I have build a pre-pre with 2 x 2SA1085 plus 2 x 2SC2547 that is so quiet, that you think it is a transformer. They have lower Cob ( 3,5pF ) then the BC337, 327 and very linear and high Hfe. E grade is up to 800 and there is an F grade of the 2SC2547 with up to 1200 !
They do have lower early voltage though.
They where designed by Hitachi in the 70th and have being made by Renesas to at least 2010, so there should be quiet a few out there. We need them in the input stage only anyway.
They do have lower early voltage though.
They where designed by Hitachi in the 70th and have being made by Renesas to at least 2010, so there should be quiet a few out there. We need them in the input stage only anyway.
Some may not have a balanced input so in this case we need unbalanced out from the Masterpiece. The most simple way is to short one input to ground and use only one output.
That works, much like an unbalanced stage but we loose 3dB of noise, common mode rejection and canceling of second harmonic. A better way is to convert the output to unbalanced. The most simple way is to use an opamp and arrange it as a differential amplifier. Frans has done that successful at the MC input of his phono stage. For a balanced-unbalanced converter this is not the optimum solution because we do not want a too low input impedance to hold the distortion in check. Do we raise the input impedance of that type of converter we get a noise disadvantage because the input resistors are in series with the signal. A more optimum solution is to put two buffers before the differential stage so we can make the input resistors on the differential opamp small. Fortunately we have very good quality buffers already on board of the Masterpiece so all we need is a relative low impedance differential Opamp. I found out that very good sound can be got from Opamps when they do not have the burden to amplify tiny signals like MC cartridges. I build good sounding MC stages with opamps too but non of them could compete in sound with a good discrete solution. Ones the signal is boosted though by a factor of say 100x the good points of an opamp come into play. More then two years ago i had the same problem and needed a balanced-unbalanced converter. First i wanted to use the Self Ultra Low Noise solution that puts a lot of buffers and differential stages in parallel to reduce the noise, much as we do at the input of the paradise.
I published the most extreme solution with 8 buffers and 4 diff amps PER channel. Scott Wurzer jumped in and persuaded me that this is not particular elegant and after some back and forth i build a converter with two B1 buffers per side and an AD797 as differential mixer. That worked pretty decent. The ADA797 is still an excellent part. Read Samuel Groner´s research about Opamps where the AD797 came out first and the LT1648 second compared to much more recent designs. We do not need additional buffers anyway and i would redesign the differential mixer by adding a class a buffer in the output. Then we are set.
That works, much like an unbalanced stage but we loose 3dB of noise, common mode rejection and canceling of second harmonic. A better way is to convert the output to unbalanced. The most simple way is to use an opamp and arrange it as a differential amplifier. Frans has done that successful at the MC input of his phono stage. For a balanced-unbalanced converter this is not the optimum solution because we do not want a too low input impedance to hold the distortion in check. Do we raise the input impedance of that type of converter we get a noise disadvantage because the input resistors are in series with the signal. A more optimum solution is to put two buffers before the differential stage so we can make the input resistors on the differential opamp small. Fortunately we have very good quality buffers already on board of the Masterpiece so all we need is a relative low impedance differential Opamp. I found out that very good sound can be got from Opamps when they do not have the burden to amplify tiny signals like MC cartridges. I build good sounding MC stages with opamps too but non of them could compete in sound with a good discrete solution. Ones the signal is boosted though by a factor of say 100x the good points of an opamp come into play. More then two years ago i had the same problem and needed a balanced-unbalanced converter. First i wanted to use the Self Ultra Low Noise solution that puts a lot of buffers and differential stages in parallel to reduce the noise, much as we do at the input of the paradise.
I published the most extreme solution with 8 buffers and 4 diff amps PER channel. Scott Wurzer jumped in and persuaded me that this is not particular elegant and after some back and forth i build a converter with two B1 buffers per side and an AD797 as differential mixer. That worked pretty decent. The ADA797 is still an excellent part. Read Samuel Groner´s research about Opamps where the AD797 came out first and the LT1648 second compared to much more recent designs. We do not need additional buffers anyway and i would redesign the differential mixer by adding a class a buffer in the output. Then we are set.
