Everyone is entitled to his own belief and opinion.
But what are our own conclusions after all the tests ?
In DIY, there are people who follow their idols 100%.
Not only circuits, but every single details, from active and passive components, to PCB shapes, enclosures, heatsinks, cables, connectors, …..
And more often than not without asking for the reasons behind those choices.
Maybe this is the guarantee to good sound without stress ? 😊
If that makes them happy, why not ?
It is only hobby, and hobby is only about pleasing oneself.
If anyone wants to build 99% close to what Hiraga has published, we have already published the Gerber files of his PCB design.
You can get most of the original transistors on the internet.
We got all ours from a DIY shop in Hong Kong, apart from the 2SK170/2SJ74GR which we have our own stock.
You can also get leaded tantalum resistors (including Shinkoh’s) relatively easily, at a premium.
And you can follow the published schematics of his power supply without much problem.
We appreciate good analogue designs, and these cannot always be judged by measurable parameters.
When we rebuild a classic design, like the Le Monstre, we also respect the original, and only want to change if it is necessary.
Otherwise we might as well design our own, and we have done that before.
If we were to build the Le Monstre again, we shall certainly only build the LM Modern, using active components as detailed in our article in post #1.
We shall use the CCS bypass as described for JFETs with Idss >4mA.
And we shall use good passive components – meaning those known to be low noise, low tempco, good reliability but of current production.
Incidentally, the RA12P is wire-wound and is a good trimmer which we shall use again, also elsewhere.
But of course, we shall use our own PCB, as we believe modern SMD components give us more freedom for a better layout.
As to power supply, it is true that the Le Monstre circuit is not outstanding for PSRR.
And we are not dealing with very large signal amplitudes.
So it makes sense to have a low noise power supply.
But the voltage does not have to exact and constant over time.
This is also reflected in the original design, with the huge bank of capacitors.
They form a very low frequency low pass filter with the CRC resistor, as well as low output impedance at high frequencies.
The Le Monstre is a fast amplifier, with -3dB bandwidth up to MHz.
IMHO, the power supply needs to be able to match that, i.e. with low Zout at high frequencies.
Feedback type voltage regulators, including so-called Super Regulators, are inductive at HF, and only rely on the output capacitor to give low(er) Zout.
So if we were to do this again, we shall most likely use a Cap Multiplier instead, maybe optionally with a simple 317/337 prereg.
It is more important to use separate supplies for each channel.
If we have to be restricted to regulators, we shall probably use 4x LT3045/3094 in parallel, with film caps at the output and not MLCC.
But we have not tried. So purely speculation, if you like.
Our own plan now is to put the LM Modern in a suitable case, with toroidal transformers and most probably cap multipliers.
And then find good homes for the LM Original and the LM Deluxe for someone who appreciate them.
If there is still sufficient interest in our PCBs and in the input JFETs (GR grade, matched, no Idss cherry picking), we shall make them available in Group Buys.
But as mentioned many times before, you don’t need anything from us to build your own.
This also concludes our contribution to the Le Monstre 2024.
Happy building and listening,
Patrick
But what are our own conclusions after all the tests ?
In DIY, there are people who follow their idols 100%.
Not only circuits, but every single details, from active and passive components, to PCB shapes, enclosures, heatsinks, cables, connectors, …..
And more often than not without asking for the reasons behind those choices.
Maybe this is the guarantee to good sound without stress ? 😊
If that makes them happy, why not ?
It is only hobby, and hobby is only about pleasing oneself.
If anyone wants to build 99% close to what Hiraga has published, we have already published the Gerber files of his PCB design.
You can get most of the original transistors on the internet.
We got all ours from a DIY shop in Hong Kong, apart from the 2SK170/2SJ74GR which we have our own stock.
You can also get leaded tantalum resistors (including Shinkoh’s) relatively easily, at a premium.
And you can follow the published schematics of his power supply without much problem.
We appreciate good analogue designs, and these cannot always be judged by measurable parameters.
When we rebuild a classic design, like the Le Monstre, we also respect the original, and only want to change if it is necessary.
