| Stefano |
| Hi all you folks, i ask for your competence for the Hiraga amp "the monster" that seems to me to be a interesting and simple amp. As anyone of you ever builded it? I'm interested to study a mosfet version to use it as car amp to economize to PSU. Any comment is welcome. |
|
|
| jazz |
stefano,
I've built this amp and i think it is very good. I don't know about a car amp though, it draws quite a bit of current Anyway if you do want to build it here's some tips: If you built this please be sure to bypass the biasing circuit of the input cascode with some good caps (or give it a really good regulator), this makes it sound quite a lot better. This is a very fast amp and to my experinece it is best to give it a psu that is build up with a large number of smaller caps in stead of one large one, a pi filter works very well and series resistance in front of the rectifier (0.33/47 OHMs) will do good as well. In the case of the mosfet version please have a look at goeff's excellent website. He has got the original articles printed there and hiraga did look at mosfets as well, there's a pcb layout to.
greetings, Joris |
|
|
| planet10 |
| quote: | Originally posted by jazz
goeff's excellent website. He has got the original articles printed there and hiraga did look at mosfets as well, there's a pcb layout to. |
URL?
dave |
|
|
| planet10 |
Thanx. A lot of interesting stuff there.
dave |
|
|
| François |
| Very cool article on the Musical Fidelity A1 |
|
|
| carlmart |
François,
Which article on the A1 you mean? There's only the A1's circuit at Geoff's site.
Or you mean the link at Mark Hennessy’s site? His is really an excellent comment on the A1.
Carlos |
|
|
| hifidaddy |
Hi,
here in Munich a lot of guys had built these around 1990, as well as the pure bipolar version "Le Classe A".
Personally, I prefer the "Le Classe A", for it does the PRAT (pace, ryhtm, and timing) thing much better, in a way otherwise Naim does. "Le Monstre" OTOH is smoother in the mid to treble region, sounding more "High End Stereo".
Where the Monster excells, is when driving 100dB horn speakers. It can be viewed as a viable alternative to a single ended triode design.
regards,
Hartmut from Munich |
|
|
| François |
| quote: | Originally posted by carlmart
François,
Which article on the A1 you mean? There's only the A1's circuit at Geoff's site.
Or you mean the link at Mark Hennessy’s site? His is really an excellent comment on the A1.
Carlos |
Mark's article.
Sorry for the confusion. |
|
|
| NU_NRG |
| Someone here know where we could find pictures of this amplifier? |
|
|
| hifidaddy |
| quote: | Originally posted by NU_NRG
Someone here know where we could find pictures of this amplifier? |
Hi,
the "Le Monstre" was a pure DIY project in that "L'Audiophile" magazin, whereas its big brother, the "Hiraga Le Classe A" was also a commercial realisation.
I try to load a pic of a commercial "Classe A", which was the first realisation of that amp, with 20W power each channel. The later revised "Le Classe A" had up to 30W per channel, and bigger lytics in the filter stage 330 000 uF (4 times).
BTW, later years of production are much prettier.
regards,
Hartmut |
|
|
| hifidaddy |
regards,
Hartmut |
|
|
| dice45 |
Hartmut and all,
agreed, the Hiraga amp had better PRaT, and "Le Monstre 8W" was more smooth and open in the mid to treble region. I had this amp running for years, making me tube nuts completely happy with it. Control in the low end was remarkable, provided listening volume was moderate.
I once heard a battery operated version, fed form 170Ah battery capacity and several Farads of PS cap capacity.
This was better than my unit, considerably, this was really something.
One warning: the input FET determining the quiescent current (by its transconductance is ultra-sensitve to electrostatic hazard; i killed my monster, performing an innocent mod (other resistor brand) and then finding the sucker sucks 4 times as much current, frying itself. :( |
|
|
| NU_NRG |
thank you hifidaddy
I'm surprised to see how simple this amp looks. It was said somewhere that it wasn't marketable because of it's huge size. If someone would decide to build one of these, how would it compare to the Zen or Aleph from Pass?
-Simon |
|
|
| jazz |
Hi there,
Hartmut and Bernhard, did you build a complete diy version of the le classe and if so which version? I built it three times as well but never managed to get rid of some low level (but audible hum) in that design and i tried everything but nothing worked so i finally settled on the le monstre. I agree with the difference in sound, to me the le classe sounds more friendly and warm. The le monstre has an incredible detail and bass though.
Stefano, one other tip due to the large differences in behaviour of the fets the load resistor of the input cascode will probably have to be tinkered with in order to get the right standing current. more resistance is more current and the other way round. Sorry by the way for not posting the url in my first reply.
Greetings Joris |
|
|
| Dozuki |
I have not read what articles of this amp are in English (yet), but I was curious if you guys are Beta matching your transistors. If so, are you matching the NPN to the PNP compliments?
-D. |
|
|
| hifidaddy |
| quote: | Originally posted by jazz
Hi there,
Hartmut and Bernhard, did you build a complete diy version of the le classe and if so which version? I built it three times as well but never managed to get rid of some low level (but audible hum) in that design and i tried everything but nothing worked so i finally settled on the le monstre. I agree with the difference in sound, to me the le classe sounds more friendly and warm. The le monstre has an incredible detail and bass though.
