Hiraga 20W class A

[damon1983]

Thank you. That's what I try to convey. I was once guilty of refusing to use SMPS just because I saw the high frequency glitches on the line that you need to be careful using filter. A small toroid to wrap a few turns on the cable from the output of the supplies, filter caps, good layout on pcb so it's less susceptible to EMI radiation, then it's all good.

I worked in designing mass spectrometers/semi conductor test equipments, we measured very low current and voltages. We used to have racks of linear regulated supplies until I decided to switch all to SMPS. They worked great, we starting to have small SMPS modules on pcb and all. It saved a lot of space, power.

Well, of cause, if people are using an imperfect supply to create sag to give certain effect on the sound like guitar amps, then it's a different story. But if stiff supply rail is what you are after, I cannot imagine you can do better with an unregulated transformer/rectifier/big filter caps. In fact, I don't think you can even tough the regulation of an SMPS.

Problem with using SMPS for normal Class AB is because you need 48V and higher, SMPS cost a lot of money for higher voltage, that make it not worth your while. Also, the normal design of the 2EF/LTP topology has much better PSRR than the Hiraga, sag is no even an issue.

But for Hiraga that use 20V rail, it's perfect to use 24V SMPS that is cheap like dirt. You just adjust down to 20V!!! You want sag, put a 0.22ohm and filter caps like the schematic of the supply used!!!

I cannot imagine the SMPS will droop more than any of the transformer. Still worry, get a 20A+ SMPS, you are not going to droop!!!

I didn't have any problem with 50 Hz and 100 Hz noises with small filter (2*0,202F).

 

[damon1983]

I can't read the language!!!

... Scary...
http://www.diyaudio.com/forums/solid-state/193-hiraga-20w-class-67.html#post4156713

 

[sergiu2009]

At last, my work is done. My Jean Hiraga Super class A

Hello my dear colleagues,
I can now proudly present, after one year of construction at this amp and allot of hours of labour for the case, wiring, listening tests... I had to say that I tested approx all the possible variants of components, and wiring (wire gauge, earthing, and cable managemnt),except for the regulated psu, Kubota regulated psu mod for input and CLC mode for output. ( keep these mods for later upgrade)...
Here are the specs:
1. Two 300w trafos, bicafillar winded (equal impedance on both rails) ,custom made, one for each channel.
2. Ultrafast soft recovery high voltage high current diodes + 484000 uF filtering (22 pcs of 22000uF/35v for each channel CRC): 66000uF+0.22ohm/30w+176000uF/35v //per rail
3.Four small 10 w paralleled secondaryes for equal impedance, two per channel giving 20 w of juice just for the input.
4. Ultrafast soft recovery low voltage low current diodes + CRCRC psu for the input psu : 10000uF/35v+2*1ohm/3w paralelld resistors+ 10000uF/35v+ 3*1ohm/3w paralelld resistor + 2*1000uF+1*680uf low resistance Nichicon caps giving a total of 22680 uF per rail.
5. All of the components are matched on the amps within 1% passive components, 1% input trannies, 5% drivers and 8 % matching for outputs. I had chosen the drivers and outputs after listening to them in the long listening and exhausting session nights, and from 5 different series of outputs and drivers with different cobs, ft, and Pd's (on semiconductors, sanken, toshibas).
6. The amp receives 23vcc for the input stage and 27.2 vcc for the output on each channel, and a bias of 1.65 A idle current (sounds way relaxed and better at this current) and its outputing on 1000hz sinusoidal signal 38w@8ohm sinus into dummi load at clip and 35w into my speakers with ease.. The idle stays rock sollid after 30 mins of varming up at 2mV on left and 4mV on right channel.
7. Soft start (diy audio version).
8. A big radiator for each output, with no insulator between each output trannie and the rads, only Prolimatech silver thermal grease and the rads completely isolated from the chasis. As you can see they are optimally oriented with the fins up and trannies bellow for the best heat dissipation because I saw when you rise the bias on this amp the temp rises rapidely, fact wich made me change two sets of rads till now, and tgis is the third pair. The is allot of heat dissipation here and if you want a longer life for the oitputs you have to go with the biggest rads that you can find.
Now they are running at 48 degrees celsius per output trannie @ 19 degree ambient in the winter..
I think that the whole amp has aprox 30-35kilos and it is really hard to lift. It has wheels for moving.