Otherwise we might as well design our own, and we have done that before.
If we were to build the Le Monstre again, we shall certainly only build the LM Modern, using active components as detailed in our article in post #1.
We shall use the CCS bypass as described for JFETs with Idss >4mA.
And we shall use good passive components – meaning those known to be low noise, low tempco, good reliability but of current production.
Incidentally, the RA12P is wire-wound and is a good trimmer which we shall use again, also elsewhere.
But of course, we shall use our own PCB, as we believe modern SMD components give us more freedom for a better layout.
As to power supply, it is true that the Le Monstre circuit is not outstanding for PSRR.
And we are not dealing with very large signal amplitudes.
So it makes sense to have a low noise power supply.
But the voltage does not have to exact and constant over time.
This is also reflected in the original design, with the huge bank of capacitors.
They form a very low frequency low pass filter with the CRC resistor, as well as low output impedance at high frequencies.
The Le Monstre is a fast amplifier, with -3dB bandwidth up to MHz.
IMHO, the power supply needs to be able to match that, i.e. with low Zout at high frequencies.
Feedback type voltage regulators, including so-called Super Regulators, are inductive at HF, and only rely on the output capacitor to give low(er) Zout.
So if we were to do this again, we shall most likely use a Cap Multiplier instead, maybe optionally with a simple 317/337 prereg.
It is more important to use separate supplies for each channel.
If we have to be restricted to regulators, we shall probably use 4x LT3045/3094 in parallel, with film caps at the output and not MLCC.
But we have not tried. So purely speculation, if you like.
Our own plan now is to put the LM Modern in a suitable case, with toroidal transformers and most probably cap multipliers.
And then find good homes for the LM Original and the LM Deluxe for someone who appreciate them.
If there is still sufficient interest in our PCBs and in the input JFETs (GR grade, matched, no Idss cherry picking), we shall make them available in Group Buys.
But as mentioned many times before, you don’t need anything from us to build your own.
This also concludes our contribution to the Le Monstre 2024.
Happy building and listening,
Patrick
To hear is not just about hearing. It is a mood, the entire organism is put in mood. And everyone knows systems that are in tune, that click in, and those that are not. This memory is longer than a few seconds. This is not, or very rarely, experienced when switching while listening.
Another experience that everyone here has had: if you have listened to a system for a long time, hours, days, or months or even years, and then change a device that takes longer than 10 seconds, you will still recognize the change. To be more precise: here too, the frequently cited, low-complexity study that claims to have proven that the ear only "remembers" a few seconds has been refuted.
Aside: tinnitus does not have much to do with ears and hearing. Often it is the resonance of teeth, or foreign dental materials (dental fillings, titanium pins, dead teeth). Tinnitus can also occur throughout the body, especially after severe poisoning or exposure to high frequency EMF, such as cell phone. It is the "swinging" of the nervous system, of the human electric network;-)
Do it as always: just listen to the systems, then change relaxed, even if it takes minutes, and continue listening. Most of the differences will only become noticeable now. And if you want to be "objective", take a few listeners and have them take notes, which can then be compared. And if (almost) everyone has heard the same thing, then the perceived difference is not due to the subject, not the listener.
And: "It is more important to use separate supplies for each channel." No.
Here it becomes important to understand hearing as a measuring instrument, and hearing as a measuring process. And experience shows - if you don't just switch back and forth while listening - that the ear perceives differences between the equal parts, the same batch: The ear recognizes a channel-separated supply by higher unrest, less cleanliness and less homogeneity. the channel-separated supply sounds dirtier, more distorted, almost inaudible in comparison. But here we can think further: if the channel-separated supply is audible, isn't the half-wave-separated supply also audible? And what about the use of half-wave separated amplifiers and the use of unequal parts (complementary transistors)? And since we have proof that the ear perceives modulation by parts: that parts modulate, do many parts modulate less or more?
And so on;-)
Another experience that everyone here has had: if you have listened to a system for a long time, hours, days, or months or even years, and then change a device that takes longer than 10 seconds, you will still recognize the change. To be more precise: here too, the frequently cited, low-complexity study that claims to have proven that the ear only "remembers" a few seconds has been refuted.
