Hi Myles,
If you have been away a while, the EVO thing can be a bit confusing. There is an EVO amp that George and others have. That is a double current pump design which is a completely different pcb. There is also "EVO Rev A Mod" that is in the BOM and, as Dario said, is simply a different combination of components. EVO Rev A Mod gives the amp a little more stability, if I remember correctly. Whether you build the FE or EVO A in the BOM, it will be a very nice sounding amp.
By the way, I'm jealous of your locations shown in your signature. Have fun with your build.
Jac
A pleasant and very informative experience
Finally on Sunday 23 February I went to visit George, together with my friend Marco. We started from Rome to the east, to the mountains, up to the Gran Sasso d'Italia.
At the foot of the mountain a small village and here George's house.
George is a person of great kindness, with a simple and humble approach (like all great personalities who must not demonstrate anything ...). His passion and competence is well known. His music room is a kind of My_Ref paradise! And not only this...
I will tell you about what I saw and heard and the emotions I felt.
The first goal of the meeting was to compare my My_Ref Fremen Edition Rev.A compensation with George's Full Evo (the one with double current pump, different operational and different compensation). Second goal was to compare its DAC under construction (Sigma-Delta, AKM, Japanese project implemented by George) with mine (NOS, R-2R ladder, Metrum Acoustic Amethyst).
Reference monitor: George's self-made speakers. I will start by talking about these (the following notes are only my first impression, more listening will be needed).
2-way standmount, Scan Speak revelator , 7' in bass reflex. I have not asked for the volume, by eye I would say about 20 liters. I asked for the filter, first order 6dB/oct with speakers impedance compensated by relative circuits and networks. The result is very good. Extended and flat frequency response. A flat response, especially in the medium-high range, I think it can be a problem if everything else does not have the right balance. None of this here.
The quality of the speakers has been enhanced by a design with the right vision and, of course, a great taste in tuning.
I think that thanks to the inclined front baffle that temporally realigns the speakers, it was possible to implement a filter with a gentle slope, but with a perfectly timed acoustic phase. So a flat frequency response is always very musical and pleasant and the monitor produces a realistic soundstage. Monitors, especially professionals, often use L-R 4th order filters to generate flat responses and great distortion control. But at the expense of aspects of reproduction that I can't give up. I almost always prefer L-R 2nd order or first order filters. But as Troels Gravesen explains, then you need the help of the speakers alignment.
What is striking is also the good dynamics: being used to listening with a 10' 94dB midwoofer was not obvious.
In short, an excellent reference monitor whose pleasant correctness was immediately appreciated.
George's Full Evo and DAC were complicit in this performance. Listening to the first pieces of music, I thought that if a 4 Ohm monitor pushed like this it was not only thanks to the Scan Speak and its excellent tuning. It also had to be thanks to the greater strength on low impedances of the Full Evo.
A strong, fast, realistic reproduction at the cost of being raw but not at the expense of pleasure. Good timbre and neutral tone.
By replacing George's DAC with Metrum Acoustics my judgment is flawed. I own it, I'm used to it, I am convinced by its technology and its implementation (this psychologically counts, like the other aspects, don't deny it!). It was natural that I recognized myself more in its sound.
However, my feeling with an R-2R ladder NOS dac like mine compared to an excellent delta-sigma is a slightly more natural sound, without that digital glare that I think I perceive with normal dacs. The result is a slightly preferable timbre.
But for my taste it was equally enjoyable, in a different way, the sound of one of the filters that George set on his music player using his AKM dac. Although I am obviously biased in favor of my dac.
George answered my question that this was one of the filters without pre-ringing and this did not surprise me.
Pause. The menu is important because I think it affects the perception of sound!
In the late morning, to warm our hearts, George welcomed us with an excellent local Pecorino cheese and serious Trebbiano d'Abruzzo wine. During the afternoon listening session we tasted frappe (a traditional Carnival period cake) with Passito Malvasia di Lipari wine. I save you on lunch ...