Almost everithing is made from scratch.

I will open a new thread with all of my findings about this amp till now and about my choices for the different components.

Thank you all for the help to all of you : 6L6, ZenMode, Ovidiu, AlexxMM, and all the other colleagues from the Hiraga 20w amp thread who helped me with good advices and sugestions.
Cheers to you all.
Altought it hasnt got an audiophille music source, now its runing on ipod classic 5.5 @ wolfson dac chip and Rockbox moded, it still blows you away with its realism, sweetness, soundstage, bass, controll, dynamics,and level of details.

Cheers
Sergiu

 

[sergiu2009]

SECOND SET OF PICS

SECOND SET OF PICS

 

[Arick]

Good job Sergiu,
It does look different!
Rick

 

[sergiu2009]

Good job Sergiu,
It does look different!
Rick

Thank you. The design is great, but sounds far better than it looks.
Now I have too look at a preamp for it. I was looking at a Kaneda preamp but there are very few comments and materials or articles on the net about it...
It is curious that I havent found a preamp from Hiraga other than for turntable.. I dont know why. Its strange.
Cheers
Sergiu

 

[bhpistorqintorpm]

@sergiu2009 that's monstrous looking monster [30W @Claas A]! interesting and solid!

phew! i almost read through the complete thread in one go! even i am planning to start a dual mono build of Jean Hiraga 30 W .
to start with , which of these two PCBs is better for the build? can anybody please help? these are from Jim' Audio listed on ebay.

Stereo amplifer PCB Hiraga super 30W class A !!!


or

Stereo amplifer PCB Hiraga super 30W class A w/ current source + kubota reg !!

 

[67339]

different hiraga amps

Did you hear Jean Hiraga's 50W amp?

Hello,
I did hear several solid state designs in France. Lost count but i did visit that shop at least 30 times. Not all of them were by Monsieur Hiraga.
Usually they were using Voice of the theatre ( altec bass bin and sandfilled wooden horn with TAD2001 driver for 600 hertz up see picture) which sounds very nice with the 20 or 30 watt Hiraga amps but also sounds nice with single ended triodes.
But the bigger amplifiers ( not by Hiraga himself) were rather complicated to build. A lot of active parts which needed to be matched i guess. I did meet several people who did have problems with these amps. Once did visit a show where one of the amps did have an '' oscilating sound ''
The 20/30 watts hiraga are a much easier bet. If you can have transistors that are kind of matched you are half way.. What i did also do is using the same '' heatsink '' for the drivers. Of course you will have to isolate them from the heatsink to not cause a short circuit !!. I remember if you touch one of the driver which will make its temperature go up DC at the output will change. Touching other driver dc would change into opposite direction!
Nobody so far did start using Lundahl choke instead of the power resistor in the power supply. Dont worry to much about some extra dc resistance and current rating. In the Hiraga there isnt much of voltage across the choke so it can handle extra current before it will saturate.
The choke(s) is (are) a small investment but typical diy others should try first because everyone is on a budget lol.
Greetings, Eduard

 

[wahab]

Hello,

But the bigger amplifiers ( not by Hiraga himself) were rather complicated to build. A lot of active parts which needed to be matched i guess. I did meet several people who did have problems with these amps. Once did visit a show where one of the amps did have an '' oscilating sound ''

This was surely the Kaneda design that was used fo a 50W class A.

Kaneda made the mistake to undercompensate his amp with a tiny 5pF capacitor since he wanted to keep a 20KHz open loop bandwith, yet with a regular 33-47pF cap not only his amp is well stabilised but the distorsion profile is also left totaly unchanged.

