R11 visually seems ok on both boards. Checked it with multimeter and i get 1.8ohms and 1.3ohms in each board respectively.
OK, R11 seems not to be the problem, but if you want to be doubly sure, just measure and select one closer to 1 ohm and replace the one that reads 1.8 ohms.
The LM3886 that's heating up may have bitten the dust, but it's best to check the MUR8xx, zeners and LM318 first before attempting to replace the chipamp. If you have to replace the chipamp, cut out all the leads and desolder each lead individually instead of trying to wiggle it out. This will sacrifice (the presumably dead) chipamp, but save the PCB pads, which is more important.
The motor current is not steady.
A simple Zener regulated supply with a resistor CCS will not drive the motor well, particularly during start up.
Better to use a real regulator with a current capacity well in excess of three times the nominal motor current.
An lm7812 would certainly meet the >=0.45A, for start up condition.
A simple Zener regulated supply with a resistor CCS will not drive the motor well, particularly during start up.
Better to use a real regulator with a current capacity well in excess of three times the nominal motor current.
An lm7812 would certainly meet the >=0.45A, for start up condition.
Hi AndrewT, thanks for the reply, the max input voltage of lm7812 is 35v and my source voltage is also 35 volts. do you think lm7812 will survive for the long run with proper heatsinking?
i will use the fan for cooling the heatsink for lm3886 because my transformer has no 12vac secondary.
Thank Andrew
sorry for the out of topic question Mauro.
i will use the fan for cooling the heatsink for lm3886 because my transformer has no 12vac secondary.
Thank Andrew
sorry for the out of topic question Mauro.
For whom interested
I've started checking interest for a third run of My_Ref Fremen Edition GB:
http://www.diyaudio.com/forums/grou...dition-gb-interest-third-run.html#post3640615
I've started checking interest for a third run of My_Ref Fremen Edition GB:
http://www.diyaudio.com/forums/grou...dition-gb-interest-third-run.html#post3640615
Link to the latest fully-discrete DIP8 opamp alternative for the MyRef Rev C or derivatives:
http://www.diyaudio.com/forums/analog-line-level/207871-lf03-discrete-opamp-2.html#post3644366
This is derived from the LF03c dual which I developed over a year ago using through-hole parts. The SMD single (which is a pure plug-in replacement for the LM318 in a MyRef Rev C with no other alterations) is the LF03s.
I've tested it in a Rev C with a modified HifiMeDIY Sabre ES9023 24/96 DAC as the digital source playing FLACs ripped directly from CD, and it essentially makes the LM318 obsolete on at least 3 audible parameters: definition, detail and transparency - while remaining completely stable with no compensation changes required or expected on almost any Rev C, Rev E or Rev C with FE compensation implementations. The detail is stunning on certain tracks - one can hear the intake of breath by vocalists and players of various wind instruments which were inaudible earlier. The downside: it is extremely revealing, and deficiencies in the recording chain can make certain tracks unlistenable, which may have been glossed over with the LM318.
The first LF03s prototypes have been built with commodity semis and generic 1% SMD resistors, and they're already a game-changer. I'll offer the present build shortly in the vendors section, and consider a premium version if there's sufficient interest. The LF03s board will also accommodate alternate BoMs, including low-noise Japanese semis and thin-film resistors, which are presently unobtainium in volume (except maybe from EBay).
Link to my post in the vendor's forums offering the LF03s and the LF03d (dual DIP8):
http://www.diyaudio.com/forums/vend...p8-opamp-myref-rev-c-similar.html#post3644382
http://www.diyaudio.com/forums/analog-line-level/207871-lf03-discrete-opamp-2.html#post3644366
This is derived from the LF03c dual which I developed over a year ago using through-hole parts. The SMD single (which is a pure plug-in replacement for the LM318 in a MyRef Rev C with no other alterations) is the LF03s.