Together with Malvasia we turned on My_Ref FE with Rev.A compensation. My realization is absolutely budget minded. I simply used the latest industrial BOM with the exception of the Amtrans that were cheap and the Mundorf I had. The suggested minimum of 2 toroidal 160 VA. Then I turned on the C9 to go back to the KZ. Finally I removed C13 since I use a passive transformer preamp.
The sound seemed slightly less raw and fast, less strong (and we expected this), but very little. The control of the My_Ref FE is there and after a minute the sound is fair and pleasant.
As mentioned, my speakers are 94 dB 8 Ohm (min. 6 Ohm) Beyma 10' midwoofer + Morel CAT 378 floorstanding 44 liters and play in a relatively small room. In the room and with George's loudspeakers I appreciated the greater strength of the Evo Full, but I don't know if I would need it.
I understood, also testified by George's experience, that for the sound I am looking for I could remain on My_Ref FE and try Dario's upgrades suggestions as soon as possible, in particular z-foil resistors and M-Lytic and maybe bigger transformers (amtrans and C13 already done).
In short, nothing new for you, experienced connoisseurs of this amplifier, but new knowledge for me: My_Ref FE, especially with Rev.A compensation, made with industrial BOM is already an incredible and definitive amplifier.
Thanks to the work of Dario, confirmed and enriched by other experienced forumers, you can refine and customize the sound. Control, precision, transparency, already excellent, can be improved. The body, the low end and naturalness can increase. And that's what I would like to do.
George's Full EVO is also a magnificent amplifier with a great feeling of energy, control and great realism and with demanding speakers it could be a must.
The high quality of George amp let me understand how good the FE is and even more after upgrades.
I will always thank Dario for his work on My_Ref. I thank George for allowing me to listen to the beautiful Evo Full.
The next step will be a meeting with George at my home in Rome. On this occasion we will have as reference my loudspeakers, less linear and rigorous than George's monitors (the sound recalls the sound signature of vintage loudspeakers like Dynaco A25 - A35, only more efficient and undistorted), but a much easier load for amps.
The main comparison will be between my My_Ref FE industrial Bom without C13 and one of George equal in all, except the replacement of LM318 with a FET chip and consequently without C9. I'm waiting to see if George's My_Ref with the replacement of LM318 and its modifications meets my tastes.
If it is possible (I can't expect George to transport part of his music room on car!) it would be nice to listen to his EVO Full and AKM dac in my room too.
We are also thinking of going, here in Rome, to the beautiful auditorium designed by Renzo Piano to listen to the national orchestra of Santa Cecilia. After all, shouldn't the emotions of live music be the aim of our passion for music machines?
Grazie a tutti
Ciao e buoni ascolti
Giacinto
Finally on Sunday 23 February I went to visit George, together with my friend Marco. We started from Rome to the east, to the mountains, up to the Gran Sasso d'Italia.
At the foot of the mountain a small village and here George's house.
George is a person of great kindness, with a simple and humble approach (like all great personalities who must not demonstrate anything ...). His passion and competence is well known. His music room is a kind of My_Ref paradise! And not only this...
I will tell you about what I saw and heard and the emotions I felt.
The first goal of the meeting was to compare my My_Ref Fremen Edition Rev.A compensation with George's Full Evo (the one with double current pump, different operational and different compensation). Second goal was to compare its DAC under construction (Sigma-Delta, AKM, Japanese project implemented by George) with mine (NOS, R-2R ladder, Metrum Acoustic Amethyst).
Reference monitor: George's self-made speakers. I will start by talking about these (the following notes are only my first impression, more listening will be needed).
2-way standmount, Scan Speak revelator , 7' in bass reflex. I have not asked for the volume, by eye I would say about 20 liters. I asked for the filter, first order 6dB/oct with speakers impedance compensated by relative circuits and networks. The result is very good. Extended and flat frequency response. A flat response, especially in the medium-high range, I think it can be a problem if everything else does not have the right balance. None of this here.
The quality of the speakers has been enhanced by a design with the right vision and, of course, a great taste in tuning.
