I have made a mistake. Instead of 50 nanovolts and 1000 nv should be ampers as a unit for current.
Input offset current and input bias current are not the same thing.
Bksabath, thanks for the capacitor list. I would say that Wima and Röderstein will fit in there too.
The Peak Atlas is actually not so bad and it will show you if the components in the right and left channel are the same provided the Peak Atlas makes a "consistent" mistake.
To check the accuracy of the Peak Atlas, buy a 1% capacitor, measure and interpolate the measurement mistake.
I already talked about how i expect the noise of the Paradise will sound. Expressed as a noise impedance it has the noise of a 8 Ohm resistor. This is very close to the physical limit of 5 Ohm. So that is a bit better then 0.4nV/qHz. Bipolars have much lower 1/F noise corner then Fets so they are cleaner in the bass. A Fet stage with the same 0.4nV/qHz spec would be more noisy under 1kHz. There are other noise sources like popcorn and flicker noise. This is hard to estimate because it depends on the quality of the wafer process. For example you could take 2 BC550C from two different processes used by two different manufacturers and they would both hold the spec for a BC550C but one of them is more noisy. The know how to find the right parts in one of the biggest trade secrets of a successful Audio company. Our only option is to build it and then look if there is any noise problem with transistors from a certain manufacturer.
Concerning balanced, our member Sampler is working on a balanced version of the Starless. When you go back on the thread you see some work he has done on that. He promised a complete schematic when the work is done.
The Peak Atlas is actually not so bad and it will show you if the components in the right and left channel are the same provided the Peak Atlas makes a "consistent" mistake.
To check the accuracy of the Peak Atlas, buy a 1% capacitor, measure and interpolate the measurement mistake.
I already talked about how i expect the noise of the Paradise will sound. Expressed as a noise impedance it has the noise of a 8 Ohm resistor. This is very close to the physical limit of 5 Ohm. So that is a bit better then 0.4nV/qHz. Bipolars have much lower 1/F noise corner then Fets so they are cleaner in the bass. A Fet stage with the same 0.4nV/qHz spec would be more noisy under 1kHz. There are other noise sources like popcorn and flicker noise. This is hard to estimate because it depends on the quality of the wafer process. For example you could take 2 BC550C from two different processes used by two different manufacturers and they would both hold the spec for a BC550C but one of them is more noisy. The know how to find the right parts in one of the biggest trade secrets of a successful Audio company. Our only option is to build it and then look if there is any noise problem with transistors from a certain manufacturer.
Concerning balanced, our member Sampler is working on a balanced version of the Starless. When you go back on the thread you see some work he has done on that. He promised a complete schematic when the work is done.
I'm not sure if this will help but here it goes, using the link you find all SiC's at Mouser sorted by price (lowest first).
http://ca.mouser.com/Semiconductors...z63xZ1z0y4ci&Keyword=sic&Ns=Pricing|0&FS=True
Starting at 1.11$
Regards,
Frans.
Last edited:
Tanks Holger I have seen the error of my way and whitout a further do it is rectified and I stand corrected.
Tanks Joakim
Atlas thing fine by me I had one for a while so no problem there.
Fairchild manufacturer is fine as well Just seem to me that Farnell just empty and refill the bin whit whatever they have and batches are not so good as stuff is mixed up over and over.
Tanks FdW That SIC from Farnell is not longher in production from manufacturer...
no problem there I have used TO 220 with Anode on Pin 1 Left ( 50 50 chance is wrong so most probably wrong )
Space alowd for Dides to be turned over if anode on pin 2.
Here is the last one
Would TO 247 2 pin foot print for diodes be a good thing? (can be done in couple of minutes).
Top layer Ground plane is not in this drawing (again a couple of minutes work as shown before)
do you need it at this stage?
I have kept capacitor size comon troug out board but not used the 5 mm
trak sizes can then be much wider as C18 and C22 are not in the way.
On the sound
I am sorry if I gave the impression of causing waves and I am more than hapy if Paradise can achieve such low noise.
I was just commenting on the sound of the groves whitout music or if you prefer the very low passages and was wondering if choosing Mica may improve or make worst the caracter of this.
As many same of my records are prety vorn out I have noticed that same capacitors sound too bright.
Tanks Joakim
Atlas thing fine by me I had one for a while so no problem there.
Fairchild manufacturer is fine as well Just seem to me that Farnell just empty and refill the bin whit whatever they have and batches are not so good as stuff is mixed up over and over.
Tanks FdW That SIC from Farnell is not longher in production from manufacturer...
no problem there I have used TO 220 with Anode on Pin 1 Left ( 50 50 chance is wrong so most probably wrong )
Space alowd for Dides to be turned over if anode on pin 2.
Here is the last one
Would TO 247 2 pin foot print for diodes be a good thing? (can be done in couple of minutes).
Top layer Ground plane is not in this drawing (again a couple of minutes work as shown before)
do you need it at this stage?
I have kept capacitor size comon troug out board but not used the 5 mm
trak sizes can then be much wider as C18 and C22 are not in the way.
On the sound
I am sorry if I gave the impression of causing waves and I am more than hapy if Paradise can achieve such low noise.
I was just commenting on the sound of the groves whitout music or if you prefer the very low passages and was wondering if choosing Mica may improve or make worst the caracter of this.
As many same of my records are prety vorn out I have noticed that same capacitors sound too bright.
Last edited:
I have used silver Mica in RIAA stages. Not all Micas are created the same though. For Example i found the CD caps a little bright sounding, maybe because they have heavy steel legs. The Micas from Reichelt have my blessing and i heard good things about the mica that RS sells under their own brand.
Here is a 6.8nF Mica i can recommend. It is expensive.Bestellen Sie Glimmer Kondensator 6800pF RS D206800/500V online mit versandkostenfreier Lieferung bei RS Components.
These one work great. I use them even in the signal path to my tube amplifiers. I use them in phono stages and they are great for miller compensation in power amps.CY 22-5 6,8N - Glimmer-Kondensator, 6,8nF, 100V, + - - 1 bei reichelt elektronik
Try the FFPF04N60S2 (the Stealth II diode), yes it is 600V but it has 15nC of reverse recovery... very quiet....
Tanks Joachim
Pointer to RS now I will get same.
reichelt electronics very tempting may get same as well.
Tanks Hesener
One more from my wish list.
Any tought on the Pre-Regulator board
Time for same sleep before tonight 12 hour shift
Board will go in the Bath tomorow PM so catch you later
Pointer to RS now I will get same.
reichelt electronics very tempting may get same as well.
Tanks Hesener
One more from my wish list.
Any tought on the Pre-Regulator board
Time for same sleep before tonight 12 hour shift
Board will go in the Bath tomorow PM so catch you later
Talk to Frans about the board. I think you and him should work on this together and he has already expressed that he likes your approach. This will be a useful module for many projects. I like it.
Here is an interesting article about what frequency response is :A Simple Frequency Response Function : Noise & Vibration Measurement Blog
Here is an interesting article about what frequency response is :A Simple Frequency Response Function : Noise & Vibration Measurement Blog
Try micas from Richard Jahre. These are awesome. A bit pricy, but worth the money. Even available in .5% tollerance.
I once fried one channel of a DL103. Over night, one regulator burned in my phonoclone. It must have been some hours that the inputs were at 0.6V.
One Channel survived!
Rüdiger