Aside: tinnitus does not have much to do with ears and hearing. Often it is the resonance of teeth, or foreign dental materials (dental fillings, titanium pins, dead teeth). Tinnitus can also occur throughout the body, especially after severe poisoning or exposure to high frequency EMF, such as cell phone. It is the "swinging" of the nervous system, of the human electric network;-)
Do it as always: just listen to the systems, then change relaxed, even if it takes minutes, and continue listening. Most of the differences will only become noticeable now. And if you want to be "objective", take a few listeners and have them take notes, which can then be compared. And if (almost) everyone has heard the same thing, then the perceived difference is not due to the subject, not the listener.
And: "It is more important to use separate supplies for each channel." No.
Here it becomes important to understand hearing as a measuring instrument, and hearing as a measuring process. And experience shows - if you don't just switch back and forth while listening - that the ear perceives differences between the equal parts, the same batch: The ear recognizes a channel-separated supply by higher unrest, less cleanliness and less homogeneity. the channel-separated supply sounds dirtier, more distorted, almost inaudible in comparison. But here we can think further: if the channel-separated supply is audible, isn't the half-wave-separated supply also audible? And what about the use of half-wave separated amplifiers and the use of unequal parts (complementary transistors)? And since we have proof that the ear perceives modulation by parts: that parts modulate, do many parts modulate less or more?
And so on;-)
In comparison between Cap Mx and Super Regulator, my opinion is that Super Regulator is a far better choice. Output impedance at high frequencies is not a problem. With several mF capacitor at the output, at 20 kHz I’ve measured output impedance 40 uΩ. 10 cm of wire is order of magnitude bigger issue. Cap Mx will have considerable rails modulation under AC load. This is the opposite of original LM supply properties.
Addendum :
We have been working on a suitable preamp for the LM that we are going to keep for ourselves.
We shall post that separately when ready.
Patrick
We have been working on a suitable preamp for the LM that we are going to keep for ourselves.
We shall post that separately when ready.
Patrick
Last edited:
For those who are interested in using the boosted LT3042 for this project, here is a good reference :
https://www.diyaudio.com/community/...elessness-with-the-lt3045.396792/post-7310304
For 4x LT3045 in parallel (even lower noise by 2x), refer to P.27 of 3045 datasheet.
There are ready made modules on the net with 6x in parallel, for 12V.
But I have not tried them.
And I would recommend using 15V 1.5A (min) preregs upstream to take out most of the mains ripples.
Patrick
https://www.diyaudio.com/community/...elessness-with-the-lt3045.396792/post-7310304
For 4x LT3045 in parallel (even lower noise by 2x), refer to P.27 of 3045 datasheet.
There are ready made modules on the net with 6x in parallel, for 12V.
But I have not tried them.
And I would recommend using 15V 1.5A (min) preregs upstream to take out most of the mains ripples.
Patrick
I have a few enquiries about kits.
As mentioned before, we are happy to offer PCBs and matched JFETs as one package.
Postage will otherwise dominate.
And no point in us buying stuff from Mouser and then send them round the world.
Extra 20% EU tax on top.
As to power supply, our version of the Didden Super Power is too difficult for average DIY.
So we are working on a 3x parallel LT3045 instead.
But we want to test first before offering together with the above.
So this will take some more time.
Unless of course you do your own PSUs or buy elsewhere.
Saves us a lot of work. 🙂
Patrick
As mentioned before, we are happy to offer PCBs and matched JFETs as one package.
Postage will otherwise dominate.
And no point in us buying stuff from Mouser and then send them round the world.
Extra 20% EU tax on top.
As to power supply, our version of the Didden Super Power is too difficult for average DIY.
So we are working on a 3x parallel LT3045 instead.
But we want to test first before offering together with the above.
So this will take some more time.
Unless of course you do your own PSUs or buy elsewhere.
Saves us a lot of work. 🙂
Patrick
What would be the difference compared to DIY audio store Super regulator pcb? Using GLED431 or extra transistors?
