Group Buy: Mauro Penasa MyRef Rev C kit

[jls001]

@sashijoseph: thanks for the pointers.

Last night I rechecked the supply voltage to the LM3886. I think I've been measuring it wrong all along. I used to measure pin 5 (V+ input) referenced to speaker negative and it showed zero. I measured pin 5 with reference to pin 7 (Ground) and it showed around +62V. Since pin 5 is shorted to pin 1 as per the circuit schematic, I checked pin 1 wrt 7 and the same +62V shows up there too. So I guess this is the reason why the chip is not sending an output signal to trigger the speaker relay since there is about 20V more voltage than the absolute continuous maximum of 42V.

Now the question is how the hell is the rail voltage almost doubled?

I didn't dare check the V- pin 4 yet, but I will devise a way to check it safely without shorting the multimeter probe to pin 5.

@quad: thanks for the link to Linuxguru's schematic. I installed Eagle and could peruse the correct schematic now. Hopefully it will translate to better troubleshooting.

 

[doors666]

you can measure v+ and v- at the rectifier diode pins also. This is the first thing you do, is the supply voltage ok? till it is, there is no point checking anything else.
the ground and speaker- should be the same as the output is ground referenced and not floating.

 

[Audiodoc]

I believe there is some negative rail voltage leaking into the ground plane which will make the voltage between positive pins 1 or 5 and ground pin 7 almost double.

You can turn of the amp and then check continuity between negative pin 4 and ground 5 if there is any short circuit resistance about 0 ohm.

Do remember to discharge the capacitors to prevent any sparking or use tape to insulate mm probes except the tip.

Change 22pf capacitor to 100pf

 

[jls001]

Please check for continuity between pin7 of 3886 and Gnd.

Just noticed R11(1 ohm) which connects GND to pin7.

If this resistor is open,it would explain the mysterious disappearance of +/- 35v on the 3886 pins.

Sir, you hit the bullseye :thumbsup:.

I removed R11 and found that it's open. Replaced with a spare 3W one and it's singing now.

Changed the other channel too and that's also working now.

It's singing

 

[Audiodoc]

Congrats. Most of the times it is a small component failure and a heck lot of troubleshooting. BUT in the end it is a learning experience

 

[jls001]

Thanks Audiodoc.

You've been exceptionally helpful. You are a blessing to the DIY folks :thumbsup:

 

[manniraj]

Thanks Audiodoc.

You've been exceptionally helpful. You are a blessing to the DIY folks :thumbsup:

That's why his handle is 'audiodoc' :clapping:

 

[jls001]

How does one use the Safety Ground tag with the CL60?

Is CL60 a thermistor? What's the ohm and power rating needed in this position?

I'm having slight channel imbalance, right being a bit louder, but I think that's coming from my buffer and not due to MyRef. Also the right channel is still working with the 6800 uF power supply caps I had used instead of the 10K Nippon Chemicon caps. I will swap those out. I'll also try with the original LM318 opamps. For now I'm still using the TLO71s on both channels. There is a slight hardness to the sound, but that could also be because the amp needs some burn in. Otherwise, what I'm hearing is nice, tight bass, very sweet midrange though tinged with a slight hard edge, and sweet highs. Detail resolution is superb. Very musical without being rolled off. I've got a "new" ECC86 tube-MOSFET Erno Borbely hybrid preamp to try it with, once I can clean up the scratchy pot. This preamp is a gift from my friend IndianEars :thumbsup: Many, many thanks for the gift, DC.

 

[Audiodoc]

That is probably at the buffer stage. What sort of pot are you using.

The CL60 is optional. Helpful when ground loops cause a hum. The myref is silent in most cases. I use the alps blue velvet but you can try cheaper pots from alps or a stepped attenuator from Hong Kong ebay.

Seller name is along1980.

The 318 will sound better. TRY them if you have problems with stability you just have to replace the compensation capacitor from 22p to 47pf or 100pf. This was what cured my amp modules.

 

[jls001]

I have a very, very faint hum audible when ear is literally pressed to driver. With another interconnect from buffer to MyRef, hum was much louder and varied when I changed the position of the IC. I used another IC and the net result is the very, very faint hum that I'm hearing now. I'll check if I still have my Gotham GAC2 IC which I remember is double shielded.

I'm using the HK eBay 21-step attenuator. It's good, better than the ALPS RK27 Bluevelvet I was using earlier. I am going to try another stepped pot bought from theaudiocrafts, Delhi.

I inserted the LM318 (branded "JRC") and it is working fine. It's starting up normally. I think the hard edge is gone too. Later I will swap to the TI LM318s.

For electrical safety the only thing I've implemented so far on MyRef is to ground the earth to the chassis. On other projects I've tried rectifier bridge-thermistor combo.

 

[hifiramr]

The hum and noise is common to all when we don't care about the wiring.Please try to do a star ground topology.That will shutdown the hum issue if this will not work try ground loop breaker.The noise may be because of several reasons that begins from input.

