Sapphire3, latest build.
C2,3 are ceramic this time, instead of mica.
C9,10 are left out.
I caved and removed the jacket from the electrytics. It's mostly cosmetic (and not all that safe!) but its supposed to sound better.
C1 is 0.33 Multicap PPFXS. They are the ones I used on the Sapphire 2.0 for the last couple of years, so are nicely aged. I find .33 works a bit better with the HD600s than .47: yes, there is less ultra-low impact but the musical balance is better overall. This evaluation will, of course, be headphone dependent.
I must say I'm sold on this whole "match your transistors" thing, even if I did not, technically, match anything .. I just discarded outliers. The improvement, at any rate, over the previous build is substantial but hard to describe. Increased soundstage precision? Yes, but that's not really the main thing. Vanishing levels of grain/glare? Maybe that's closer to it. It's in the way the amp slots away each musical line so gracefully now, softly yet with absolute authority. Previous iterations of the Sapphire, while clean and powerful, never quite reached this level of refinement.
C2,3 are ceramic this time, instead of mica.
C9,10 are left out.
I caved and removed the jacket from the electrytics. It's mostly cosmetic (and not all that safe!) but its supposed to sound better.
C1 is 0.33 Multicap PPFXS. They are the ones I used on the Sapphire 2.0 for the last couple of years, so are nicely aged. I find .33 works a bit better with the HD600s than .47: yes, there is less ultra-low impact but the musical balance is better overall. This evaluation will, of course, be headphone dependent.
I must say I'm sold on this whole "match your transistors" thing, even if I did not, technically, match anything .. I just discarded outliers. The improvement, at any rate, over the previous build is substantial but hard to describe. Increased soundstage precision? Yes, but that's not really the main thing. Vanishing levels of grain/glare? Maybe that's closer to it. It's in the way the amp slots away each musical line so gracefully now, softly yet with absolute authority. Previous iterations of the Sapphire, while clean and powerful, never quite reached this level of refinement.
Attachments
Last edited:
I second that suggestion.
Taking it apart results in very small springs falling out and never getting back in the right place. I speak from experience and I use a LOT of those switches. Additionally I would add one more resistor from output to ground of maybe 500k so that when it starts to NOT make before break you dont have an open circuit to ground. Otherwise I like them for the price.
Taking it apart results in very small springs falling out and never getting back in the right place. I speak from experience and I use a LOT of those switches. Additionally I would add one more resistor from output to ground of maybe 500k so that when it starts to NOT make before break you dont have an open circuit to ground. Otherwise I like them for the price.
Housekeeping stuff:
1. I have a spare chassis (HLLY) if anyone wants one. It's exactly the right size for the Sapphire, I should know 'cus that's what I put mine in. New, unused.(photos here)
2. I have two sets of Sapphire 2.0 boards left, which can also be made available as a kit. I need the spot on my shelf to put the stack of Sapphire 3 boards. Clearance! ($20/$60 kit)
3. Sapphire 3.0 boards/kits will be half price if you are upgrading from a previous version. ($20/$60 kit) I will bin/screen the transistors for you.
1. I have a spare chassis (HLLY) if anyone wants one. It's exactly the right size for the Sapphire, I should know 'cus that's what I put mine in. New, unused.(photos here)
2. I have two sets of Sapphire 2.0 boards left, which can also be made available as a kit. I need the spot on my shelf to put the stack of Sapphire 3 boards. Clearance! ($20/$60 kit)
3. Sapphire 3.0 boards/kits will be half price if you are upgrading from a previous version. ($20/$60 kit) I will bin/screen the transistors for you.
Last edited:
Passed the smoke test and my voltages are within a gnat's breath of the values in the instructions. Input offset is exactly zero on both channels. Added resistors 19 and 20 as per the modification described in posting 324; I have some low level spitty noise, but I'm burning this in beside my computer (and plugged into my UPS with all its wall warts) which is asking for trouble.
Now to get some hours on this baby.
Now to get some hours on this baby.
Attachments
Nice build! Feel free to post more pics, as I think others could find the example useful.
I'll take a stab at the source of the spitty noise, as I use the same case, also placed close to a desktop PC: the problem is the black paint/coating on the metal doesn't conduct electricity, so you have to file away the paint both at the ground lugs and around the area where the case parts join together, particularly the front and rear sections.