Greetings Joris |
Hi Joris,
I had commercial 20W and 30W Hiraga Classe A at home, and there was absolutely no hum. You should check your ground routing. There are quite a lot of DIY Monsters in Munich, and all of them worked very good. Most folks I know use a regulated power supply for the Monster, which they say sound better than half a Farad of caps. The regulation is that published in L'Audiophile magazin, consisting of a zener, a small signal bipolar, and a power mosfet. I could search for the schematic and scan it, if anybody is interested.
regards,
Hartmut |
|
|
| jazz |
Hi,
Harmut, thanks for you're reply I'd be very interested in the regulated psu and am sure to try it. I am currently designing a project in which i plan to build several amps in a certain style and both hiraga's are apart of this so any tips or hints are welcome. Due to the cost and complexity the project will span a couple of years though. As for that hum, I'm by now suspecting it has something to do with the casing i'm using but changing that will mean a complete redesing and has to wait for the definite plans of my amp project.
regards, Joris |
|
|
| hifidaddy |
Hi,
this power supply for power amps has been designed for Kaneda style class A power amps, but the voltages can easily be changed to your needs, just by changing the voltage reference, which is just a zener diode. For Hiraga, 20..25Volts each rail are good, for "Le Monstre", about 12..14 volts. As the Hiraga and the monster only idle currents in the one amp region, at least for the monster the second Mosfet can be omitted.
regards,
Hartmut |
|
|
| ralf |
Hi,
has anyone expereiences with both, and can tell something about the differences?
Ralf |
|
|
| hifidaddy |
BTW, anybody detected the error in above FET power supply schematic ??
You get 5 electronics engineering experience points (short 5 EEEP ) for that ;-)
regards,
Hartmut from Munich |
|
|
| rickpt |
| the 2sk30 on the negative power suply is not conected properly... |
|
|
| hifidaddy |
regards,
Hartmut |
|
|
| rmgvs |
I did build Hiraga 20 W and Monstre versions for several occasions. As well do I have one original manufactured 20 Watter for almost 15 years now.
The 20/30 Watter to my ears sounds more friendly and tubelike, almost warm. It has a beautiful midrange and a full bottom end and a silky smooth top end. The Monstre has more detail and depth into the music but can sound thin or harsh on occasions (but in its best moments outperforms the 20 W). I also build the Monstre using heavy batteries and 1,5 Farad of buffer: better still but maintaining forementioned peculiarties. I do not agree that the Monstre sounds like a triode single-ended tube amplifier, only in the transparancy of the sound I can see some similarities.
Interesting to know that I have build some cheap versions of the 20 Watter, using a moderate power supply (around 16 Volts DC plus and minus) and using standard 3055/2955, BD139/140 etcetera. Such an amplifier also has a remarkably good sound to it, bettering almost anything commercially available be a good margin. So for a few hunderd Euro's you can build yourself a pretty power amplifier. |
|
|
| rmgvs |
And I forgot:
I did not build any Pass amplifiers but did have a diy-built version of the later DC single-ended mosfet design (with the big power resistors) that gave around 5 Watts of RMS power.
The sound was not up to my expectations (of course not bad, but way behind the Hiraga) and had a horrible hum in it as well. I do not know about the quality of building and/or used components though and my preamplifier was not up to the task (insensitive and lowish impedance this Pass amplifier). So, not a fair comparison this one.
Rudy |
|
|
| Tyimo |
Hello!
I have a strange, a bit stupid question!
Is it possible to create a SE version from the original PP Hiraga Class A (or may be from The Monster) as for example the JE Labs made from the Williamson PP tube amp? ( it could be not the best comparison.....)
I would like to hear the diffrence between the 2 operations and because I am an enthusiast of the SE amps.
Thanks!
P.S.: I built the Monster and it is a realy good amp....... |
|
|
| ericfe1975 |
hello,
after building several class ab an a amplifiers, i have no doubt:
class a is the more natural an sweet sound.
now i want to build "le monstre" but since the transistors of the original schematic doesn't exists anymore, i like to know if some of you have the replacements.
thanks for your answer.:D |
|
|
| scottw |
| quote: | | now i want to build "le monstre" but since the transistors of the original schematic doesn't exists anymore, i like to know if some of you have the replacements. |
This may be of interest:
http://web.vip.hr/pcb-design.vip/hiraga.html
scottw |
|
|
| jacco vermeulen |
I have had a second hand commercial 30w Hiraga, for a large part because it is so beautifull.
Also i built a number of L'audiophile designs; the Preamp of Akihito Kaneda, Le Monstre, L'Ampli Fou.
I even made an effort on the Nemesis, a reason for reading Susan Parker's work with interest.
pcb's and most of the parts i got from the L'audiophile store in Paris, in those days in the Rue de Belfort.
The comparison between Jean Hiraga and Nelson Pass designs is not just i think.
I had the commercial Jean Hiraga JH50 PushPull tube amplifier on ESL's for a couple of days, my ears thought it to perform much better on Quads. Should be, cost a heap more than the 30 watter Hiraga.
I own an Aleph Null, bought it when only ocassionally a second hand commercial Hiraga was offered on sale in an audio store.
I built Aleph 4's, currently adding a number more.
So, i need not mention my preference.
Le Monstre i heard in the L'Audiophile store on typically French very high efficiency speakers, its a brilliant amplifier as the Hiraga is.
To me, putting Le Monstre in a car is the same as the difference between class A and class AB Mosfet amplifiers.
Try stuffing a high efficiency speaker system in a car, supposing you can then savor the beauty of the amplifier over the sound of the motor running and bad vehicle acoustics.
For enjoyment i drive a Chevy van with raised roof. I am 6'2" and able to stand in the 15' x 6.5' livingroom of the car.
The V8 only runs 1500 rpm when doing 60 mph, i can not hear the airco humming.
It has double 4 channel modified amplifiers, for a large part biased in class A with multiple decade MHz Sanken ringemitter output devices that tap from an added battery and two 1 Farad caps.
Sounds quite nice, but i bet a Le Monstre in a regular home on +100 db/watt quality units sounds a lot better. |
|
|
| Panelhead |
I have a strange problem in my Le Monstre. The amp is now where I measure 0.0 mv hum. But this is only without input cabling connected. Once the cable are connected it goes up to 0.2 - 0.3 mv with unterminated shielded short cables. With unshielded it is 0.6 - 0.8 mv ac.
