Parasound JC3 Phono

[john curl]

What a bunch of 'crybabies' ! Of course, it is NOT designed for MM cartridges. The GAIN is too high! It is a COMPROMISED DESIGN, but not TOO compromised for MC cartridges. I am surprised that anyone who listened to it would not like it, especially with what is usually promoted here, as a worthwhile project. Erno Borbely is a good engineer, and he made a BUILD-ABLE KIT for others to make. I might make a similar design for Parasound, on occasion. I do NOT give them VENDETTA level designs. It is impossible to teach them how to build one at distance.

 

[RCruz]

No, too much input capacitance from numerous paralleled JFETs. If you get the noise low enough for MC use, you have probably ended up with too much JFET input capacitance for MM use. Bear in mind, this capacitance is nonlinear.

Cheers,
Bob

Hi Bob

This capacitance nonlinearity depends on what ? Frequency ?

 

[Joachim Gerhard]

You could imagine a Fet as a very bad solid state capacitor.

 

[ticknpop]

I built the Borbely 320 with matched 2sk147GR and 2sj72GR quads rather than the 170/74GR the kit included. The cascoding reduces the non-linear capacitances of the parralled devices and the second stage - it sounds quite fine with 4 of his super shunt regulators and Relcap teflon 1% Riaa caps. As John states it's designed for kit building for those of us not having a decent stash of unobtainable jfets to match (Jacco, John, Nelson, Charles, and Erno excepted).

 

[tvi]

A little Googling

AUDIO TECHNICA AT33EV
Coil impedance: 10 Ohms (1kHz)
Coil inductance: 22 uH (1kHz)

Kiseki Purpleheart Sapphire
42 Ohms 120uH

Audio Technica AT-F3 III
Coil Impedance 12Ω/12Ω (DC/1kHz)
Coil Inductance 25µH (1kHz)

Denon DL304
Output Impedance 40 ohms
Coil Inductance 50µH @ 1kHz

haniwa HCTR01
Input Impedance 0.8Ω (1kHz)
Inductance 1.3µH (1kHz)

Dynavector DV DRT XV-1S
Impedance R=6 ohms, L=18 uH

goldring Eroica LX
Internal Inductance 12 µH
Internal Resistance 8 Ohm

Grado (all Moving Iron)
Statement1 Reference1, Master1, , Sonata1 & Platinum1
Output at 1kHz 5CM/sec. 0.5mV
Inductance-2mH
Resistance-2 ohms

Rgds
James

 

[scott wurcer]

Hi Bob

This capacitance nonlinearity depends on what ? Frequency ?

No voltage, at these levels irrelevent. You should stick with listening impressions if you are not interested in the engineering details of what is involved.

 

[scott wurcer]

You could imagine a Fet as a very bad solid state capacitor.

Please at least do a first order estimation of a low order non-linearity with an excitation of milli-volts with a DC bias of a few volts. We are talking about the proverbial -120dB distortion that is obvious to the spouse in the kitchen.

 

[Bonsai]

All those parallel FET's do look like an overkill, but as an exercise in a low noise 'all FET' inpuit stage, I suppose it has its merits. Re the input capacitance, Scotts' comments aside, why not cascode the front end?

 

[Werner]

Grado (all Moving Iron)
Statement1 Reference1, Master1, , Sonata1 & Platinum1
Output at 1kHz 5CM/sec. 0.5mV
Inductance-2mH
Resistance-2 ohms

Grado's inductance spec for their Statement series is wrong.

I once measured one an obtained 60 uH.

It worked well enough into a current-input phonostage.

 

[john curl]

First of all, the Grado cartridges are NOT MC cartridges, they will OVERLOAD a typical MC input stage.
Second, we have been using paralleled high input capacitance jfets for the last 30 years or more. They work for typical MC cartridges, but will NOT work with MOVING MAGNET cartridges. The dynamics are VERY DIFFERENT. I am surprised at how little many here seem to know about this. Perhaps, phono playback was popular before your time. However, many years of experience shows that, within reason, paralleled jfets are an ideal way to step up MC cartridges.
What about non-linear capacitance? Well, this is where we try to CASCODE, and if possible, operate the input devices with a reasonably high voltage, in order to LINEARIZE the INPUT CAPACITANCE.
The Borbely design is an approach to this effort, the only drawback is the INPUT DEVICE VOLTAGE. It should be higher, if possible. Special high Vp jfets would work optimally as the input cascode, if they are available. However, there are these sorts of reasons why a Vendetta Reseach or a more exotic preamp is so complex when it comes to biasing the cascodes, and current sources.

 

[gk7]

All those parallel FET's do look like an overkill, but as an exercise in a low noise 'all FET' inpuit stage, I suppose it has its merits. Re the input capacitance, Scotts' comments aside, why not cascode the front end?