That worked for me at any rate. The remaining noise source, I have found, is my volume control which is an open frame stepped attenuator. It needs to be in a metal can shield of its own, something I haven't bothered with yet.
I'll take a stab at the source of the spitty noise, as I use the same case, also placed close to a desktop PC: the problem is the black paint/coating on the metal doesn't conduct electricity, so you have to file away the paint both at the ground lugs and around the area where the case parts join together, particularly the front and rear sections.
That worked for me at any rate. The remaining noise source, I have found, is my volume control which is an open frame stepped attenuator. It needs to be in a metal can shield of its own, something I haven't bothered with yet.
The low level spitty noise is in both channels and does not seem to attenuate with the volume control until about halfway round. My RCA jacks are isolated from the back panel; the headphone jack is isolated from the front panel. My RCA cables are well-shielded, connected back to an Audioengine DAC which runs quietly, its headphone out is spitty-noise free.
My volume control is a stepped 50k unit with Takman carbon film resistors. I did not buy the shield for it - I suppose I shall, but I wonder if there is room for it. Its ground is connected only to the - connections on pcb ins and outs. Should I ground the volume pot also to a chassis ground post? I've two ground posts on the bottom chassis, both were filed to bare white aluminum and I've tested ground continuity from one to the other with my meter. The posts are stainless steel bolts.
My wiring pairs are all tightly twisted; and my panel switch wiring is twisted and isolated by stainless steel braid that is connected to chassis ground.
I'm very pleased with the sound, spitty noise aside. The first half hour it sounded pretty rough, but has been opening up nicely with detail and soundstage. I used a .33uF multicap as input coupler, thinking about getting some more exotic types to try (Parts connexion has them on sale!).
I made another change in addition to resistors 19 and 20; I figured since I'm going to put the 3.0 boards in later, that I'd try MF resistors in place of the Kamaya CC, partly because as I was building the boards I realized that the through-holes are quite tight, and replacing the resistors later was going to be difficult. I've hunted high and low for info on induction in MF resistors and can't find anything that says I shouldn't try them (and I've used them in other circuits), so in they went. I'm very happy with what I'm hearing, but it's all a voyage of discovery.
My volume control is a stepped 50k unit with Takman carbon film resistors. I did not buy the shield for it - I suppose I shall, but I wonder if there is room for it. Its ground is connected only to the - connections on pcb ins and outs. Should I ground the volume pot also to a chassis ground post? I've two ground posts on the bottom chassis, both were filed to bare white aluminum and I've tested ground continuity from one to the other with my meter. The posts are stainless steel bolts.
My wiring pairs are all tightly twisted; and my panel switch wiring is twisted and isolated by stainless steel braid that is connected to chassis ground.
I'm very pleased with the sound, spitty noise aside. The first half hour it sounded pretty rough, but has been opening up nicely with detail and soundstage. I used a .33uF multicap as input coupler, thinking about getting some more exotic types to try (Parts connexion has them on sale!).
I made another change in addition to resistors 19 and 20; I figured since I'm going to put the 3.0 boards in later, that I'd try MF resistors in place of the Kamaya CC, partly because as I was building the boards I realized that the through-holes are quite tight, and replacing the resistors later was going to be difficult. I've hunted high and low for info on induction in MF resistors and can't find anything that says I shouldn't try them (and I've used them in other circuits), so in they went. I'm very happy with what I'm hearing, but it's all a voyage of discovery.
@udailey
No I haven't and the different pin out (ECB vs CBE) makes the connection a bit of a fuss unless you re-spin the board layout.
@Stan
It's very unlikely metal film vs. carbon will make any difference whatsoever. Actually I'm moving some of the parts (R17,18,19,20) from carbon comp to metal film in the Sapphire3 BOM since the accuracy is more reliable.
The noise you hear is explained as follows: when you turn up the volume control and leave the input disconnected from the source, all the input wiring from the opamp in+ pin out to the RCA jacks floats up to about 50kohms referenced to the chassis/GND/COM.