This is with a double shielded power cord. Using an unshielded power cord the measured hum can reach 8 mv depending on cable orientation.
Connecting the preamp to the cables does not help, the measured hum is the same. The preamp has an output transformer with a 38 ohm DCR.
This is the only amp I own that has a problem with hum pickup. The input impedance is around 23K. All my other amps do not have a problem with input cable hum pickup.
On my 100 db efficient speakers the amp is almost dead quiet without cables. I can play with cable dressing and get it lower, but never as quiet as it is without cables.
Laying the cables close to the rear of the amp gives big hum, pickup from power transformer. This is a nice Victoria Magnetics toriod.
Also another goofy question. I have read about the bypassing of the cascode transistors to improve the sound. I tried it in another version, using the PCD Design recommended replacement transistors. It degraded the sound quality.
In this one, original transistors, it has not been tried. Did I do it incorrectly, or is it not needed if the power supply is well built.
One thing learned, this amp can sound great on the right speakers. But it is a devil in exposing power supply issues. Just moving a ground wire connection a couple inches can make the hum go from 0.0 mv to 2 - 3 mv on the outputs.
Anyone thinking of building this amp should be aware. The thing will require some tweaking to get the best out of it. I doubt if anyone has ever built one and had great sonics to begin with.
George |
|
|
| gatedriver |
Hallo Dieter,
you have lots of free space on that board, you could shift the components much closer together and make a tighter layout on the board. that means you need less board area and get a shorter signal path and less parasitics.
try to avoid sharp corners (smaller than 90 deg), make smooth areas (use polygones in your layout software).
you can probably cut the board are in half..
or more, if you use SMD resistors. 1206 or MELF is really easy to solder, no problem for any DIY.
also, the bypass caps are very close to the input stage and there are none close to the output.
Regards
Timo |
|
|
| dieter12 |
Timo,
the board is not so big... Shure, that with the corners is clear, its the first try. No SMD, we would prefer Tantal RsThanks for your commends...
Dieter |
|
|
| ericfe1975 |
thank you guys!
big thanks to appolon scotw and dieter (in no particular order)
i will now study closer all that information
:clown: :D |
|
|
| pred |
Hi everybody,
I am also interested in building le Monstre. Before doing that I wanted to look for "European" transistor substitutions and also to do some spice simulations. Although I am electrical engineer I am not, at least in audio, looking for a perfect amplifier.
But what bugs me the most is the simulation of the output impedance. I used the schematics similar to one posted here with Tip3055 tip2955 or replaced them with smaller MJE15030 or D44H11. I always got the output impedance of 2-4 Ohm. That seems a lot to me for the amplifier with the feedback. Do you have any comment on that?
Since we speak of replacements is there any for jfets SJ SK with some 2nxxxxx or similar that is more likely to be found? The same question is for every other?
Pred
PS
I was using LTSpice if it is of any significance. |
|
|
| plovati |
Regarding to quite high output impedance 2 ohm seems ok, because the monstre exit on collector and was designed to be a 300B solid state equivalent, so with tipical tube damping factor.
Me too prefer to build on ewith current european transistor, but not easy to find replacement for these old toshibe product. High gm input fet that are critical for results seems impossible to find.
Here my list of equivalent (first item of each group is the original one). Any suggestion on equivalent?
POWER NPN 50V 7A 60W 15MHz 2SD844 TOSHIBA
POWER NPN 50V 7A 60W 15MHz 2SD1063 TOSHIBA
POWER NPN 50V 7A 60W 15MHz TIP35A ST
MEDIUM PNP 2SB716 TOSHIBA
MEDIUM PNP ~MJE350 ON SEMI
SIGNAL NPN 2SC1775E TOSHIBA
SIGNAL NPN ??
JFET NCH 2SK170Y TOSHIBA
JFET NCH 2SK364 TOSHIBA
SIGNAL PNP 2SA872E TOSHIBA
SIGNAL PNP ??
JFET PCH 2SJ74Y TOSHIBA
JFET PCH ??
MEDIUM NPN 2SD756
MEDIUM NPN ~MJE3440 ON SEMI
MEDIUM NPN 2SC2632 TOSHIBA
POWER PNP 2SB754
POWER PNP TIP36A ST
__________________
Piergiorgio |
|
|
| Panelhead |
www.mcminone.com can supply all the original transistors. I built one using the recommended substitutions listed on the PCB-Design website. Later found the original transistors and built another.
The one built with the original transistors sounded much better. It also was more stable.
One tip I missed. Keep your ground length paths equal from rail to rail and channel to channel. Small differences in the ground paths will increase hum and effect dc offset.
Poor layout and build can cause big problems with stability. It will zing your eardrums.
George |
|
|
| danieljw |
hello people
I just built "Le Monstre" using origional transistors
tried it last night.... I have to say the amplifier sounds really really good
has any one else tried it ???
pictures to come.
-Dan |
|
|
| Tyimo |
Hi Dan!| quote: | | I have to say the amplifier sounds really really good has any one else tried it ??? |
Yes, me too, but with unregulated giga PSU like the original. Not bad. It is good for a reference PP amp.
Greets:
Tyimo |
|
|
| danieljw |
Hi There,
Yes i tested this amp and found very good results both using a reactive speaker load and a resistive load.
Have you built the 20W or 30W hiraga ? I want to build that one next
-Dan |
|
|
| Tyimo |
I built the Le Monster, but I know the Class A (20W version) also very nice sounding amp. Maybe with a bit warmer sound.
Tyimo |
|
|
| bluevas |
| what is the good speaker for le monstre ? Any suggestion? Because it's only 8 watt. thank you guys |
|
|
| danieljw |
Well
Hiraga Mentioned using the onken style enclosures and voice of the theatre enclosures with this amp
but they are a little large and hard to get :)
i personally have used 86dB speakers with no problems but that is a 5 x 8m room.