I've tested it in a Rev C with a modified HifiMeDIY Sabre ES9023 24/96 DAC as the digital source playing FLACs ripped directly from CD, and it essentially makes the LM318 obsolete on at least 3 audible parameters: definition, detail and transparency - while remaining completely stable with no compensation changes required or expected on almost any Rev C, Rev E or Rev C with FE compensation implementations. The detail is stunning on certain tracks - one can hear the intake of breath by vocalists and players of various wind instruments which were inaudible earlier. The downside: it is extremely revealing, and deficiencies in the recording chain can make certain tracks unlistenable, which may have been glossed over with the LM318.
The first LF03s prototypes have been built with commodity semis and generic 1% SMD resistors, and they're already a game-changer. I'll offer the present build shortly in the vendors section, and consider a premium version if there's sufficient interest. The LF03s board will also accommodate alternate BoMs, including low-noise Japanese semis and thin-film resistors, which are presently unobtainium in volume (except maybe from EBay).
Link to my post in the vendor's forums offering the LF03s and the LF03d (dual DIP8):
http://www.diyaudio.com/forums/vend...p8-opamp-myref-rev-c-similar.html#post3644382
Last edited:
if space and money are no problem, I always use several voltage regulators to keep the delta t (temp) down as much as possible. That saves on a big heatsink too.Let's say 24V 18V and then 12V. A bit much but it works very well. In today's time the regulators are dead cheap
Sorry - I have to retract that. It's not yet ready for prime time in a MyRef Rev C - the LF03s is only marginally stable in a MyRef Rev C. It's not a direct plug-in replacement for the LM318, LF01 or LM318+LF07 in a MyRef Rev C for now.
It sounds just great, but the LM3886 heats up in at least one channel, indicating marginal stability. So it joins a long list of high-performance opamps, including the OPA627, AD797, LME49710, LME49990, etc., that are unstable in a MyRef Rev C.
The LF03s is still perfectly usable as an upgrade to most monolithic bipolar-input DIP8 single opamps for most small-signal and line-level applications.
The LF03s is a Class-A fully-discrete topology. Most of the details are in the LF03 thread, but here are the key differences:
1) Input LTP in the LM318 consists of two darlingtons, while the LTP in the LF03 is a low-noise NPN pair. Darlingtons are not great on both noise and distortion, which is one of the reasons why some Musical Fidelity preamps that use the LM318 bypass the internal darlington LTP and use external discrete low-noise NPNs with the collectors connected to the offset null terminals (pins 1 & 5) of the LM318.
2) The non-inverting and inverting signal paths in the LF03 are almost exactly matched right until the output stage. This allows the distortions in the forward and feedback paths to cancel each other to the extent possible.
3) The output-stage darlington in the LF03 runs at nearly constant current, which is not the case in the LM318.
Of course, all the other advantages of discrete opamps also hold - very linear resistors (thin-films are trivial to use), excellent C0G/NP0 compensation caps, high quality semis of all kinds, output-stage thermal isolation from the small-signal stages, higher output-drive currents from SOT89 output stage, etc.
The LM318 wins on (asymmetric) slew-rate (50V/usec) and GBW (15 MHz), but I don't believe either is hugely important to sonics beyond a point; the LF03s has respectable GBW in the 8-10 MHz range. Most faster opamps aren't stable in the MyRef, and those that are stable (TL071) don't necessarily have better sonics than the LM318.
The LF03s is stable (after tweaking the internal dominant pole caps yesterday) in the MyRef now, and has better definition, micro-detail and transparency - all without losing Class-A character.
This is a follow report from yesterday' "stop I have a problem Houston".
Do you now have two versions? One for General purpose use and one for MyRef?
Or is the new compensation going to be universally adopted for all versions?
Yes, there are two different prototypes - the version which is stable in the MyRef has 2x 47pF Miller caps in the VAS. The first version had 2x 27 pF, and about 3 dB lower THD20 in simulation (shouldn't really make an audible difference - it's almost all H2).