The op amp's current noise is converted to a voltage signal across that 50kohms, generating additional output noise. In addition, the sections of wire at 50kohms (and the volume control generally) are prone to any kind of noise pickup, capacitive or inductive, again in proportion to the impedance between that bit of metal and the chassis ground. Basically, all amps will do this to a lesser or greater extent, though it will be a lot less when 10k volume control is used for example, and will also be less if the op amp is configured for lower gain. It may likely vary with the op amp used (NE5534 vs OPA134). The Sapphire3 has a relatively high impedance input section and high-ish gain, so may be slightly worse in this respect than some.
The point is though that with a source component actually attached, i.e. the practical listening configuration, the problem goes away because the input impedance is all shorted though the output impedance of the source, so 50 kohms turns into ~0 ohms, and its just the worse case (about 12kohms) when the volume control is turned up to 70-80%.
If it really, really bothers you, then use a 10k volume control.
All that said, I'm not 100% sure on a case-by-case basis whether the noise you hear originates from the op amp input itself (intrinsic, you can't do anything about it short of using an op amp with less current noise) or is picked up extraneously. If the latter, carefully shielding all the input section may improve matters.
No I haven't and the different pin out (ECB vs CBE) makes the connection a bit of a fuss unless you re-spin the board layout.
@Stan
It's very unlikely metal film vs. carbon will make any difference whatsoever. Actually I'm moving some of the parts (R17,18,19,20) from carbon comp to metal film in the Sapphire3 BOM since the accuracy is more reliable.
The noise you hear is explained as follows: when you turn up the volume control and leave the input disconnected from the source, all the input wiring from the opamp in+ pin out to the RCA jacks floats up to about 50kohms referenced to the chassis/GND/COM.
The op amp's current noise is converted to a voltage signal across that 50kohms, generating additional output noise. In addition, the sections of wire at 50kohms (and the volume control generally) are prone to any kind of noise pickup, capacitive or inductive, again in proportion to the impedance between that bit of metal and the chassis ground. Basically, all amps will do this to a lesser or greater extent, though it will be a lot less when 10k volume control is used for example, and will also be less if the op amp is configured for lower gain. It may likely vary with the op amp used (NE5534 vs OPA134). The Sapphire3 has a relatively high impedance input section and high-ish gain, so may be slightly worse in this respect than some.
The point is though that with a source component actually attached, i.e. the practical listening configuration, the problem goes away because the input impedance is all shorted though the output impedance of the source, so 50 kohms turns into ~0 ohms, and its just the worse case (about 12kohms) when the volume control is turned up to 70-80%.
If it really, really bothers you, then use a 10k volume control.
All that said, I'm not 100% sure on a case-by-case basis whether the noise you hear originates from the op amp input itself (intrinsic, you can't do anything about it short of using an op amp with less current noise) or is picked up extraneously. If the latter, carefully shielding all the input section may improve matters.
Last edited:
Low level spitty noise cured
I'm not sure what did it, because I did a number of things.
Took the amp cover off and shortened the wires to the LED, also twisted them and connected them at the bridge rectifier instead of at the circuit board.
Bunched the power connections at the circuit board together to get them as far from the electrolytic caps as possible.
Added a ground connection from the headphone jack to chassis ground.
Gently pried the input leads away to stand off from the stepped volume pot, as I read that these are sensitive and can pick up schmutz.
Moved what other wiring I could further away from any possible sources of EM interference.
Filed the anodization off the case corners to connect ground throughout the case.
Found a shielded power cord, swapped that into use.
Moved the amp to another room but now it's back beside my computer, though a little further away from the modem and wireless hub.
Now dead quiet. Sounds great. And Sapphire 3.0 sounds better? I can hardly wait.
I'm not sure what did it, because I did a number of things.
Took the amp cover off and shortened the wires to the LED, also twisted them and connected them at the bridge rectifier instead of at the circuit board.
Bunched the power connections at the circuit board together to get them as far from the electrolytic caps as possible.
Added a ground connection from the headphone jack to chassis ground.
Gently pried the input leads away to stand off from the stepped volume pot, as I read that these are sensitive and can pick up schmutz.
Moved what other wiring I could further away from any possible sources of EM interference.
Filed the anodization off the case corners to connect ground throughout the case.
Found a shielded power cord, swapped that into use.
Moved the amp to another room but now it's back beside my computer, though a little further away from the modem and wireless hub.
Now dead quiet. Sounds great. And Sapphire 3.0 sounds better? I can hardly wait.