-Dan |
|
|
| HKC |
Hi to all
What idss class (GR/BL/V) of the 2SK170/2SJ74 should be used for the Hiraga "le monstre" power amp to get the best sounding result?
Best Regards |
|
|
| lineup |
| quote: | Originally posted by HKC
What idss class (GR/BL/V) of the 2SK170/2SJ74 should be used
for the Hiraga "le monstre" power amp
to get the best sounding result? |
hey.
This nobody knows. you just would have to try and listen.
As you ask about 'sound'.
If you mean technically best data of operation into a specific LOAD
it may be easier to try to test or simulate an optimal bias of input stage.
The short answer is, the input FETs should be biased
to get the desired Class A idle current in Output stage.
This current level, would be your choice.
would be different for different wanted power output into different LOADs.
4, 6 or 8 Ohms speakers
This version of Hiraga, in image, uses 2SK246 / 2SJ103
and the bias is set to like 4 mA, for ~ 1 Ampere in output stage.
But I am sure you use some other version / setup
for the Hiraga monster.
 |
|
|
| HKC |
| quote: | Originally posted by lineup
hey.
This nobody knows. you just would have to try and listen.
As you ask about 'sound'.
If you mean technically best data of operation into a specific LOAD
it may be easier to try to test or simulate an optimal bias of input stage.
The short answer is, the input FETs should be biased
to get the desired Class A idle current in Output stage.
This current level, would be your choice.
would be different for different wanted power output into different LOADs.
4, 6 or 8 Ohms speakers
This version of Hiraga, in image, uses 2SK246 / 2SJ103
and the bias is set to like 4 mA, for ~ 1 Ampere in output stage.
But I am sure you use some other version / setup
for the Hiraga monster.
|
Hi Lineup
Thank you for your answer.
Yes I meant technical best data. My speaker load is 8 Ohm and I will use parts as specified on original schematic which the input JFets are 2SK170 & 2SJ74
Best Regards |
|
|
| danieljw |
Hi All,
Lineup = interesting answer :)
HKC = see the link in my signature line or google the "class a amplifier site" it has the info in the jean hiraga index.
-Dan |
|
|
| lineup |
| quote: | Originally posted by HKC
Hi Lineup
Thank you for your answer.
Yes I meant technical best data. My speaker load is 8 Ohm and I will use parts as specified on original schematic which the input JFets are 2SK170 & 2SJ74
Best Regards |
| quote: | Originally posted by danieljw
Hi All,
Lineup = interesting answer :)
HKC = see the link in my signature line or google the "class a amplifier site" it has the info in the jean hiraga index.
-Dan |
I forgot.
I think you should buy/use 2SK170BL + 2SJ74BL
They should be like max 8-10 mA, I think.
They are also the easiest variant to find. Nelson Pass uses often 2SK170BL
( GR variant is only like half as much current. And this may be too little )
Refereing to image in my post.
Depending on different JFET, you may even have to use 220 or 470 Ohm potentiometer.
The procedure for seeting like 1 A current in output, can take some time.
For 8 Ohm speakers, this would be my initial current setting.
When you have 1 Volt DC across both 1 Ohm resistors, you will have 1 Ampere running.
What factors effects (most) how much current/voltage across 1 Ohm resistors?
Several things:
1. What JFETs.
And different exemplars of same 2SK170 can differ.
2. THe potentiometer.
220 Ohm = more current.
470 Ohm pot less current.
3. Value of '390 Ohm's resistors.
Higher value = more ampere in output.
Lower value less current.
====================================
Visit Dan's homepage, like he said. Inpprvious post.
I am also sure he can guide you well
how you should do, to trim and adjust Le Monstre
in a safe and good way.
... because if you are unlucky ... you may do a BurnOut
.. at first Power on
:D |
|
|
| Lumba Ogir |
lineup,
can you estimate the open-loop gain and feedback level of your version? |
|
|
| lineup |
| quote: | Originally posted by Lumba Ogir
lineup,
can you estimate the open-loop gain and feedback level of your version? |
I have not built Le Monstre.
I have only looked at schematic and read a bit.
but,
some day soon, I thought I should setup this nice Class A amplifier
in my MultiSim spice. (torrent download pirate cracked pro version)
And then I may tell about open loop and feedback level
and THD, AC bandwidth and show the harmonics Fourier spectrum.
How much 2nd, 3rd .. 7th dist ...
I guess this is a rather ear-friendly distortion (Class A + JFETs).
Right now I am a bit busy.
To do a spice and fair simulation of this classical amp is easy, very easy.
Because use very common transistors/jfet.
Such that there are very good virtual spice models for.
And is not very complex.
Mr. Hiraga really knew about Keep-It-Simple-And-Good :cool:
regars :) |
|
|
| Lumba Ogir |
| thank you lineup, I am waiting... |
|
|
| lineup |
| quote: | Originally posted by Lumba Ogir
lineup,
can you estimate the open-loop gain and feedback level of your version? |
| quote: | Originally posted by Lumba Ogir
thank you lineup, I am waiting... |
Lumba.
You should really have a read on a Nelson Pass paper from 2005.
.. unless you have already studied this great paper :cool:
See his investigation on feedback factor, in 4 rather classical amplifiers.
JLH is one of them.
I have attached a little teaser image.
Go download at www.passdiy.com
The PLH Amplifier: The classic JLH - Pass style
Then you thank me, lineup, for telling.
But most of all give a nice thought to Our One and Only Nelson.
For doing a good audio work .. and sharing it with us :cool: |
|
|
| Lumba Ogir |
lineup,
well, Hiraga`s design is simply genial, I`ve never seen anything so eccentric neither before nor after.
You are about to raise the open-loop gain, I would go in the other direction... |
|
|
| HKC |
| quote: | Originally posted by danieljw
Hi All,
Lineup = interesting answer :)
HKC = see the link in my signature line or google the "class a amplifier site" it has the info in the jean hiraga index.