Here is another design with discrete output stage by Bonsai :http://hifisonix.com/wordpress/wp-c...niversal-Small-Signal-Class-A-Buffer-V1.0.pdf
Here is a design that does all that : ClassA to 3mA, high output current capability ( 100mA ),
very fast also settling, low distortion. It has a floating ground plus servo. It is simple and the LME49713 can be ordered with a metal case.
I do not know if the low input impedance on the negative Pin of 58 Ohm creates any problems. This is a current feedback Opamp.
The CCSs can be Fet cascodes, see Dr.Evil´s design. There is an option to bias negative and positive and also switch the biasing off for comparison.
very fast also settling, low distortion. It has a floating ground plus servo. It is simple and the LME49713 can be ordered with a metal case.
I do not know if the low input impedance on the negative Pin of 58 Ohm creates any problems. This is a current feedback Opamp.
The CCSs can be Fet cascodes, see Dr.Evil´s design. There is an option to bias negative and positive and also switch the biasing off for comparison.
If we go with 2 x 2SA1085 plus 2 x 2SC2547 on inputs we are looking at about 0.3 nV/rtHz input noise at 1KHz plus 3dB penalty from balanced. Won't compromising this En at inputs with cabling and layout/shielding will be major challenge anyway so I personally can live with that additional 3dB of noise by simply taking one output. The real beauty of this masterpiece is that we already have 93dB common mode rejection at 50Hz to start with for bjt version. In sims 0.28mVpp common mode at inputs produces only 0.096mVpp at outputs, while differentially it gives 4.32Vpp, hence ratio of 44906 or ~93dB. I'm saying this loud as there is not much consistency describing CMR even in op amps datasheets. This is a really nice bonus we get for floating common emitters. Fet version numbers are sky rocking because of higher input independence and floating sources too. Ok, this is in sims, it will deteriorate in real build from mismatches but we get a picture, so se outputs are not so compromised as in case with two completely separate stages running balanced in. #5169 converter looks really good. That can be put on pcb too, making it recommended for bjt version and optional for those who want to have it all in fet version. Multi purpose pads for input bjt's is a good idea too, just 1 additional hole like in multimeter hfe measuring connector would do the job.
This are beautiful numbers ! Sure, we only get the benefit with heavy shielding and good cabling plus matching to 0.1%. With the ultra low Rbe`devices noise is not a problem. Thanks for approving on my balanced-unbalanced converter. The LME49713 is a very good sounding part.
I somebody does not like an opamp in the chain using only one output does the job too.
I somebody does not like an opamp in the chain using only one output does the job too.
Incredible specs from those LME49713 (1900V/us slew....1.9nV/ √ ^Hz (typ) noise)
I was wondering if the LME49713 has same hidden problems like the LT1028 for which the data sheet state a very low voltage noise but glide a bit over the poor current noise.
I am still paying full attention to the progress with the Paradise any updates / developments. When could we get a PCB for it?
About the Masterpiece and your earlier mention to same reference tracks
Masterpiece by the Temptations is a pretty good Album (STML11229)
Recorded in 1973 so free from too much electronic processing.
In one of the tracks the band is hand clapping and as this was done on a single take (everybody playing at same time) stage position of the band members is pretty well defined (It is actually so good that with decent stage one can tell how tall the band members are).
Maybe this could be a pretty good indicator of Phase shifts taking place on the Preamplifier that would compromise the tri dimensional stage.
The Temptations - Law Of The Land - YouTube
ON Ofset Iwas wondering if is possible to Offset the Offset by ofsetting the suply voltages.
I had a LAC powered MC stage and (bateries don't charge evenly) I was using a buffer to get the 0V virtual ground....(got stuck here I know a bit more about music than electronics)
Bksabath, the LME49713 is dead quiet in this application. We do not use it as the MC pre-pre input but at the output of a phono stage at line level. My current privat phono stage has that chip at the output and i have listened to it just yesterday. Yes, the LT1028 is best for low impedance use but of totally different design then the LME49713. There is also a time gap of over 20 years between this critters and technology has moved on.
Yes, i love the Temptations.
The DC offset problem will be solved.
Yes, i love the Temptations.
The DC offset problem will be solved.