I think that thanks to the inclined front baffle that temporally realigns the speakers, it was possible to implement a filter with a gentle slope, but with a perfectly timed acoustic phase. So a flat frequency response is always very musical and pleasant and the monitor produces a realistic soundstage. Monitors, especially professionals, often use L-R 4th order filters to generate flat responses and great distortion control. But at the expense of aspects of reproduction that I can't give up. I almost always prefer L-R 2nd order or first order filters. But as Troels Gravesen explains, then you need the help of the speakers alignment.
What is striking is also the good dynamics: being used to listening with a 10' 94dB midwoofer was not obvious.
In short, an excellent reference monitor whose pleasant correctness was immediately appreciated.
George's Full Evo and DAC were complicit in this performance. Listening to the first pieces of music, I thought that if a 4 Ohm monitor pushed like this it was not only thanks to the Scan Speak and its excellent tuning. It also had to be thanks to the greater strength on low impedances of the Full Evo.
A strong, fast, realistic reproduction at the cost of being raw but not at the expense of pleasure. Good timbre and neutral tone.
By replacing George's DAC with Metrum Acoustics my judgment is flawed. I own it, I'm used to it, I am convinced by its technology and its implementation (this psychologically counts, like the other aspects, don't deny it!). It was natural that I recognized myself more in its sound.
However, my feeling with an R-2R ladder NOS dac like mine compared to an excellent delta-sigma is a slightly more natural sound, without that digital glare that I think I perceive with normal dacs. The result is a slightly preferable timbre.
But for my taste it was equally enjoyable, in a different way, the sound of one of the filters that George set on his music player using his AKM dac. Although I am obviously biased in favor of my dac.
George answered my question that this was one of the filters without pre-ringing and this did not surprise me.
Pause. The menu is important because I think it affects the perception of sound!
In the late morning, to warm our hearts, George welcomed us with an excellent local Pecorino cheese and serious Trebbiano d'Abruzzo wine. During the afternoon listening session we tasted frappe (a traditional Carnival period cake) with Passito Malvasia di Lipari wine. I save you on lunch ...
Together with Malvasia we turned on My_Ref FE with Rev.A compensation. My realization is absolutely budget minded. I simply used the latest industrial BOM with the exception of the Amtrans that were cheap and the Mundorf I had. The suggested minimum of 2 toroidal 160 VA. Then I turned on the C9 to go back to the KZ. Finally I removed C13 since I use a passive transformer preamp.
The sound seemed slightly less raw and fast, less strong (and we expected this), but very little. The control of the My_Ref FE is there and after a minute the sound is fair and pleasant.
As mentioned, my speakers are 94 dB 8 Ohm (min. 6 Ohm) Beyma 10' midwoofer + Morel CAT 378 floorstanding 44 liters and play in a relatively small room. In the room and with George's loudspeakers I appreciated the greater strength of the Evo Full, but I don't know if I would need it.
I understood, also testified by George's experience, that for the sound I am looking for I could remain on My_Ref FE and try Dario's upgrades suggestions as soon as possible, in particular z-foil resistors and M-Lytic and maybe bigger transformers (amtrans and C13 already done).
In short, nothing new for you, experienced connoisseurs of this amplifier, but new knowledge for me: My_Ref FE, especially with Rev.A compensation, made with industrial BOM is already an incredible and definitive amplifier.
Thanks to the work of Dario, confirmed and enriched by other experienced forumers, you can refine and customize the sound. Control, precision, transparency, already excellent, can be improved. The body, the low end and naturalness can increase. And that's what I would like to do.
George's Full EVO is also a magnificent amplifier with a great feeling of energy, control and great realism and with demanding speakers it could be a must.
The high quality of George amp let me understand how good the FE is and even more after upgrades.
I will always thank Dario for his work on My_Ref. I thank George for allowing me to listen to the beautiful Evo Full.