-Dan |
Hi Daniel
Your Le Monstre link is very interesting.
According to the power supply schematic. You use 2 LM338K for both of the upper and lower section. As far as I know, LM338K is a positive regulator therefore, should there be 1 LM338K and another negative regulator to form the regulator section instead of 2 LM338K? In addition, the two voltage rails are specified as -12V. Is this typing error? I think they should be +12V & -12V respectively. Correct me if I am mistaken.
Back to the Le Monstre schematic. Is the 100 Ohm trim pot use to adjust the voltage at output? Please advice.
Thanks & regards |
|
|
| HKC |
| quote: | Originally posted by lineup
I forgot.
I think you should buy/use 2SK170BL + 2SJ74BL
They should be like max 8-10 mA, I think.
They are also the easiest variant to find. Nelson Pass uses often 2SK170BL
( GR variant is only like half as much current. And this may be too little )
Refereing to image in my post.
Depending on different JFET, you may even have to use 220 or 470 Ohm potentiometer.
The procedure for seeting like 1 A current in output, can take some time.
For 8 Ohm speakers, this would be my initial current setting.
When you have 1 Volt DC across both 1 Ohm resistors, you will have 1 Ampere running.
What factors effects (most) how much current/voltage across 1 Ohm resistors?
Several things:
1. What JFETs.
And different exemplars of same 2SK170 can differ.
2. THe potentiometer.
220 Ohm = more current.
470 Ohm pot less current.
3. Value of '390 Ohm's resistors.
Higher value = more ampere in output.
Lower value less current.
====================================
Visit Dan's homepage, like he said. Inpprvious post.
I am also sure he can guide you well
how you should do, to trim and adjust Le Monstre
in a safe and good way.
... because if you are unlucky ... you may do a BurnOut
.. at first Power on
:D |
Hi Lineup
Thank you for your detailed reply.
I have some 2SK170bl/2SJ74BL available to use for this power amp. I have to measure to see what is the idss of these JFets.
What is the bias current at the output stage of the original design of the Le Monstre? I would like to set the bias current same as the original design.
Thanks again. |
|
|
| Sigurd Ruschkow |
Using MicroCap, I simulated the le Monstre, see schematic attached, and CL gain is 24 dB and OL gain is only 37 dB.
THD is about 0,07% at 1kHz and 8Vpeak output.
It is quite amazing that OL gain is flat up to 1MHz!
I have used the GR grades of SK170/SJ74 with Idss = 4mA,
and the pot is 200 Ohm, as to get about the same drain
currents in the JFETs as in the original design.
Lumba,
if you download the free eval version of MC, you can simulate the le Monstre - it is easy as pie.
You can get my MC file if you wish.
Sigurd
| quote: | Originally posted by Lumba Ogir
lineup,
can you estimate the open-loop gain and feedback level of your version? |
| quote: | Originally posted by Lumba Ogir
lineup,
well, Hiraga`s design is simply genial, I`ve never seen anything so eccentric neither before nor after.
You are about to raise the open-loop gain, I would go in the other direction... |
|
|
|
| Sigurd Ruschkow |
500m - 600mV over 1 Ohm is the Ib value,
which is too little, IMO, to get 8W into
8 Ohms. Into 4 Ohms you will only have half of what you get at ( Ohms (into class A).
I have played with the Hiraga le Monstre and am running at
2.5A idle current and that at about 20V rails using 0302/0281 Onsemi transistor. Needs some heavy duty fan cooling :cool:
| quote: | Originally posted by HKC
What is the bias current at the output stage of the original design of the Le Monstre? I would like to set the bias current same as the original design.
Thanks again. |
|
|
|
| Sigurd Ruschkow |
JLH = John Lindsey Hood and not Hiragas amp.
Output impedance for the le Monstre is a few Ohms.
Sigurd
| quote: | Originally posted by lineup
Lumba.
You should really have a read on a Nelson Pass paper from 2005.
.. unless you have already studied this great paper :cool:
See his investigation on feedback factor, in 4 rather classical amplifiers.
JLH is one of them.
I have attached a little teaser image.
Go download at www.passdiy.com
The PLH Amplifier: The classic JLH - Pass style
Then you thank me, lineup, for telling.
But most of all give a nice thought to Our One and Only Nelson.
For doing a good audio work .. and sharing it with us :cool: |
|
|
|
| Lumba Ogir |
HKC,
The bias currents:
Input 1mA
Output 500mA
Sigurd,
nice to have you around. Thanks again.
Would this work? |
|
|
| Sigurd Ruschkow |
Lumba,
I have attached my le monstre file. To use it in MicroCap
just rename the file to *.cir
and you can play with it, and also add your own ideas
as you have another 25 components to add to my CIR before
the evaluation version says stop.
Enjoy!
Sigurd
| quote: | Originally posted by Lumba Ogir
HKC,
The bias currents:
Input 1mA
Output 500mA
Sigurd,
nice to have you around. Thanks again.
Would this work? |
|
|
|
| lineup |
| quote: | Originally posted by Lumba Ogir
HKC,
The bias currents: Input 1mA Output 500mA |
Hello there.
Looks like another guy did the job for me. So I wont have to simulate and find out feedback factor ;)
Now.
Almost any classical amp, is around in several variants. Often with some exchanged, 'more modern' transistors. We see designers make Revised versions. This is not the first, original version. But often improved in some way.
I refer to such a new variant. And this is what I will do, further in this topic.
Because, if I built Le Monstre, there is no way I would be able to use the original devices, anyway.
It is this new version I mean. With 2SK246 / 2SJ103 as good new substitute JFET pair.
Hiraga Le Monstre - Suggested new transistor substitutes.
Even if the numbers, currents can/will be different for the original, my advices, as method, can be used for any Le Monstre version. I mean what factors will determine the idle currents in output stage.
