Parasound JC3 Phono

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Joachim,

Thorsten, the measurements are from an Olympos in a Triplanar arm. I have now changed to a Titan i and will repeat the measurements with a bipolar phonostage i made. Sooner or later i will change the arm too. I have a Schröder Artemiz and wait for a Spiral Groove Centroid.

Sounds like you are having WAY TOO MUCH fun.

To tame the main arm resonance at ca. 7 to 9Hz i will try a subsonic filter or i could fill more oil into the damper of the Triplanar or both.

I found once the arm is damped enough or a sufficiently steep subsonic filter is usedthe subjective sound seems to loose something, maybe I'm just used to several nS worth of 8Hz jitter and a bit of AM for good measure?

Ciao T
 
Thorsten, yes, vinyl is my hobby so i really have a lot of fun here. Yes, there is a bump at 8 Hz but funny enough it sounds just fine. When i increase the damping other things suffer.
I have a subsonic filter on my mono subwoofer under 20Hz so the cones move very little anyway. I did the measurements with a simple bipolar stage. I worked a bit on the scaling and used some 1/12 octave smoothing so the levels come out in correct proportion. What you see is the noise of a single medium power bipolar. The 1kHz tone is produced by my Titan i and i am using the rare DIN 45 549 test record from 1978. That vinyl is much more quiet then modern ones i tried. I heard that Telefunken had a special lownoise recipy that is lost forever.The speed is Veff = 5.6cm/sec and Vsp = 8cm/sec. Under this conditions the noise is ca.75dB down at 1kHz. With 4 bipolars in parallel we whould get something like -80dB. I do not think that more then 4 in parallel
whould give us any more benefit concerning noise unless you really want massive Gm for a low distortion all in one go RIAA.. The bipolars have much more Gm then Fets anyway.
My open loop Fet stages have around -80dB distortion at nominal input whereas i get
-100dB distortion with degenerated ( to the same Gm as the Fet input ) bipolar input.
See also that the bipolar input has less 1/F noise at least in this case. I had trouble with flicker and popcorn noise on some bipolars though, even ones with low Rbee`so choosing the right bipolars is not easy either. P.S. I had a slight hum problem i could not get away quick so there is something at 50Hz and harmonics. The 300Hz artifact in the red trace is here too so Thorsten maybe correct that this comes from a mechanical arm resonance. Lets see how my other arms behave when i find the time.
 

Attachments

  • Noise of bipolar Phonostage.pdf
    7.9 KB · Views: 142
Joachim,

I did the measurements with a simple bipolar stage. I worked a bit on the scaling and used some 1/12 octave smoothing so the levels come out in correct proportion. What you see is the noise of a single medium power bipolar.

Yes, these can be pretty quiet. Bipolars do have quite a few undesirable side-effects more than J-Fets, especially in input stages...

The 1kHz tone is produced by my Titan i and i am using the rare DIN 45 549 test record from 1978. That vinyl is much more quiet then modern ones i tried. I heard that Telefunken had a special lownoise recipy that is lost forever.The speed is Veff = 5.6cm/sec and Vsp = 8cm/sec. Under this conditions the noise is ca.75dB down at 1kHz.

Well, 8cm/S peak are around 4dB higher than the common 5cm/s peak "0dB" level. So it would be around -71dB at 1KHz.

My open loop Fet stages have around -80dB distortion at nominal input whereas i get -100dB distortion with degenerated ( to the same Gm as the Fet input ) bipolar input.

Well, your open loop LP and Cartridge look like -35dB 2nd HD to me, so the difference between -80dB and -100dB HD would seem to be even more academic as the differences in SNR.

Ciao T
 
It would be better to allow a 6dB bump at 10Hz, rather than adding extra filtering. You can prove this to yourself by looking at the transient response of different approaches. In this case, adding extra high pass filtering will give inferior transient response, even if the amplitude is relatively flat.
 