Earthing (Grounding) Your Hi-Fi - Tricks and Techniques

 

[jls001]

The hum and noise is common to all when we don't care about the wiring.Please try to do a star ground topology.That will shutdown the hum issue if this will not work try ground loop breaker.The noise may be because of several reasons that begins from input.

Earthing (Grounding) Your Hi-Fi - Tricks and Techniques

Thanks for the suggestion. I completed fitting the ground isolator as per Rod Elliot design. I used a 35A bridge rectifier with a 10 Ohm/5W resistor and a 100nF/1000V capacitor. I could not get the CL60 thermistor today so I've simply shorted that place on the circuit board. I'll fit it later when I can buy it.

The hum level has decreased from what I could hear yesterday but it's not absolute zero. It's really very faint now and I should really stop being bothered about it. I'll let this dog lie and listen to music.

Transformer also bolted down now and LEDs finally installed

Top cover still not screwed down as I want to use the Texas Instruments LM318 after a few days of listening with the JRC LM318. Opamp rolling in progress

 

[hifiramr]

Yes thats great but sometimes we cannot completely mute the hum very slight noise will be there that may come but as you said less bother about it will be fine.

 

[jls001]

Here's a very detailed description of how to design your own LM3886 based amplifier, received from Keith:

A Complete Guide to Design and Build a Hi-Fi LM3886 Amplifier - Circuit Basics

It's amazingly detailed in its explanation of every aspect of amplifier design.

Today's update:
I've installed DC blocking circuitry for mains supply recommended by Rod Elliot. Link here: Mains DC and Transformers

I've used 3510 bridge and two 6400 uF/50V caps instead of the 4700 uF recommended in the above link. I honestly don't know how much this blocker will help but I made it anyway as I already had all the needed parts. This can also be built with your choice of discrete ultra fast switching diodes, but using a rectifier bridge module works out significantly cheaper, and in fact offers better performance especially in current rating (35A bridge is common and costs a measly Rs 25-35, whereas one rectifier diode like MUR860 (8A) will set you back by Rs 45. And you need four to form a full wave bridge. Higher current rated diodes will cost substantially more.

 

[keith_correa]

Here's a very detailed description of how to design your own LM3886 based amplifier, received from Keith:

A Complete Guide to Design and Build a Hi-Fi LM3886 Amplifier - Circuit Basics

It's amazingly detailed in its explanation of every aspect of amplifier design.

Following that along with Tom Christiansen's article should make for a very nice sounding amplifier.

 

[kaushik]

the complete quide is too theoritical..but has good know how for newbies

 

[rikhav]

Had the pleasure of listening to myref amp build by jls001 in my setup.
He was updating me about how it was sounding since the time he started using the amp in his setup and his views were very positive about the amp
But since I had never heard a LM3886 based amp I had no idea how it sounds. All I knew was that it is not a good option for power hungry amps and whose impedance dipped very low
So obviously had a certain picture built in my mind about the amp

But the moment I heard it in my setup I was in for a big surprise. The midrange was so life like and the drive was more then enough for my speakers. There was no loose bass which is generally due to less grunt by the amp.
My speakers are 8 ohms
It was not digging the last bit of details and shade less dynamic then what I have heard but it was not at all a issue atleast for me and the kind of music I listen to
Maybe someone listening to heavy rock would wish for more dynamics and more resolution in the highs

To conclude the sound was such that you forget dissecting sound and just end up enjoying the music. Extremely musical with a very good flow and no way one can say it was extremely rolled off at both ends of the frequency spectrum (typical with budget tube gear)
In all I did not miss my odyssey amp at all while the amp was in my setup

Maybe in the future I can connect my ess9018 dac and see how it sounds with the myref. Not using it right now as the sound tends to be a little on the brighter side with my current amp and speakers

 

[jls001]

Thanks for your impressions, Rikhav :thumbsup:

I've been meaning to pen down my impressions but I plunged headlong into another amp build right after I completed building the MyRef Rev C.

The headlining feature of this amp is the beautiful midrange. Vocals, piano, violin, sax all sound special and have that extra presence much like mid-centric tube amps. This imparts it warmth.

The bass digs deep and has heft, and even on my small standmounts I can discern a shadow of something deeper, perhaps an octave lower than I'm used to hearing. A "shadow" because the speakers won't go much lower than 45-50 Hz.

The highs compared to my big Class A initially sounds rounded off, but on extended listening it's easy to not miss that last bit of higher frequency sparkle and sizzle.

As Rikhav said, this amp has a flow that makes it very musical. One can sit down for extended listening without strain.