I'll do some further comparative auditioning between the 27 pF and 47pF prototypes in a line amp or similar, and standardize on the 47 pF version if they're audibly indistinguishable (which seems to be my preliminary audible impression).
There's a minor wrinkle that may cause a lead time of a couple of weeks - the Kyocera 47pF C0Gs that I already have are 0805 size, which I can hand-assemble on the 0603 pads on the board, but I'd prefer to use 0603 47 pF C0Gs, which are on their way.
Q: Is a 200V-rated C0G likely to be more linear than a 50V-rated C0G, for use as a 47pF Miller cap? Or is it immaterial, given that the few pF of junction capacitance that it shunts is highly non-linear with Vcb anyway? Are there any SMD film-cap alternatives in 0603 (the Panasonic ECHU seems to stop at 0805 size) ?
I am wondering, it seems the phase characteristic is an important factor. Lots of higher bandwidth op amps have some funny phase characteristics. Probably as close as possible to an ideal integrator means a stable MyRef operation? The constant current aspect seems interesting because this is an advantage, however, I am wondering how constant it is comparing against the LM318.
I kind of worry when people use "class A sound" description. I have ran a comparison of my tweaked MyRef against a class A amp, and my impression is that it beat the class A amp, in terms of revealing the emotion in the music giving a more perception of live performance. Most people describe class A as having a clean sound, the class A amps I have listed to have a more uniform coloration spectrum, which also was my impression with the original MyRef.
I kind of worry when people use "class A sound" description. I have ran a comparison of my tweaked MyRef against a class A amp, and my impression is that it beat the class A amp, in terms of revealing the emotion in the music giving a more perception of live performance. Most people describe class A as having a clean sound, the class A amps I have listed to have a more uniform coloration spectrum, which also was my impression with the original MyRef.
hi soongsc, would it be possible to ask you for the list of changes you made to your myref? sorry, it's too time consuming to go through the thread looking for what you might have already mentioned.
having test built siva's miniref, i too feel that myref is colored in comparison. not only is miniref more transparent, it's also more neutral in it's distortion profile. (you can also adjust it's character with op-amp swapping which is nice)
needless to say though, they are both world class designs.
having test built siva's miniref, i too feel that myref is colored in comparison. not only is miniref more transparent, it's also more neutral in it's distortion profile. (you can also adjust it's character with op-amp swapping which is nice)
needless to say though, they are both world class designs.
The comparison was quite a long time ago, and I have not played it for some time. The last time a tweaked it was changing the current gain to be more compatible with a 6 ohm load. This made the sound even cleaner with a speaker of 6 ohm load, and probably more to tolerate 4~8 ohm variation range.
Other major changes that I can remember was different power supply for the high voltage rail, removed the input capacitor, modified the input so that the input impedance curve shape is flat in the audio range to more closely match my interconnect.
I have also worked on a chip amp which in many ways is as clean as the tweaked MyRef. Routing of the return and reference grounds were critical to accomplishing that.
Other major changes that I can remember was different power supply for the high voltage rail, removed the input capacitor, modified the input so that the input impedance curve shape is flat in the audio range to more closely match my interconnect.
I have also worked on a chip amp which in many ways is as clean as the tweaked MyRef. Routing of the return and reference grounds were critical to accomplishing that.
Last edited:
Thanks - I'm just waiting to try a discrete opamp like the LF03c or LF03d in the MiniRef 3886 prototype that I have. It should also work in your MiniRef 1875 prototype, but be prepared to tweak C10/C20 upwards to maintain stability - I already have 100pF (way higher than the nominal 22pF indicated on the MiniRef 1875 schematic), so it should not be a problem, but we'll see how it goes. 33pF FKP2 would be a good starting point for C10/C20 on a MiniRef 1875, especially if you want to roll some fast or marginally-stable opamps.
