Almost everything said here is NOT BS. It is right on track, for the most part. The input stage is a MINOR modification to the Vendetta RESEARCH SCP-1, introduced more than 25 years ago. with the same surface area of silicon on the input stage, made by the same manufacturer. The only significant difference is that I used the equivalent of 5 ohms, where Syn08 used 1 ohm as source resistors. That changes the effective noise from 9 ohms (typical Vendetta) to 6 ohms (Syno8). There is also some minor low frequency contribution from the Vendetta servo but it is minor, IF the IC is working properly.
This is like increasing the compression ratio of an automobile engine and claiming a whole new design, as the HP increases slightly.
This is like increasing the compression ratio of an automobile engine and claiming a whole new design, as the HP increases slightly.
john curl said:
Yeah, right, you wish
Read again:
Bottom line, 0.31nV/rtHz noise both measured and simulated, 25% less than the best case Vendetta, measured 0.005% THD 20-20KHz, measured 0.005% IMD 19+20KHz, allows DC coupling to the output, 46dB headroom, no JFET matching or sorting required to achieve all these, and... sounds good (according to my observations and a few other guys that already buit this) Use it with the RIAA correction/gain stage of your choice. Can go down to 0.25nV/rtHz with sorted JFETs...
Now, could you post your Vendetta results (before I do
) and tell us how much of your life you spent matching JFETs?BTW, the equivalent source resistor is 0.5ohm and the equivalent resistor noise for the input stage is 4.3ohm (for 10mA Idss devices). The rest, to 6 ohm, is coming from R16 divided by the open loop gain of the opamp, the servo and the folded cascode transistors noise. I can increase the source resistors to 2.2 ohm, the feedback resistors to 87ohm (also adding another 6dB to the headroom, to 52dB) and the overal noise will almost remain the same!
EDIT: Nobody has ever seen the official Vendetta detailed schematic.
Think about it Syno8! I use a matched pair of 2SK147 V Toshiba Fets or the 2SK146
Then I match for Idss with a matched pair of 2SJ72 V Toshiba Fets or the 2SJ73. All devices are measured for 1/f noise and noisy devices used elsewhere.
ALL I have to do is match for Idss which has been in the range 15-18ma for the last 25 years to 1 ma or less.
You do have a more accurate sumulation than I have ever done, but not a really a different circuit.
I compromised by adding a 10 ohm WIREWOUND mil grade multi-turn pot, center tapped to ground between the sources, and this is where the additional noise comes from almost completely. IF I had used a 2 ohm WIREWOUND mil grade multi-turn pot instead, I would have matched your noise level within 1 ohm. However, I use the pot to balance the 2'nd harmonic distortion due to the inherent imbalance in the complementary input fets. Later, versions used a 20 ohm wirewound pot, with 2 paralleled ten ohm resistors to make the equivalent of 10 ohms. I am now going to drop it to 1-2 ohms, just to match you. Thanks for the tip.
Then I match for Idss with a matched pair of 2SJ72 V Toshiba Fets or the 2SJ73. All devices are measured for 1/f noise and noisy devices used elsewhere.
ALL I have to do is match for Idss which has been in the range 15-18ma for the last 25 years to 1 ma or less.
You do have a more accurate sumulation than I have ever done, but not a really a different circuit.
I compromised by adding a 10 ohm WIREWOUND mil grade multi-turn pot, center tapped to ground between the sources, and this is where the additional noise comes from almost completely. IF I had used a 2 ohm WIREWOUND mil grade multi-turn pot instead, I would have matched your noise level within 1 ohm. However, I use the pot to balance the 2'nd harmonic distortion due to the inherent imbalance in the complementary input fets. Later, versions used a 20 ohm wirewound pot, with 2 paralleled ten ohm resistors to make the equivalent of 10 ohms. I am now going to drop it to 1-2 ohms, just to match you. Thanks for the tip.
I use Borbely's 320 phono stage with cascoded 4 x 2sk147GR and 4 x 2sj72 GR matched within 0.3ma IDSS for the input stage and no source resistors. I think that puts it at about 4.25 ohms equivalent input R.
( Runs 6.5ma and 7.1ma per device on each channel with drain resistors appropriately adjusted ). The 10 ohm trim pot which was left out when the servo was tested and found to have enough range to get rid of the offset.
Got the matched 2sj 72and 2SK147 devices almost 20 years ago for Erno's 1985 stand alone 585 moving coil headamp that never got finished and got the second chance to use them last year in the 320 rather than the specified 4x 2sk170GR and 2SJ74.
( Runs 6.5ma and 7.1ma per device on each channel with drain resistors appropriately adjusted ). The 10 ohm trim pot which was left out when the servo was tested and found to have enough range to get rid of the offset.
Got the matched 2sj 72and 2SK147 devices almost 20 years ago for Erno's 1985 stand alone 585 moving coil headamp that never got finished and got the second chance to use them last year in the 320 rather than the specified 4x 2sk170GR and 2SJ74.
Also use 1% Rel cap Teflons -for equalization and bias bypass along with 4 of Erno's super -shunt regulators ( one per gain stage ) along with pre-regulation - Input stage performance depends on the combination the active device characteristics, circuit topology, power supply performance, and the passive supporting components all contributing to the overall sonic signature.
Thanks to John and others in this forum for sharing some of their accumulated experience, opinion, and voodoo charms which all add to the knowledge base and competitive jousting
Thanks to John and others in this forum for sharing some of their accumulated experience, opinion, and voodoo charms which all add to the knowledge base and competitive jousting
john curl said:
I suspect you are playing dumb, anyway, you are hopeless... Read your own list of issues and see how they were addressed. It's much more than simply noise.
You are beaten at all performance parameters. The winner is not me, but 40 years of technology advances, making your handmatched design obsolete. The last and only thing you can do is to claim "Vendetta sounds better", which I personally can't debate.
john curl said:
What do you mean? I am a heretic. One owner of a high end store spit on me as soon as he heard that I am an electronics engineer. Orthodox engineers who learned by studying typical diagrams and answering multi-choice tests call me crazy audiophile.
I don't care. I keep working.
Sorry Syn for off-topic.
syn08 said:
I suspect you are playing dumb, anyway, you are hopeless... Read your own list of issues and see how they were addressed. It's much more than simply noise.
You are beaten at all performance parameters. The winner is not me, but 40 years of technology advances, making your handmatched design obsolete. The last and only thing you can do is to claim "Vendetta sounds better", which I personally can't debate.
syn08:
When you post like this I lose respect for your designs. This kind of derogatory posting is offensive to many members and shouldn't be tolerated. Please contribute for the common good or be silent.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.