A dmm is generally not good at measuring HF and certainly not VHF/UHF.
They cannot be relied on to indicate oscillation.
Output offset is a DC voltage that is fed to the load.
Output noise is an unwanted AC signal that is fed to the load.
Both should be low.
I recommend targets of <10mVdc and <0.3mVac.
The closer you can get to zero the better.
BTW,
0.3mVac only applies for normal sensitivity speakers. High sensitivity speakers need lower noise. Headphones need much lower noise.
They cannot be relied on to indicate oscillation.
Output offset is a DC voltage that is fed to the load.
Output noise is an unwanted AC signal that is fed to the load.
Both should be low.
I recommend targets of <10mVdc and <0.3mVac.
The closer you can get to zero the better.
BTW,
0.3mVac only applies for normal sensitivity speakers. High sensitivity speakers need lower noise. Headphones need much lower noise.
Oh linuxguru 's silence here kills me. I know he knows the fix (http://www.diyaudio.com/forums/chip-amps/54571-my-audiophile-lm3886-approach-69.html#post2842257). You lucky sir this is not Facebook or I'd have tagged you 
Painfull though, this is my second read of this 4300 post entire thread.
Painfull though, this is my second read of this 4300 post entire thread.
Ok. It was still better but now it's fried I guess. It was playing fine and I was playing it via a bulb tester. And then I restarted and now it's dead. When turning on the relay makes a sound, the speakers make a pop and then everything is silent. Offset was -18Vdc on both channels (they are on the same PCB). I guess we're at square one. I replaced with new TL071 and still the same.
That is a symptom not a cause.. Shorting them will only show other symptoms and the cause will remain.
Thanks for the reply Troy. Where do you suggest I start? I had hooked it up to an LM1036 circuit. The AGND of myref and the GND of LM1036 were separated by a 0.5uF cap.
Later I found that the LM1036 PCB was bad (tried it with my tested LA4440 amp and fixed the dry solder).
Now I've disconnected everything and the myref is on it's own. I'm just going to use it without a volume. My guess is that there was some start-up pop of the LM1036 or the discharge between the two grounds that broke things. Should I try to replace R35 and R11 or should I try something else? Your response is really appreciated.
Later I found that the LM1036 PCB was bad (tried it with my tested LA4440 amp and fixed the dry solder).
Now I've disconnected everything and the myref is on it's own. I'm just going to use it without a volume. My guess is that there was some start-up pop of the LM1036 or the discharge between the two grounds that broke things. Should I try to replace R35 and R11 or should I try something else? Your response is really appreciated.
I have no idea what your pre- board which drives the MyRefs is doing, or what its optimal ground arrangement should be, etc.
However, in a stock MyRef driven by a external source, R11 (the 1 ohm lift resistor) usually blows open when there is some kind of over-current situation in the LM3886, causing excess current to flow through its ground pin to the lifted ground, through R11 to PGND. The most common cause in my experience is plain human error during connection of the transformer leads - usually the accidental exchange of the AC1/AC2 and PGND leads from the trafo to the board.
In about 50% of such cases (maybe more), you can just replace R11, check the power connections carefully next time, and power up. With luck, nothing else would have blown.
There's not much impact on SQ from the slight differences in layout alone, between v1.2, 1.3 and 1.4 boards - they're all fairly close if you use exactly the same BoM.
The main difference is from the component flexibility and larger components that can be accommodated in the later boards. For instance, v1.4 can accommodate at 15mm pitch Wima FKP3 at C4, which is difficult or impossible in the other boards. The same applies to larger resistors in v1.4 - it allows the use of Holco H4 and some other physically large resistors at some locations, which is beneficial to sonics (not because those resistors are larger, but because they happen to be more linear/better).
Oddly, my favourite (and reference) build at the moment is a classic Rev C assembled on a v1.2 Twisted Pear board - but with a slightly tweaked parts list which uses some parts that were never on any published BoM at that time. However, I can replicate that exact BoM on the later boards if needed.
The main difference is from the component flexibility and larger components that can be accommodated in the later boards. For instance, v1.4 can accommodate at 15mm pitch Wima FKP3 at C4, which is difficult or impossible in the other boards. The same applies to larger resistors in v1.4 - it allows the use of Holco H4 and some other physically large resistors at some locations, which is beneficial to sonics (not because those resistors are larger, but because they happen to be more linear/better).
Oddly, my favourite (and reference) build at the moment is a classic Rev C assembled on a v1.2 Twisted Pear board - but with a slightly tweaked parts list which uses some parts that were never on any published BoM at that time. However, I can replicate that exact BoM on the later boards if needed.
Last edited:
Siva, where is the latest schematic for this?
I'm daunted by having to search through 400+ pages.
The current schematic is almost exactly the same as Mauro's c.2005 original Rev C (Rev3) schematic:
http://www.diyaudio.com/forums/chip-amps/54571-my-audiophile-lm3886-approach-10.html#post684359
Most of us here use the parts designations from Russ White's adaptations of Mauro's schematic for the Twisted Pear V1.0 to 1.2 PCB layout - the differences are small, and mainly involve a few additional bypass caps (which are no longer used). My V1.3 and V1.4 boards are adaptations of the Twisted Pear boards, and again have a few additional optional bypass capacitors (with new parts designations).
Some parts values in use have changed over the years as a result of experimentation. The ones I recollect are:
R2 = 10k..12k
C7 = 10nF
C21 = 22nF (optional)
apart from the compensation components C10, C32, C34, R42 and R43, for which the default Rev C values are:
C10=22pF
C32=150pF
C34=10pF
R42=47 ohms
R43=22k
The default Rev C compensation gives splendid sonics, so if you want something that just works, you can just stick with it. Everything after that is experimental, and there's no real consensus that any of the later mods are better-sounding across the board.
Note that the MyRef Rev C allows only one real opamp choice, the LM318. The TL071 and a few others will work, but with inferior sonics or transparency. I do have a few hybrid/discrete modules which consist of an LM318 and various Class-A output stages, which are generally a modest improvement over the stock LM318.
If you're big on opamp rolling, the Rev C is probably not for you - the MiniRef 1875 may be more to your liking.
Edit: The changes I made in parts selection for the Rev C (with the Ultimate BoM as the base) include the following:
C4: 100nF/100V/1% ERO KP1830
C5: 100nF/250V Wima MKS3
R7: 12.1k/1% PRP GP1/4
R10: 347R Takman REY25
R12: 2k7/0.5W/5% Allen-Bradley Carbon Composition
C9: 220uF/4V Black Gate PK
C13: 1.0uF/100V EFC Wesco SP1213TX MKP-type axial
R3: 0.47R/5W Futaba MPC74
ZD1, ZD2: On Semi 1N5349B 12V/5W Zener
and maybe a few other other non-critical changes.
Last edited:
I just wanted to humbly make available my single side "REV C" monobloc design which is suitable for REV-A/B/C (with a little simple re-arrangement).
PDFs
The PDFs are suitable for the toner transfer method as the copper side is not mirrored, and the component side is.
Please enjoy it if you are capable of etching your own boards.
If anyone makes a good amp with it then this journey has been worth it for me.
Thanks to everyone who has made this a lot of fun (and very educational) for me. The results are largely thanks to you. Thank you Mauro most of all.
Cheers!
Russ
EDIT: fixed the link.
Someone have this pdf?
I would suggest you to check the Rev. E / FE freemen edition instead...
Hi.
Do you have the link of the thread?
Thanks!