceed about 5-10 pF. Even if you use a "Y" connector (which I do not prefer) with a loading plug on one leg of the Y, the added capacitance would still be under 30 pF IF you use a good quality Y adapter. In either configuration, shielding is not a factor.
|
Interesting thought about the 5:1 attenuator. It seems to me that it might very well do the trick, but I would have some concerns.
One concern is the possibility that the higher source impedance seen by the phono stage at low and mid frequencies, compared to direct connection of the cartridge, would result in some degradation of the phono stage's signal-to-noise performance, as well as increased susceptibility to hum pickup in the phono stage input circuit. Note in one of the links I provided earlier that the cartridge's DC resistance is spec'd at 530 ohms.
Also, obviously, the attenuator would have to be constructed so as to provide good shielding.
Finally, stray capacitance in the attenuator and its input connector would reduce the 100 pf saving at least a little bit, perhaps by 10 pf or so as a rough ballpark guess.
Best regards,
-- Al |
Just got another wild idea, actually from another post where the Soundsmith folks are recommending the use of attenuators between the cartridge and pre-amp. As I posted there, in the late 70's we did this with MM cartridges to increase the effective phono overloads of pre-amps. The sound improvement was very noticable.
It occurs to me that a 5:1 attenuator would isolate the pre-amp input capacitance from the cartridge, leaving only the cable capacitance. This would give me another 100 pF to play with, at the expense of more required gain, which is in adequate supply with the GCPH. This would also isolate the very high impedance of the cartridge at the LC resonant point, another issue that affects the stability of the phono preamp input stages. With the attenuator, the phono pre-amp sees a pretty constant 10 K source impedance throughout the frequency range, not the nonlinear 1k to 25K it sees unattenuated. |
Thanks for everyone's feedback. Incredibly useful.
Looks like no matter what cartdridge I end up with, low capacitance cables may still be required. As Al rightly points out, large low Q resonant peaks at 40KHz can be just as damaging as ones at 15 KHz.
Its also clear that Needledoctor sales folks may be more interested in selling cartridges that providing correct information.
Thanks to Johnnyb for the cable suggestions, esp the Cardas versions. The Golden Presence has an incredibly low capacitance of 12 pF/ft x 4.1 ft (1.2 m) for a total of 50 pF. They are still a bit pricey but good know they are available as a fallback. Not sure the Blue Jeans stuff meets my criteria. I would like to stay with silver plated Cu/Teflon insulation which may mean I'm building my own. I have some DH Labs BL-1 Silversonic cable that is 17 pF/ft which may work. |
When I was shopping for a phono pre a year ago, I came across a great deal on a demo Musical Surroundings Phonomena II. However, I also found out that the lowest input capacitance setting on that phono pre is 200 pF for a total of 300 with TT cable, so I passed. I got the Jolida instead and its wide range of capacitance (for MM), resistance (for MC), and gain settings made it easy to dial in my current rig and any LOMC, HOMC, MI, or MM I may get in the future.
And yes, the Needledoctor guy was wrong. The 150MLX wants to see 100-200 pF total including the tonearm cable. At least, in my case, knowing the Technics cabling is spec'd at 100 pF, I get the best sound when the loading into the phono stage ranges from 0 to 100 pF, and preferably 50-ish. |
DHL, thanks for providing the references and the additional comments. Yes, the Hagerman paper is somewhat simplistic, most notably in not addressing the fact that for MM's too little capacitance can result in too little upper treble. Good inputs from Johnny. His finding that optimal results occurred with 147 pf, right in the middle of the recommended range, would seem to further confirm that the salesperson you spoke with was mistaken. Re MC's, the response of the cartridge at audible frequencies will of course be pretty much insensitive to load capacitance. A point to keep in mind, though, is that the amount of load capacitance may still have significant audible consequences (with lower capacitance generally being better), due to effects that the ultrasonic resonant peak may have on the phono stage, and due to constraints that it may impose on the choice of resistive loading. See the post by preamp and cartridge designer Jonathan Carr (JCarr) dated 8/14/10 in this thread. Regards, -- Al |
04-17-12: Dhl93449
Johnyb:
Where can I get a 100 pF RCA/DIN phono cable? Most manufacturers of aftermarket versions don't spec the capacitance. ADC made a 100 pF special version back in the day, but this is NLA and I have never seen a used one.