The next step will be a meeting with George at my home in Rome. On this occasion we will have as reference my loudspeakers, less linear and rigorous than George's monitors (the sound recalls the sound signature of vintage loudspeakers like Dynaco A25 - A35, only more efficient and undistorted), but a much easier load for amps.
The main comparison will be between my My_Ref FE industrial Bom without C13 and one of George equal in all, except the replacement of LM318 with a FET chip and consequently without C9. I'm waiting to see if George's My_Ref with the replacement of LM318 and its modifications meets my tastes.
If it is possible (I can't expect George to transport part of his music room on car!) it would be nice to listen to his EVO Full and AKM dac in my room too.
We are also thinking of going, here in Rome, to the beautiful auditorium designed by Renzo Piano to listen to the national orchestra of Santa Cecilia. After all, shouldn't the emotions of live music be the aim of our passion for music machines?
Grazie a tutti
Ciao e buoni ascolti
Giacinto
George is a craftsman. I am sure his full setup is very special. It is all about the weakest link. Usually that is the loudspeaker.
I used Thiel CS-3.5 with an Evo and a FE. They were Thiel last product with European drivers. Vifa woofer, Scanspeak mid, and Dynaudio tweeter. First order slopes and time aligned. It was hard to move on.
I think the amps are more linear and less distorted than any loudspeaker.
I used Thiel CS-3.5 with an Evo and a FE. They were Thiel last product with European drivers. Vifa woofer, Scanspeak mid, and Dynaudio tweeter. First order slopes and time aligned. It was hard to move on.
I think the amps are more linear and less distorted than any loudspeaker.
Giacinto is exagerating..
And I would like to thank You two, for the extremely pleasant and refreshing encounter.
I had realized that I had to do with long time experienced people with high culture and very fine taste.
That the FE had become their preferred choice (after many adventures) is talking for itself.
And the FE of Giacinto had provided a very nice, equilibrated sound.
The Metrum dac does sound warm and musical. From my side I perceive a ~NOS sound (with it's pro's and contra's) and tried to break the 'magic' with some HQP upsampling..) For me it is getting better that way, but I think we will return to this several times..
My test bed is a Yanasan DM4499, at this moment.
We listened with laptop server, Allo renderer, Gentooplayer system, (everywhere), & HQP in DSd 256.
The dac is fed with an external clock (NDK SDA), and is in dsd-direct modality.
And I'm still playing with the configurations..
Ciao! G
And I would like to thank You two, for the extremely pleasant and refreshing encounter.
I had realized that I had to do with long time experienced people with high culture and very fine taste.
That the FE had become their preferred choice (after many adventures) is talking for itself.
And the FE of Giacinto had provided a very nice, equilibrated sound.
The Metrum dac does sound warm and musical. From my side I perceive a ~NOS sound (with it's pro's and contra's) and tried to break the 'magic' with some HQP upsampling..
) For me it is getting better that way, but I think we will return to this several times.. My test bed is a Yanasan DM4499, at this moment.
We listened with laptop server, Allo renderer, Gentooplayer system, (everywhere), & HQP in DSd 256.
The dac is fed with an external clock (NDK SDA), and is in dsd-direct modality.
And I'm still playing with the configurations..
Ciao! G
Dario evidently did a long and patient job demonstrating unmatched taste and balance. I only made his industrial BOM clean and limited myself to the minimum requirements for every other component.
Only now do I understand that many upgrade suggestion in Dario's tutorial are simply the right one. The effects described for each upgrade are perfectly replicable (at least in the majority of setups) as in a scientific experiment!
I really have to retrieve two FE boards to build a version giving me some little luxury. And I still have to listen to your FE with LM318 replaced.
Ciao Giacinto
Jac, Dario
Thank you for the BOM information for the amplifier. The avatar looks across the valley into the Kokanee glacier. We have just built a new house here. It is the most quiet and serene place I have experienced in my life. The complete opposite of PV, which is now our winter holiday place when needed.
Just organizing my workplace, then some practice on some gainclones, then onto the FE build.