The front end _is_ cascoded in Ernos design.

 

[IngemarR]

As gain can easily be set with a resistor, the design is made for both MM and MC cartridges. From the manual:
"

The EB-199/320 is the Top-of-the-Line ALL-FET MM/MC phono preamp with 44dB for
MM and 54, 64 or 73dB gain for MC. It can handle all low-output MC pickups on the​

market. RIAA accuracy is better than 0.1dB.
"

This Borbely design is only for MC cartridges, just like a Vendetta would be.

 

[IngemarR]

Well, the 2SJ74:s are still available, so it is hard for me to agree with you on this. One just has to look an be prepared to pay a little more than back in the old days. The high Vp JFETs used in the cascode are also available - at least in the lower Idss ranks but that is fine as Erno usually uses 2mA in his input stages.

Also, this design is almost impossible to make by amateurs, because of the lack of parts that are necessary, just like a Vendetta.

 

[IngemarR]

Huh, that does not correlate with what I have read about the 320 kit.

Unfortunaltely, I have not heard it in a well built version with good PS regulators and components, and now the kit is not available anymore....

Here's what another user writes about the 320 kit:
"

Erno, the FET phono stage and shunt regulator have fully broken in now and it is the best phono
stage for MC’s that I have ever heard. I get all the ambience of the recording site and if the record
is quiet I get the quietest backgrounds that I have ever heard.
I was listening to recordings that were 25 years old and have been surprised on how much I have
missed. I don’t think there is a commercial phono stage that sounds any better whether it is tube
or transistor.
I have the phono stage loaded with a 250 ohm resistor for the Koetsu and I use 47k with the​

Grado. Both sound very nice
"

I built that Borberly design i the 90 ies some time...must say it was completely dead in the water compared to my other setup with the Hiraga Prepre and a ecc 83 tube riaa with split riaa network...(LA-audio)...It was dull and boring...

 

[IngemarR]

But the input JFETs are all cascoded which should remove the Miller effect, and thus keep the input capacitance very low, right?

No, too much input capacitance from numerous paralleled JFETs. If you get the noise low enough for MC use, you have probably ended up with too much JFET input capacitance for MM use. Bear in mind, this capacitance is nonlinear.

Cheers,
Bob

 

[RCruz]

No voltage, at these levels irrelevent. You should stick with listening impressions if you are not interested in the engineering details of what is involved.

Hi Scott

I am very interested in this discussion about topologies that might be good for a MC input stage.

It is possible that my questions might seem noob to you but they are very important for my overall understanding of what is beeing discussed here.

I know jfets suffer from miller and that can be tamed with cascoding. Untill now I only used a single cascoded jfet in my phonos so I do not know about paralleled devices.

IMO using higher Rs values also reduces miller effect but we I noticed higher noise coming from the resistors.

As you can see, my questions rise the subject and I normally get severall opinions that help me form my opinion.

Quote from JC "What about non-linear capacitance? Well, this is where we try to CASCODE, and if possible, operate the input devices with a reasonably high voltage, in order to LINEARIZE the INPUT CAPACITANCE."

 

[ticknpop]

the Idss of the Borbely 320 input devices are 3,5 to 4.0 ma and when supplied as matched quads did not use individual source resistors

my 320's 2sk147/j72 input stage are idss 7.0 each run without source resistors with R3 + R6 drain resistors set to 110 ohms to account for the 28 ma input stage current

John - rather than the self cascoding with the high Vp 2sj103-k148 devices Erno used what voltage would you use to bias the cascodes - Erno used 10 volts in his 1985 MC preamp with the Rohm 2SD 786 and SB 737 as cascodes. I expect the trade off is dissipation versus voltage if the input devices are high Idss hence the choice of Gr grade devices. The k147-j72 series had bigger pacakages allowing higher device dissipation the the k170-j74 in addition to their lower noise structures. I expect the higher dissipation of the 147-j72 was also of benifit in the Vendetta in using the high idss V grade devices you used.

 

[MiiB]

IngemarR.. Trust me on this...I know 100 % What I am talking about....

I had at the time a complete Forsell reference turntable and Arm with several different cartridges...Shinon RED's and Koetsu's. The Borberly is Dull.. Plain Dull...The PRE-PRE-- and Tube RIAA.. was a totally different experience...Simply much much more music..!!

To some the Borberly may just be the best, but then what have they compared to..??

 

[SY]

Well, this is where we try to CASCODE, and if possible, operate the input devices with a reasonably high voltage, in order to LINEARIZE the INPUT CAPACITANCE.

But doesn't that increase gate leakage current?

 

[Bonsai]

The front end _is_ cascoded in Ernos design.

Sometimes I am an ********.

Apologies Erno!