For me the most important is the sound, the feeling I get from it.
Can the super regulator from DIY Store deliver 1.5A current ?
Or you mean why it is difficult ?
Patrick
Or you mean why it is difficult ?
Patrick
Last edited:
Yes DIY store regulator can deliver up to 1 A current.
What changes you made in this regulator that it can deliver 1,5 A or more current? Yes I mean why it is difficult and you think that diy users can not make it so easily?
Do you share the schematic?
We use a discrete Darlington with the same power devices as the Le Monstre Modern.
It can easily deliver up to 10A with sufficient heat sinking.
The previous GB experience with the Dao headphone amp told me I should not do this again.
You can get something similar from Belleson.
About 70€ per regulator, without power device.
So regulators alone ~300€ ?
If you believe the Store version is sufficient, feel free to go ahead.
🙂
Cheers,
Patrick
It can easily deliver up to 10A with sufficient heat sinking.
The previous GB experience with the Dao headphone amp told me I should not do this again.
You can get something similar from Belleson.
About 70€ per regulator, without power device.
So regulators alone ~300€ ?
If you believe the Store version is sufficient, feel free to go ahead.
🙂
Cheers,
Patrick
On the same subject, you can try to use TIP122/127 for the pass devices.
That will give you more current, but not more power dissipation.
And of course you will need larger caps.
I have not tried.
You should ask Jan at the support thread.
https://www.diyaudio.com/community/threads/super-regulator.247281/
Patrick
That will give you more current, but not more power dissipation.
And of course you will need larger caps.
I have not tried.
You should ask Jan at the support thread.
https://www.diyaudio.com/community/threads/super-regulator.247281/
Patrick
Last edited:
Our two deluxe modules, as in Version 1 in post #146, have already found a good new home.
https://www.diyaudio.com/community/threads/hiraga-le-monstre-2024.413301/post-7757582
🙂
Patrick
https://www.diyaudio.com/community/threads/hiraga-le-monstre-2024.413301/post-7757582
🙂
Patrick
I'm coming back from vacation next week, I'm going to the Alps tomorrow to rest from the sun and the sea (and crowds) and then I'll post the scheme for the positive and negative series reg 😉 .No worries, you will also have green LEDs.
Great that someone else volunteers to take care of things.
But of course that particular design has already been published :
https://www.diyaudio.com/community/threads/hiraga-le-monstre.5462/post-6192683
We have another simpler solution planned.
Hopefully test results in a month or so.
Patrick
But of course that particular design has already been published :
https://www.diyaudio.com/community/threads/hiraga-le-monstre.5462/post-6192683
We have another simpler solution planned.
Hopefully test results in a month or so.
Patrick
Well actually he already did.
https://www.diyaudio.com/community/threads/super-regulator.247281/post-5986623
And so did Walt :
https://www.diyaudio.com/community/threads/super-regulator.247281/post-7687738
Patrick
It is possible without the darlington transistor, it is only necessary to ensure sufficient current through the CCS to drive the pass transistor. Normally, it is necessary to dimension the CCS components for that current and that is all that needs to be done. In the new version I drive KSC5242/KSA1962 with ZTX450/550 and there is no current limit, everything depends on the pass transistor. This is also a disadvantage because the classic version has a current limitation by itself.
In our version, the CCS is less than 3mA, and not 40mA+ as in post #177.
It works with both Darlington and with Sziklai.
Using TIP122/127, you can actually use the DIY Store boards with sufficiently larger heatsinks.
See post #178.
Of course none of these were used by Jean Hiraga.
His was just CRC with a huge amount of capacitance.
Anything else is individual interpretation.
As was our Didden regulator.
But of course everyone is free to choose what he likes best.
Patrick
It works with both Darlington and with Sziklai.
Using TIP122/127, you can actually use the DIY Store boards with sufficiently larger heatsinks.
See post #178.
Of course none of these were used by Jean Hiraga.
His was just CRC with a huge amount of capacitance.
Anything else is individual interpretation.
As was our Didden regulator.
But of course everyone is free to choose what he likes best.
Patrick