On Rikhav's system it sounded very similar to mine but at another friend's place, it had a different character. His speakers are nominally 4 Ohms. My build is wired for 8 Ohm speakers so I knew that it was less than optimal to drive a lower load. Nevertheless, we tried driving the 4 Ohm speakers and what we heard made the five jaded audiophiles who made up the informal listening panel sit up and listen seriously. The chatter in the room stopped abruptly. It came in for much praise with everyone liking it a lot. I felt humbled and a bit embarrassed and actually questioned in my mind whether my friends were giving polite appreciation. But as the evening progressed and got much longer than expected with me playing disc jockey, it became amply clear that my friends were really enjoying themselves. Surprisingly on this system, the highs sounded just right, and not a bit recessed like it sounds on my system or Rikhav's. However the bass was on the higher side as the speakers were deliberately positioned quite close to the front wall for bass reinforcement since my friend's tubed amp does not give sufficient bass weight to his speakers. Mids sounded good there too.

Imaging is very good, with voice dead center. Image depth is decent and is on the more laidback side of things. Separation of voices and instruments is clean even when music becomes complex.

The build was plagued with lots of frustrating failures but in the end it turned out to be very worthy of the time, effort and money spent. I believe the premium components carefully chosen by Linuxguru contributes to the refined sound.

The designer Mauro Penasa is a genius and Linuxguru who had optimised and laid out this board and hand picked the critical components for best sound is another genius. Thank you, gentlemen.

 

[jls001]

Some further experiments in tuning the sound: since the LM318 opamps in the MyRef circuit are derived from the +/-35V rectified DC using a zener to drop the voltage to +/-12V, and no further regulation is carried out on this +/-12V, I tried supplying only the opamps with external regulated DC.

First up, I tried powering it with the OPS power supply set to +/-10V. This is my DCB1 power supply temporarily disconnected and diverted to the MyRef. One OPS was used to supply to both the channels of the MyRef opamps. There was a significant change in the sound signature. It seems to flatten the frequency response from the earlier spotlighted mids and mid bass bump. My main gripe was the recessed highs in the MyRef. The regulated supply brought very good extension in the high frequencies. There was clearly better resolution. Imaging depth and specificity improved substantially. However, the magical midrange was gone. So was the mid bass bump which added a nice weight to the music. The sound signature was more akin to my class A or D amps. I already have that sound signature, and losing the beautiful midrange was too much of a loss. So I reverted to the original supply.

On hindsight, I should have played around with different voltage levels as I had observed dramatic changes in sound signature when different supply voltages are used. I had observed this while building the Blues Power Shunt Regulator designed by doors666.

The other power supply I have at hand is the above-mentioned Blues Power Shunt regulated supply hard wired for +/-13.3V. The sound with this supply was more to my liking, though here too the midrange lost much of its charm.

So this exercise taught me that there seems to be some sort of optimal supply voltage value for a particular circuit. 13.3V is what I like for powering my Kuartlotron buffer and Hypnotoad MC phono preamp, but it's not ideal for the MyRef. More trials could probably lead to the subjective ideal, but for that I need to expand the size of MyRef cabinet to accommodate the extra circuitries. This is not so trivial so I'll rest here. The magical midrange of MyRef is good enough to forgive it's omissions.

 

[linuxguru]

I believe the premium components carefully chosen by Linuxguru contributes to the refined sound.

The designer Mauro Penasa is a genius and Linuxguru who had optimised and laid out this board and hand picked the critical components for best sound is another genius. Thank you, gentlemen.

My pleasure, and thanks for the review. I checked back here after a long gap, hence the delayed reply.

Most of the work in optimization of components is due to others at diyAudio, including Dario Inserra and many others. My contributions in component selection are relatively small and ongoing. I'll however claim credit for the layout optimizations in v1.3 and 1.4 of the PCB, as well as my ongoing effort with the LF01, LF07, etc., to improve the sonics of the LM318 with a Class-A buffer - I hope to resume that activity shortly, now that I have a better listening room and location to test and audition stuff without disturbing the sleep of others.

I just revisited auditioning an early v1.3 build of mine against a vintage Technics SU-4, of which I had recently upgraded some signal-path components. The SU-4 is a compact bookshelf integrated amp, which one might even call a gainclone, since it uses an STK4121 hybrid module. It is a sweet-sounding classic Technics amp, and gets more detailed with my upgrades, but still remains a shade below the MyRefs, both in detail and dynamics. However, it has a tone-control section which has still not been upgraded or bypassed, so it still has room for sonic improvement.

Finally, the MyRefs seem to sound better with full-rangers (1-way speaker setups) like the TABAQ, in some cases with a small helper tweeter (1.5-way). They will work just fine with multi-way setups, too. You may also have to reduce the chipamp rail voltages below +/- 28V for long-term reliability with 4-ohm loads.

Addendum: Dario changed the opamp supply a bit in the Fremen Edition (FE), adding a series-shunt regulator. I haven't auditioned the change, but it apparently changes the sound signature significantly. You may want to check with FMs who have FE builds to get their opinion - I know of at least one, Badrisuper, who built an FE a while ago. I was planning to include that change in the next PCB revision of the MyRef, but it got a bit crowded in a through-hole PCB - I may still revert to it later. I did use a different series-shunt regulator design in the MiniRef v1.02 (a lower-powered stereo PentaWatt nested-gainclone based on the LM1875), and that also seems to work fine for sonics, with a wide variety of opamps.