FYI the ADC arm's internal wiring is 30 pF (headshell to DIN in the arm), so the DIN/RCA cable has to be about 70 pF to have a total interconnect of 100 pF. The input capacitance on the GCPH is fixed at 100 pF. If your tonearm cable is already 30 pF and the GCPH input is 100 pF, then you want the phono interconnect to be 70 or less. Kimber Kable gives the capacitance specs for *all* their interconnects including phono interconnects. Their TAK Cu is 47.1 pF, so that would make your entire phono signal chain 171 pF, which is nearly ideal. All the phono cables in their standard line are 1M long and 47 fP, give or take a couple tenths of a picofarad. Nordost also publishes specs, and their Frey 2 is 25 pF per foot. Cardas also publishes specs. Their Neutral Reference is 20 pF per foot. So here are a few off the top of my head that offer low enough capacitance to dial in the right overall total for your phono signal chain. I'm sure there are more makers that publish, but I knew that Nordost and Kimber interconnects tend to have low capacitance. I was pleasantly surprised by the Cardas. That may explain why his wire is popular for tonearms. |
Blue Jeans Cables has interconnects that are 12.2 pF/ft. They might make you a set with DIN plug, or you could buy a set and incorporate them. They are relatively inexpensive, but many people report excellent, good-sounding results.
Some people also change the input capacitance on their MM input. 25 or 50 pF should help, of course you'd have to modify your pre. It's not only AT, some other mfg also recommend total load of < 200pF.
Regards, |
Johnyb:
Where can I get a 100 pF RCA/DIN phono cable? Most manufacturers of aftermarket versions don't spec the capacitance. ADC made a 100 pF special version back in the day, but this is NLA and I have never seen a used one.
FYI the ADC arm's internal wiring is 30 pF (headshell to DIN in the arm), so the DIN/RCA cable has to be about 70 pF to have a total interconnect of 100 pF. The input capacitance on the GCPH is fixed at 100 pF. They use all surface mount capacitors so exchanging or modding input caps in this phono preamp is near impossible.
I think your setup with a total of 147 pF is right on the money, as the references indicate the lower the better.
To get to 70 pF in the interconnect would mean I have to build my own cable, or build an interface box with a short cable to the arm/DIN and RCA bulkhead connectors. I have low cap RCA/RCA interconnects I can use from the box to the pre-amp. Or I could shorten the cables I have to under 12" and mount new RCAs. But that limits where I can place the preamp; it would need to be immediately adjacent to the turntable.
A lot of work just to use this cartridge though. I am leaning back to a HO MC or maybe MI
|
Swap to a 100 pF arm cable if you can. I've been using an AT150MLX for 4 years and love it. I currently have a JoLiDa phono stage with highly adjustable capacitance. I have a Technics SL1210 M5G TT whose arm cable capacitance is 100 pF. I have played with the input capacitance in my preamp and found that 0 through 100 pF works OK, but my personal preference is an input of 47 for a total of 147 pF (or 150). However, 200 total is good enough that it could come down to system matching. I definitely wouldn't want to run it into 300 pF or above.
I hope you can get it to work for you. The AT150MLX has a very musical balance of robustness and body on the one hand and spatial cues and detail on the other. It's also an excellent tracker. |
Al:
Thanks for your input. Since posting I have done more digging into this and it's a bit more complex than I first thought.
The site you reference is a bit misleading because his plots don't take into account the effect of the series resistor on the shape of the peaking (he assumes a Q factor of 1).
I found this reference by Rod Elliot to be a bit more realisitic:
http://sound.westhost.com/articles/cartridge-loading.html
The problem is that some manufacturers use the electronic tuning response to compensate for mechanical resonances of the stylus tip/cantalever/motor devices in the cartridge, so that the overall actual response is not just predicted by the RLC network of the elctrical components. See, for example:
http://www.tnt-audio.com/sorgenti/load_the_magnets_e.html
So the electrical modeling would suggest that the lower the capacitance, the better. But for some MM carts this might result in a reduced high end response due to mechanical effects.
Not sure how any of this is related to the performance of the AT 150 mlx, but I am begiining to remember why I don't like the concept of MM carts, and may re-consider using them. MCs don't have this issue due to the much lower inductances and impedances. They are tolerant of higher cap loading because of this. |
I can't say for sure, and I have no experience with the particular cartridge, but I'm highly skeptical of what the salesperson told you. Although the Audio-Technica site does not indicate a spec for the inductance of the cartridge coils, this review, which is linked to at their site, indicates 450 mH, while the specs here for the VM version (not sure what that is) indicate 350 mH. Plugging those values into the calculator provided here shows that in conjunction with the 150 pf mid-point of the recommended load capacitance range a high frequency resonant peak will exist at 19.4 kHz in the first case, and 22 kHz in the second case. Those seem like sensible values. Increasing the total capacitance to 350 pf (250 pf cable + 100 pf phono stage) would reduce those numbers to 12.7 and 14.4 kHz, respectively, and would also increase the magnitude of the resonant peak. While it is possible that the design is intended to use that resonant peak to compensate for what otherwise would be a dip or rolloff in the upper treble, and my understanding is that it is not uncommon for that to be done, in this case I am doubtful. And, of course, what sense would it make for them to provide a load capacitance spec that excludes cable capacitance, without providing an indication of how much cable capacitance is being assumed? Personally, I wouldn't worry about going a little over the 200 pf number, perhaps to a total of 250 pf or so, but I would not be comfortable with 350 pf. Regards, -- Al |