Regards,
MM
Thank you for the BOM information for the amplifier. The avatar looks across the valley into the Kokanee glacier. We have just built a new house here. It is the most quiet and serene place I have experienced in my life. The complete opposite of PV, which is now our winter holiday place when needed.
Just organizing my workplace, then some practice on some gainclones, then onto the FE build.
Regards,
MM
just checking
Hi all,
Back in 2016 I bought 2 myRef boards, and the time has finally come to actually make the amplifier. I have everything installed on the boards and I will do the first tests this weekend!
The only thing before i proceed is that I want to make sure that all the wiring is correct. I understand that in principle the board is designed to be floating however it does feel safer to me to have some sort of grounding and the ground lift should prevent any hum issues to my mind. I have attached a picture of the wiring scheme I intend to use (only a single channel for clarity), the green lines are signal wires and audio output and the red wires are the power side of things. The heatsink is not electrically connected ot the rest of the chassis so it connects into the star ground, and the socket contains a fuse and switch. Does anybody see anything which needs correcting?
Thanks Mauro and all others for making this design btw! up to now it has been a very nice project to work on. And once everything is finished up I will of course post some pictures!
kind regards,
Joost
Hi all,
Back in 2016 I bought 2 myRef boards, and the time has finally come to actually make the amplifier. I have everything installed on the boards and I will do the first tests this weekend!
The only thing before i proceed is that I want to make sure that all the wiring is correct. I understand that in principle the board is designed to be floating however it does feel safer to me to have some sort of grounding and the ground lift should prevent any hum issues to my mind. I have attached a picture of the wiring scheme I intend to use (only a single channel for clarity), the green lines are signal wires and audio output and the red wires are the power side of things. The heatsink is not electrically connected ot the rest of the chassis so it connects into the star ground, and the socket contains a fuse and switch. Does anybody see anything which needs correcting?
Thanks Mauro and all others for making this design btw! up to now it has been a very nice project to work on. And once everything is finished up I will of course post some pictures!
kind regards,
Joost
Attachments
Hey everyone,
I tried firing up one channel tonight however something very strange is happening.
With no inputs or outputs connected i get the following behaviour
- on power up everything seems to work fine, the led turns on and i hear the relay click
- once i touch basically any component with the multimeter (for example the output), the relay immediately clicks and the board turns off
- when I remove the power after a short time the relay clicks again, the led turns on for a short time and then the relay clicks again and the led goes off again...
I'm quite puzzled by this behaviour i must say. I can post some pictures tomorrow, and as the multimeter sets of the relay I can't measure voltages i'm afraid.
Thanks in advance for any advice!
Joost
I tried firing up one channel tonight however something very strange is happening.
With no inputs or outputs connected i get the following behaviour
- on power up everything seems to work fine, the led turns on and i hear the relay click
- once i touch basically any component with the multimeter (for example the output), the relay immediately clicks and the board turns off
- when I remove the power after a short time the relay clicks again, the led turns on for a short time and then the relay clicks again and the led goes off again...
I'm quite puzzled by this behaviour i must say. I can post some pictures tomorrow, and as the multimeter sets of the relay I can't measure voltages i'm afraid.
Thanks in advance for any advice!
Joost
Try to disconnect the groundbraker.
Connect PGND with a direct, short thick wire to the chassis. Be sure that the heatsinks are well connected to the chassis, too.
Or.
Do not connect ground breaker, and your red wires on diagram.
If the heatsinks are ~floating originally, (is it the reason for the red extra wire to them?) >> Then eliminate red extra wire and connect sinks to HSground on the boards. Very short direct piece of metal.
All this for now, one time temporary. See if it works that way, then we can proceed.
Again: I suspect oscillations, because of absolutely non-existent chassis connection for schielding.
Establish that, low inductance short wires.
Then in a second run we will think of the PE connections.
Ciao, george
Connect PGND with a direct, short thick wire to the chassis. Be sure that the heatsinks are well connected to the chassis, too.
Or.
Do not connect ground breaker, and your red wires on diagram.