As I mention in a previous post
Also, without knowing the original current,
I suggested an initial value of 1.0 Ampere for 8 Ohm speakers.
It seems a good value to start with .. and can be increased later, when we got everything working and want to try with more current.
Sigurd suggests 2.5 A, which seems a bit high to me. Of course for speaker Z dips (downto 4-5 Ohm) and for 4 Ohms speakers we can really need this high currents.
What me and Sigurd agree on, is that the original 0.5/0.6 Ampere is a bit low. At least for modern speakers and operating at the higher end of the power (say 5-7 Watts RMS/8Ohm), and that peformance probably would be better with a bit more current.
| quote: | Originally posted by Lumba Ogir
Would this work? |
Can you explain your idea more, Lumba.
I do not get it from your image.
lineup :) |
|
|
| Lumba Ogir |
lineup,| quote: | | I would think that this 2SK246 / 2SJ103 version is really by Hiraga. But I do not know. |
You do not know.| quote: | | Sigurd suggests 2.5 A, which seems a bit high to me. |
It`s fine...for a while.| quote: | | Can you explain your idea more, Lumba. I do not get it from your image. |
I blame the software.
It is an attempt to linearize the driver by increasing its emitter resistance. |
|
|
| lineup |
| quote: | Originally posted by Lumba Ogir
lineup,
I blame the software.
It is an attempt to linearize the driver by increasing its emitter resistance. |
I see.
Yeah, the optimal value can be different, at different output levels + different loads (normally I do not use pure resistors for load, when I listen to my amps)
Nelson Pass puts it like this:| quote: | There is a sweet spot
for a every transistor in a given circuit.
For a special job to do. |
This means an optimal resistance to set the 'best' average current of operation.
Such a resistor is almost always a compromize.
For the best operation at low levels of output and a high.
For normal (1kHz)audio frequencies music and for higher frequencies.
When designing such a blameless, good compromize amplifier
that will work with good quality in many situations,
it may require a bit of tweaking up & down, until finding 'the optimal compromize'.
lineup :) solen lyser här i norra sverige nu, Lumba! |
|
|
| Sigurd Ruschkow |
I actually have used 2.8 A at 20V rails for a long time and as long as one has proper cooling, there are no problems.
I do use a 350mm long heatsink with 3x 80mm mains powered fans for a stereo setup. See attached image.
I tried your idea with degenrating the drivers, Lumba, but distortion increased.
SIgurd
| quote: | Originally posted by Lumba Ogir
lineup,
You do not know.
It`s fine...for a while.
I blame the software.
It is an attempt to linearize the driver by increasing its emitter resistance. |
|
|
|
| Lumba Ogir |
lineup,
an emitter resistor is the simplest linearization technique, although very effective, proved by listening tests, the only test method I trust. It`s extremely reliable and relevant. |
|
|
| Lumba Ogir |
Sigurd,| quote: | | I tried your idea with degenrating the drivers, Lumba, but distortion increased. |
That`s a very good sign, THD rises as it should.
lineup,
start planning for the implementation! |
|
|
| HKC |
| quote: | Originally posted by danieljw
hello people
I just built "Le Monstre" using origional transistors
tried it last night.... I have to say the amplifier sounds really really good
has any one else tried it ???
pictures to come.
-Dan |
Hi Daniel
Would you please tell where did you get the original transistors for Le Monstre? I tried to order from MCM Electronics but they do not ship parts outside USA.
Best Regards |
|
|
| HKC |
| quote: | Originally posted by Sigurd Ruschkow
500m - 600mV over 1 Ohm is the Ib value,
which is too little, IMO, to get 8W into
8 Ohms. Into 4 Ohms you will only have half of what you get at ( Ohms (into class A).
I have played with the Hiraga le Monstre and am running at
2.5A idle current and that at about 20V rails using 0302/0281 Onsemi transistor. Needs some heavy duty fan cooling :cool:
|
Hi Siqurd, Lumba & Lineup thank you guys for giving your comments about idle current. I would like to set the current @ 1A as suggested for 8 Ohm load speakers so in this case, do I need to change the value of the emitter resistors of the power transistors? If yes, what is the resistors value would be?
Best Regards |
|
|
| lineup |
| quote: | Originally posted by HKC
I would like to set the current @ 1A as suggested for 8 Ohm load speakers so in this case, do I need to change the value of the emitter resistors of the power transistors?
If yes, what is the resistors value would be? |
| quote: | Originally posted by lineup
The procedure for seeting like 1 A current in output, can take some time.
For 8 Ohm speakers, this would be my initial current setting.
When you have 1 Volt DC across both 1 Ohm resistors, you will have 1 Ampere running.
What factors effects (most) how much current/voltage across 1 Ohm resistors?
Several things:
====================================
1. What JFETs you use.
And different exemplars of same 2SK170 /SJ74 can differ.
2. THe potentiometer.
50 OHM or (47) pot = most current,
100 Ohm pot = normal
220 Ohm = less current.
3. Value of 1 k Ohm's resistors.
Higher value = more ampere in output.
Lower value less current.
====================================
Visit Dan's homepage, like he said. In previous post.
I am also sure he can guide you well how you should do, to trim and adjust Le Monstre. in a safe and good way.
... because if you are unlucky ... you may do a BurnOut .. at first Power on :D |
You have already forgotten my post ;)
I tell 3 things that will effect How much current in output.
1)
- I suppose, you use GR variants of JFETs.
- They will suit the original schematic best.
With something like ~ 1.7 mA in input stage.
Across 1 kOhm resistors at Drain side.
Measure voltage across those resistors, without adding the output stage.
The output stage is:
(2SB716 / 2SB756 / 2SD844 / 2SB754 transistors.)
Balance with trimpot until about same voltage across both positive and negative 1 kOhm resistor.
Trimpot center should be attached to GND, when you have not attached output stage.
1 Volt across 1 kOhm = 1 mA
1.7 Volt across 1 kOhm = 1.7mA, this is the value that should give like 1 Ampere across 1 Ohm
(for the schematic with 2SK246 / 2SJ103, I refered to earlier, we should use 2SK170BL / 2 SJ74BL .... something like 4 mA in input, across 390 Ohm resistors in DRAINS)
2)
The potentiometer, should be chosen to give like 1 - 2 mA in your JFETs.
In the original 100 Ohm trimpot.
If you want more current ( measure ), then you may have to change to 47 Ohm trimpot.
If you want less current, change to 220 Ohm trimpot.
3)
When you have like 1-2 mA. Or better 1.7 - 1.8 mA.
Then, you could temporarily replace both 1 kOhm resistors
with 2k2 trimpots.
Start with a low value in them pots, before you power on.
Then you can carefully increase the trimpots resistance value, on both sides.
While all the time balance for 0 Volt at output terminal.
When you have reached 1 Volt across BOTH 1 Ohms resistors,
you can replace those two 2k2 potentiometers with suitable resistors.
With closest resistor value.
I suggest you should NOT change the 1 Ohm resistors, in output
It is nice to know that 1 Volt is = 1 Ampere. For control!
-------------------------
So, sum is:
1. Try to get 1-2mA in input 2SK170GR / 2 SJ74GR (or 2SK170BL / 2SJ74BL )
This you do with use the best suitable potentiometer in '100 Ohm' place.
Can be 47, 100, 220 trimpot.
2.. Change the 1 kOhm resistors value
(1.2 kOhm will give more current in output, 820 Ohm will give less)
until you get 1 Volt across BOTH 1 Ohm.
Balance all the time with trimpot.
lineup :)
-------------
Ps.
Hiraga was clever. He has chosen 1 Ohm and 1 kOhm resistors
to make it easy for us.
1 Ohm and 1 kOhm values we often have at home. We can buy many of those values, and get them for lower price.
And this way we can easy measure.
1 Volt into 1 Ohm = 1 Ampere, 0.6 Volt = 0.6 A.
Same with 1 kOhm:
1.00 Volt = 1.0 mA, 1.70 Volt = 1.7 mA, 2.00 Volt = 2.0 mA |
|
|
| HKC |
Hi Lineup
Yes I overlooked your post. I am sorry for that.
I need to take some time to go through all the details. Once again thank you for sending the information, very appreciated.
Have a great weekend. |
|
|
| lineup |
| quote: | Originally posted by HKC
Hi Lineup
Yes I overlooked your post. I am sorry for that.
I need to take some time to go through all the details. Once again thank you for sending the information, very appreciated.
Have a great weekend. |
Yes, it has been many other posts here. Which is good when people are interested :)
There is nothing so bad as when nobody posts, when you need advice.
I wrote my last post very quickly.
But hope you can get the procedure I would use.
--- Your first setup should be without output stage.
When you get 1.7 Volt across BOTH those '1 kOhm' resistors ( using 820, 1000 or 1200 Ohm)
Then you may attach output stage.
Without loudspeaker, PLEEEAAASEEEE
Instead you HAVE TO USE a dummy load power resistor.
One big 5-10 Watt power resistor, in the place of Speaker.
The value can be like 10 - 22 Ohm.
Take what you have.
also first speaker you attach to an amplifier under construction
.. :att'n: should be some old cheap one, something you wont miss much,
if it get destroyed :D :D
:att'n: |
|
|
| Lumba Ogir |
lineup,
you are a dreamer, right, floating around... |
|
|
| danieljw |
| quote: | Originally posted by HKC
Hi Daniel
Your Le Monstre link is very interesting.
According to the power supply schematic. You use 2 LM338K for both of the upper and lower section. As far as I know, LM338K is a positive regulator therefore, should there be 1 LM338K and another negative regulator to form the regulator section instead of 2 LM338K? In addition, the two voltage rails are specified as -12V. Is this typing error? I think they should be +12V & -12V respectively. Correct me if I am mistaken.
Back to the Le Monstre schematic. Is the 100 Ohm trim pot use to adjust the voltage at output? Please advice.
Thanks & regards |
| quote: | Originally posted by HKC
Hi Daniel
Would you please tell where did you get the original transistors for Le Monstre? I tried to order from MCM Electronics but they do not ship parts outside USA.
Best Regards |
Hello HKC,
Wow this thread got interesting all of a sudden !
The power supply uses 1 LM338K per rail. and is +/- 12V d.c.
The transistors were sourced by another diyaudio member via ebay.
If you do a search in ebay you may be able to get some still.
If not I have some of the transistors left.
-Dan |
|
|
| HKC |
Hi Daniel
May I buy your leftover transistors? If possible I want the following quantity:
2SB754: 6 Pcs.
2SB716: 6 Pcs.
2SC1775: 6 Pcs.
2SA872: 6 Pcs.
2SD844: 6 Pcs.
2SD756: 6 Pcs.
I have 2SK170BL/2SJ74BL so I do not need these JFets.
Please advice.
Thanks |
|
|
| danieljw |
Hi HKC,
I better send you a PM. |
|
|
| HKC |
Hi Daniel
I received your pm. I will advice shortly.
Thanks & have a nice day |
|
|
| Sigurd Ruschkow |
Lumba - what are your thoughts regarding inserting degeneration resistors for the?
Wanna go OL?
Sigurd
| quote: | Originally posted by Lumba Ogir
Sigurd,
That`s a very good sign, THD rises as it should.
lineup,
start planning for the implementation! |
|
|
|
| Lumba Ogir |
Sigurd,
briefly, it improves the sound subjectively but we should not discuss this matter as the chances of consensus are negligible. |
|
|
| Sigurd Ruschkow |
Concensus? Huh, what is the point in concensus? Not sure there would be much progress i if concensus was the goal...
I am all ears!
Sigurd
| quote: | Originally posted by Lumba Ogir
Sigurd,
briefly, it improves the sound subjectively but we should not discuss this matter as the chances of consensus are negligible. |
|
|
|
| danieljw |
here are some test results from my le monstre amp
1KHz sine full output |
|
|
| danieljw |
| 10kHz sine full output (loaded 8 Ohm) |
|
|
| danieljw |
| 10kHz square 8 Ohm load |
|
|
| danieljw |
| stability when clipping @ 1kHz 8 ohm load |
|
|
| danieljw |
| 300 kHz sine @ 8 ohm Down approx 3dB here |
|
|
| Sigurd Ruschkow |
Lumba, I read that post a few days after you wrote it. That is why I thought you wanted to go OL and not only degenrate-
Would be interesting with a non global feedback le Monstre
(I have one in MC) and compare sonics.
My simulated OL le Monstre has 25dB gain, 100kHz BW,
and THD can be 2nd, 3rd, or a mix of your choice and THD is about 0,5% at 1W/ 8 Ohm.
Sigurd
|
|
|
| HKC |
| Here below is my pcb layout of Le Monstre, comments are welcome: |
|
|
| Sigurd Ruschkow |
HKC- here are some ideas:
* make the tracks that carries heavy currents MUCH wider. As wide as you can make them, ie make copper pours.
* Put the input JFETs so that they "kiss" eachother, so you can later on glue them together
* Move the power resistors to the sides and put the drivers close to the output transistors
* Move the long feedback wire more to the middle of the PCB.
* Move the cascode transistors more to the middle an d closer to the JFETs.
* Maybe a ground plane?
Hirga has a 47u cap on each rail. Did you skip that or did I miss it?
You might want to decouple the cascode resistors with a small cap (say 100n) so that the cascode voltage going in to the small BJTs is low pass filtered. It can give slighly better sonics.
Attached is my layout which has more PS decoupling (incl input section low pass filtering) than the original le Monstre and also an input filter and a feedback capacitor. I also use TO220 power resistors with a heatsink instead of the standard cylindical ones,
and have LEDs power-on indication.
Size is 160x100mm.
Sigurd
| quote: | Originally posted by HKC
Here below is my pcb layout of Le Monstre, comments are welcome: |
|
|
|
| Lumba Ogir |
Sigurd,
it is a beautiful design, excellent sonics and driving capability, indicating genuine sense for audio. The feedback is probably not very harmful here.
Running the FETs at 1mA, instead of 2-3mA, I suspect Hiraga`s intention was to keep the open-loop gain low.
Emitter resistors raise THD, which does not look nice in simulations. |
|
|
| qwad |
:confused: l'm intrigued by this design, just one question though, can other devices be used ?; i.e: outputs, or the small signal. l ask this as in one schematic l see the use of bc550/ 560 devices besides the j fets......... cheers, live long and prosper
T.C. |
|
|
| Sigurd Ruschkow |
Lumba-
Zeners are too noisy in that position - add a cap to filter it out,
or do as I plan to do, and use a blue LED.
The sound of the le Monstre much depends on the power supply as with most power amps,
and also depends on the type of passive components one uses.
As I am running at approx 2.5A idle current, I have skipped the 8 Ohm series resistor in the PS and used a choke instead,
and also used a dual bridge rectifying circuit with plenty decoupling.
Sigurd
| quote: | Originally posted by Lumba Ogir
Even better: |
|
|
|
| Sigurd Ruschkow |
Yes, you can. The design is rather forgiving when it comes to different types of semis.
The output devices can for ex be the very good 0281/0302 from OnSemi or the bigger brother of these.
Some even use the 2N3055 and its complementary...
But I like to use the much faster ones from Onsemi.
The cascode transistors can be 2SC2240/SA970 or smiliar,
and the JFETs can be of other types, too, but you need
to see to that the current is approx correct through them by changing the pot.
You could even use JFETs for the cascode transistors.
The drivers I have not played with as I use the original types.
Sigurd
| quote: | Originally posted by qwad
:confused: l'm intrigued by this design, just one question though, can other devices be used ?; i.e: outputs, or the small signal. l ask this as in one schematic l see the use of bc550/ 560 devices besides the j fets......... cheers, live long and prosper
T.C. |
|
|
|
| Lumba Ogir |
Sigurd,
zeners are nasty noise generators causing ugly distortion. They stay on the top of my blacklisted parts toghether with ceramic capacitors and opamps. IC voltages references should be used. LEDs are noisy too, with a zener-like behavior.
| quote: | | The sound of the le Monstre much depends on the power supply as with most power amps, |
All power amps.| quote: | | I have skipped the 8 Ohm series resistor in the PS and used a choke instead, and also used a dual bridge rectifying circuit with plenty decoupling. |
A very wise decision. |
|
|
| Lumba Ogir |
qwad,
Hiraga selected the parts with extreme care.
It would be of crucial importance to use some J-FETs from Toshiba.| quote: | | You could even use JFETs for the cascode transistors. |
No. Not at all. Those should be bipolars. |
|
|
| Sigurd Ruschkow |
Lumba -
Here we disagree. JFETs work fine for cascode work. I have many amp with JFETs for cascodes. Even MOSFETs could be used.
How rock solid the cascode voltages are, can make a difference aswell as how clean they are.
I guess your 5V "Zener" is not a Zener but something else, right?
2Sj74 and 2Sk170 are still widely available in both Europe and the US - get them while the last.
Sigurd
| quote: | Originally posted by Lumba Ogir
qwad,
Hiraga selected the parts with extreme care.
It would be of crucial importance to use some J-FETs from Toshiba.
No. Not at all. Those should be bipolars. |
|
|
|
|