My Triplar is a bit heavy so i end up a bit too low. I already use the best fitting counter weight as close to the arm as i can.
I tried several subsonic filters over time. The sound can be made tighter that way but i always end up to switch it off. The subsonic filter in my sub is integral to the active electronics i bought somewhere else so it is not that easy to modify. I may up ending to build my own subsonic crossover anyway when time allows.
Thorsten, i agree mostly that ultra low distortion is overrrated. I usually do not even measure my stages recently for distortion but i simulate them so that i am safely in the ballpark. As John already put down is that an open loop class A stage has ever diminishing distortion at lower levels anyway. I apreciate a quiet phonostage though on my 100bB efficient speakers i use to play loud at times. That bipolar stage i measured i am using right now and i can not hear noise from it at the listening seat. The slight hum of my poweramps is louder.
 
Joachim,

Thorsten, i agree mostly that ultra low distortion is overrrated.

I personally tend to aim my circuitry (excluding power amp's) for < 0.3% THD in production (with variations due to tubes etc. accounted for) at around 2V out.

The "continously decreasing harmonics" you mention can be seen very well in JA's measurement of my recent Phono:

Abbingdon Music Research PH-77 Phono Equaliser Measurements | Stereophile.com

With a totally inappropriate 600 Ohm Load (which will also take all the bass out - 120Hz -3dB) and 1V out you get 3% THD, but it looks "textbook open-loop class A":

666AMRfig06.jpg


With a more reasonable load (1V/100K) it looks like this:

666AMRfig05.jpg


It does not look much different with 10K either.

The bottom line is, well below 0.3% 2nd HD I prefer to worry about the distortion in speakers, microphones (never mind LP's themselves) rather than that of the of the electronics.

I leave pursuing 0.0001% THD to others, I already made stuff like that many times and the low distortion did not make it sound better as such (which may be related to the > 30% 2nd HD in the human ear at around 90dB SPL).

Ciao T
 
Last edited:
There is excellent sounding equipment out there that measures great too. A fruitless discussion, sorry. I burned my fingers on that many times. Fortunately an average music listener with good ears and taste does not care much about the measurements when it does not hum nor hiss or produces any other nasties like switch on transients etc. Neither a "designed" sound nore strict neutrality influence in any way the commercial success of a product i am afraid. There is simply no magic pill or i am too stupid to identify it.
 
Hi,

I read in the Stereophile text that your RIAA preamp has a +0.5 to +0.75dB rise between 40 and 50Hz... is it intentional ?

Well, it's either that or my trusty HP Scientific, my copy of P-Spice AND my AP2 are broken, right?

Overall, if you draw a +/- 0.5dB tolerance field the the PH-77 RIAA EQ falls right inside, 20Hz-20KHz. I could have made it +/- 0.1dB as well.

Production units are usually < 0.1dB tolerance from the target.

Ciao T
 
diyAudio Chief Moderator
Joined 2002
Paid Member
In older days they were showing the 20Hz-1KHz region wide on SP. Now they squeeze it not to draw the eye on 1/f & hum. Is it really only 70dB down at 20Hz your noise floor with just 48dB gain on your test equipment also? Or is it -90dB? Its too tight to the left to see on their chart.
 
Hi,

In older days they were showing the 20Hz-1KHz region wide on SP. Now they squeeze it not to draw the eye on 1/f & hum. Is it really only 70dB down at 20Hz your noise floor with just 48dB gain on your test equipment also? Or is it -90dB? Its too tight to the left to see on their chart.

The LF noise is hum, likely a ground loop or such via the AP2.

Using inputs shorted with 10R at 72dB gain nothing sticks it's head above -80dB re. 0.5mV (it is basically the same with a turntable and cartridge). Plugging the inputs into the AP2 with an attenuator to give 10R Z-Out and I have a -70dB 50Hz hum and tons of higher harmonics. Maybe JA had the same issues?

Ciao T
 
Last edited:
Member
Joined 2006
Paid Member
Hi John

In my 48dB Salas Riaa I am using a 8.5mA Idss 2SK170BL cascoded with a BC550C for the input.

It is powered at 28v and the colector resistor is a 3k3r.

For Rs under the jfet I am now using a 33r Takman carbon resistor.

This is my second ultra silent build... The previous one (I built for my Wife TT) also used a 33r Takman for Rs.

The noisier build (used for another TT) has a 33r Kiwame at the same position.

I do not have the means to do propper (or any) noise measurements so this is only my DIY subjective listening opinion..... It seems I finally found the way to make this wonderfull sounding Riaa dead quiet. :)

PS: This post is intended to thank you for your previous input about excess noise :)
I am sorry for the off topic but I could not help it.
 
Last edited:
I would like to comment about the RIAA curve. Much is made of changes to it, often without real basis.
The RIAA curve is a COMPROMISE that was made perhaps 60 years ago to get an AVERAGE over a number of curves developed by different companies often in different countries.
If we had to do it over again, a continuous 6db/octave rolloff starting at 50 Hz would have been a better choice, but we have to live with what we have. Now what about the frequency extremes? Over the decades, many approaches to low frequency record warp have been attempted. One approach is a multi-pole high pass filter at 15 Hz or so. This has been discarded over the years. A second approach is the single pole 20Hz filter in the IEC 'standard'. I find this approach almost useless, because it just reduces the bass frequencies, and IF it really helped, you have a more serious problem with IM in the phono cartridge, itself, not being effected by the rolloff. It is better to fix the tone-arm phono cartridge combination, rather than try to roll off the warp frequencies electrically, after the fact.
When it comes to very high frequencies, the so called Neumann compensation is not really appropriate, because the tuning of the system cutter feedback already addresses the issue, and the extra boost just gives a high frequency emphasis, that might be good sounding, but it is not as accurate at 20KHz.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Some slightly tweak to subjective given where they want to head their phono stage sonics as a total. Some reminding a BBC speaker curve with a bit boosted mid-bass and a mid to highs gradual shave off (over much smaller amplitude bracket than a loudspeaker of course), or a bit tilted up/down as a whole. I think you tend to tilt a wee bit down?
 
John,

I would like to comment about the RIAA curve. Much is made of changes to it, often without real basis.

Plus the actual accuracy on the cutting side is rarely considered.

Does anyone hear actually believe cutting lathes really offer +/-0.05dB EQ accuracy?

It is better to fix the tone-arm phono cartridge combination, rather than try to roll off the warp frequencies electrically, after the fact.

Yup, absolutely. However getting damoing right can be interesting.

When it comes to very high frequencies, the so called Neumann compensation is not really appropriate, because the tuning of the system cutter feedback already addresses the issue, and the extra boost just gives a high frequency emphasis, that might be good sounding, but it is not as accurate at 20KHz.

First, proper Neuman cutters (not Audiophool BS modified systems) use motional feedback and suppress any peaking. The WE and related cutters more common in the US are another story.

Second, my Phono has this under remote control. Once, during development I mixed up something in the spec and as a result our Software guy had the "neumann EQ" relay drive reversed. Before this listening session I did not a full set of measurements but we listened instead.

Everyone completely and absolutely hated the so-called "enhanced RIAA" EQ which we thought was with the extra EQ. When I measured the following day, it turned out the "bad enhanced RIAA" was actually the stock RIAA withput HF boost, while the "good sounding original RIAA" was actually the one with Neuman compensation.

So the ears have it...

Neuman EQ here please.

Ciao T
 
Hi,

Some slightly tweak to subjective given where they want to head their phono stage sonics as a total. Some reminding a BBC speaker curve with a bit boosted mid-bass and a mid to highs gradual shave off (over much smaller amplitude bracket than a loudspeaker of course), or a bit tilted up/down as a whole. I think you tend to tilt a wee bit down?

Well, what I intensely dislike are situations where the tolerances can accumulate and causes a tilted up (even slightly) response. By tilting it down a little (remember, we are still 20Hz-20KHz +/-0.5dB for each and every production machine) the sound becomes subjectively a little nicer and even all accumulated tolerances cannot tilt it the other way.

Ciao T
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.