If the heatsinks are ~floating originally, (is it the reason for the red extra wire to them?) >> Then eliminate red extra wire and connect sinks to HSground on the boards. Very short direct piece of metal.
All this for now, one time temporary. See if it works that way, then we can proceed.
Again: I suspect oscillations, because of absolutely non-existent chassis connection for schielding.
Establish that, low inductance short wires.
Then in a second run we will think of the PE connections.
Ciao, george
I grounded the heatsink (see attached picture). And I still have the same problem, there was one time when it did not turn off immediately and then I measured a offset of 5.4 mV over the output tabs. I moved the setup over to another room where it is not right next to all kinds of noisy electronics (computer etc) but that did not help.
I also included some pictures of the board which I hope are clear enough! As you can probably see my soldering skills are not very good but I did check everything for continuity and everything does seem to be properly connected!
EDIT: the power wires are now twisted and this also does not help
I also included some pictures of the board which I hope are clear enough! As you can probably see my soldering skills are not very good but I did check everything for continuity and everything does seem to be properly connected!
EDIT: the power wires are now twisted and this also does not help
Attachments
Last edited:
Joost,
Ok.. The photos show a correct assembly and grounding now.
What happens if You short the input pins (a jumper on the connector) and do the same?
And again, what are You touching exactly? If You touch sensitive inner nodes, your description shows the correct "response"... Also valid for inputs..
What is not normal that it stays in that condition, until You turn off.
The described turn-off behaviour is again clear sign of an oscillating system.
Reassuming: after turn on, it's activating OK. If you do not touch patts, it remains ok, I guess?
I would say, it is not 'politically correct' to touch other than the output pin, in these conditions..
Correct work configuration is : a load connected to the output, (anything from 10ohm downwards), and some lowish impedance connected to input (like 1kohm).
Maybe try what happens that way.
Anyway, though it is ufficially not a correct modality, generally these units remain stable also 'everything open'.
Question: both modules are behaving the same?
Tentative suggestion: given that it shows signs of instability, are You sure about all the parts quality/ value in the compensation scheme? (capacitor values etc)
George
Ok.. The photos show a correct assembly and grounding now.
What happens if You short the input pins (a jumper on the connector) and do the same?
And again, what are You touching exactly? If You touch sensitive inner nodes, your description shows the correct "response"... Also valid for inputs..
What is not normal that it stays in that condition, until You turn off.
The described turn-off behaviour is again clear sign of an oscillating system.
Reassuming: after turn on, it's activating OK. If you do not touch patts, it remains ok, I guess?
I would say, it is not 'politically correct' to touch other than the output pin, in these conditions..
Correct work configuration is : a load connected to the output, (anything from 10ohm downwards), and some lowish impedance connected to input (like 1kohm).
Maybe try what happens that way.
Anyway, though it is ufficially not a correct modality, generally these units remain stable also 'everything open'.
Question: both modules are behaving the same?
Tentative suggestion: given that it shows signs of instability, are You sure about all the parts quality/ value in the compensation scheme? (capacitor values etc)
George
Second update,
The other board seems to be working well, relay clicks on powerup, offset is varying between 0.4 and 0.8 mV and if I connect some old speakers a slight hum is audible without any input connected!
The board was connected up exactly the same way as the first board.
EDIT: @Joseph K, I only saw your responses after posting
The relay also clicks when I only touch the output tabs, and you are right that the system stays on if I don't touch anything.
I will resolder the resistor. And I will recheck all values! As for quality is used the industrial bom (from mouser) with only C9 upgraded (Cerafine) .
The other board seems to be working well, relay clicks on powerup, offset is varying between 0.4 and 0.8 mV and if I connect some old speakers a slight hum is audible without any input connected!
The board was connected up exactly the same way as the first board.
EDIT: @Joseph K, I only saw your responses after posting
The relay also clicks when I only touch the output tabs, and you are right that the system stays on if I don't touch anything.
I will resolder the resistor. And I will recheck all values! As for quality is used the industrial bom (from mouser) with only C9 upgraded (Cerafine) .
Last edited: