Hey there, I revisited most of Borbely's circuits in the last few days.
While I admire the overall technical excellence, I wonder whether the extra complexity produces better 'sound'. Undoubtedly the cascode operation yields superior performance. Does it contribute significantly to subjective improvement?
Has anybody built and compared the 'sound quality' of his versus a lesser design?
Any thoughts?
Mark.
While I admire the overall technical excellence, I wonder whether the extra complexity produces better 'sound'. Undoubtedly the cascode operation yields superior performance. Does it contribute significantly to subjective improvement?
Has anybody built and compared the 'sound quality' of his versus a lesser design?
Any thoughts?
Mark.
Mr. Borbely uses a self biased JFET-JFET cascode which demands precise Vt selection plus precise device pair matching. Single die matched JFETs like the 2SK389BL are available for the input diff-pair, but single devices with higher Vt like the 2SK246BL must be used in the cascode, and if both of these cascode JFETs do not have the correct Vt and precisely matched parameters, distortion results. It is very difficult and expensive for DIY'ers to acquire and test JFETs to the precision required.
I use a 2SK389BL diff input, but use a LED biased MAT02 single die NPN bipolar pair for the cascode, plus a 2SJ109BL with a LED biased MAT03 single PNP bipolar pair for the complementary symmetrical cascode. The high gain and stable Vbe of the MAT0X bipolars generate low distortion over a wide input range without complicated matching.
Some of my amps directly connect the 2SK289BL and 2SJ108BL like Mr. Borbely's circuits. Some of my amps use mirrored current sources to each diff-pair with a shared JFET load for temperature tracking to match the JFET inputs.
Mr. Borbely’s phono amps look very clever, but demanding on JFET transistor selection and matching.
I use a 2SK389BL diff input, but use a LED biased MAT02 single die NPN bipolar pair for the cascode, plus a 2SJ109BL with a LED biased MAT03 single PNP bipolar pair for the complementary symmetrical cascode. The high gain and stable Vbe of the MAT0X bipolars generate low distortion over a wide input range without complicated matching.
Some of my amps directly connect the 2SK289BL and 2SJ108BL like Mr. Borbely's circuits. Some of my amps use mirrored current sources to each diff-pair with a shared JFET load for temperature tracking to match the JFET inputs.
Mr. Borbely’s phono amps look very clever, but demanding on JFET transistor selection and matching.
Thanks for the inputs.
Hey Grey, Borbely first published circuits without using cascode. He added them on at a later time ("all-out" as he put it). No doubt I like the technical merits. But wonder whether it's truly necessay and how much it 'changes' the sound. Have you built jfet preamps (complimentary type design) with and without cascode? Do they sound different? Mark.
Hey Grey, Borbely first published circuits without using cascode. He added them on at a later time ("all-out" as he put it). No doubt I like the technical merits. But wonder whether it's truly necessay and how much it 'changes' the sound. Have you built jfet preamps (complimentary type design) with and without cascode? Do they sound different? Mark.
to further make grey's point, for some applications that use sufficiently high rails, the choices are either to cascode or do a balanced bridge design to get sufficient voltage swing.
also, you will see in past threads that performance in some key areas (for example, noise) is improved when the the drain to source voltage is kept below (or within) a certain range.
similar to linesource's comments, i am most happy with designs that use a bipolar cascodes above the jfets.
acquiring the parts
and doing the critical matching 
needed for jfet /jfet cascodes is more headache than i can stand at this point.
i've built some of borbely's preamp line stages and loved 'em.
i've been very happy with a borbely's servo 50 as my primary power amp. i use regulators for the front end and lots of bias on the output mosfets.
one day i'll try a jfet VAS stage and no global feedback, but i'm not quite there yet ... still studying and reading, soaking up john curl's pearls and waiting for charles hansen to drop some more ...
mlloyd1
also, you will see in past threads that performance in some key areas (for example, noise) is improved when the the drain to source voltage is kept below (or within) a certain range.
similar to linesource's comments, i am most happy with designs that use a bipolar cascodes above the jfets.
acquiring the parts
and doing the critical matching needed for jfet /jfet cascodes is more headache than i can stand at this point.i've built some of borbely's preamp line stages and loved 'em.
i've been very happy with a borbely's servo 50 as my primary power amp. i use regulators for the front end and lots of bias on the output mosfets.
one day i'll try a jfet VAS stage and no global feedback, but i'm not quite there yet ... still studying and reading, soaking up john curl's pearls and waiting for charles hansen to drop some more ...
mlloyd1
Folks, thanks for the inputs again.
I located a shop which still carries 2SJ72GR, misc Jfets, and Rohm 2SB737 & D786. Very curious about Borbely's design and tempted to invest on these hard-to-find devices.
John, you use a lot of fets in your designs. This may sound silly - do they really sound that different? Like your old JC1 circuit with those J??? fets. Lost some sleep since I couldn't get them way back when I was still at university!
Mark.
I located a shop which still carries 2SJ72GR, misc Jfets, and Rohm 2SB737 & D786. Very curious about Borbely's design and tempted to invest on these hard-to-find devices.
John, you use a lot of fets in your designs. This may sound silly - do they really sound that different? Like your old JC1 circuit with those J??? fets. Lost some sleep since I couldn't get them way back when I was still at university!
Mark.
You still haven't specified a particular design, but I will say that I'm not comfortable with Borbely's habit of sitting the cascode JFET right on top of the gain device, leaving little Vds for the gain device. That said, the designs I can call to mind at the moment are pretty sound designs, otherwise.
Grey
Grey
john curl said:
Yes, John, they are very good deal! It is sad that people do not want to spend the little money his kits cost.
They are not just a good deal economically, they sound good to!
At 70+, who knows how long Erno will keep going.
Erno's 320 MC/MM RIAA amp is one heck of a low noise design! Now that he has removed the 8 source resistors, noise will be even lower
Maybe not as low as your's, John, but close.
At a noise level like a 10 Ohms (I think it was that) resistor, not many MC RIAA amps will beat that.
John, what input JFETs did you use to get 10 Ohms? SK146/7?
Sigurd
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.