 

[sergiu2009]

Hello my friends,
Thanks bhpistorqintorpm for the apreciation.
Both those pcb's will work. The kubota regulation version has a suplimentar filtering for the input. I heard from a colleague here that the kubota version sounds cleaner..
You ca try whaterver you want. The best for me was dual mono,from the pictures I posted, with separate psu for the input, BIG radiators and bigger quiescent current.
Cheers
Sergiu

 

[67339]

Hello

Hello,
A real pity that the people who dont have or cannot read all the publications about these French amps like the 8 watt LE MONSTRE, 20/30 watt Hiraga, Nemesis ( one mosfer power amp) and the 50 watt Kaneda and all its variations think they can improve these designs in a few hours. The 50 watt Kaneda was a tricky one. Some did work, some didn't. If you dont need the power i would go for the Hiraga 20 or 30 watts. The big difference between these 2 is the power supply. The 30 watt did have a few extra volts Dc in the supply, a higher current but most of all expensive power caps.
Greetings, Eduard

 

[bhpistorqintorpm]

Hello my friends,
Thanks bhpistorqintorpm for the apreciation.
Both those pcb's will work. The kubota regulation version has a suplimentar filtering for the input. I heard from a colleague here that the kubota version sounds cleaner..
You ca try whaterver you want. The best for me was dual mono,from the pictures I posted, with separate psu for the input, BIG radiators and bigger quiescent current.
Cheers
Sergiu

thanks for the response sergiu2009! i am planning to just the way you said, but might opt for separate cabinets each channel.

 

[damon1983]

Hello,
A real pity that the people who dont have or cannot read all the publications about these French amps like the 8 watt LE MONSTRE, 20/30 watt Hiraga, Nemesis ( one mosfer power amp) and the 50 watt Kaneda and all its variations think they can improve these designs in a few hours. The 50 watt Kaneda was a tricky one. Some did work, some didn't. If you dont need the power i would go for the Hiraga 20 or 30 watts. The big difference between these 2 is the power supply. The 30 watt did have a few extra volts Dc in the supply, a higher current but most of all expensive power caps.
Greetings, Eduard

Le Monstre 8W need to do.

 

[67339]

Hello,
If 8 watt will do just build it. Because of the lower current and voltages compared to the Hiraga amps the French did experiment a lot with car batteries. Most present diy people will try to save money on the power supply. In the eighties there were some serious capacitor manufacturers in France and some were even asked to make special caps like the ones used in the Hiraga le classe A. There was no ebay with cheap copies lol
Greetings, Eduard
p.s where are the choke users?

 

[damon1983]

Hello,
If 8 watt will do just build it. Because of the lower current and voltages compared to the Hiraga amps the French did experiment a lot with car batteries. Most present diy people will try to save money on the power supply. In the eighties there were some serious capacitor manufacturers in France and some were even asked to make special caps like the ones used in the Hiraga le classe A. There was no ebay with cheap copies lol
Greetings, Eduard
p.s where are the choke users?

I did made maket of Le Monster 8W with 0,2F summary filter (CRC, across 0,5 om) a year ago, but I didn't have normal FET's (it had big Idss=approx 2,2 mA). Now I can get good FET's with normal Idss.
p/s: sorry for offtop.

 

[sergiu2009]

Hello,
A real pity that the people who dont have or cannot read all the publications about these French amps like the 8 watt LE MONSTRE, 20/30 watt Hiraga, Nemesis ( one mosfer power amp) and the 50 watt Kaneda and all its variations think they can improve these designs in a few hours. The 50 watt Kaneda was a tricky one. Some did work, some didn't. If you dont need the power i would go for the Hiraga 20 or 30 watts. The big difference between these 2 is the power supply. The 30 watt did have a few extra volts Dc in the supply, a higher current but most of all expensive power caps.
Greetings, Eduard

Hi eduard,
I have readed those articles and did wanted to make at first the kaneda amplifier, but I have read that it was really tricky to adjust especially because of the tight matching between the transistors and their complements and the psu's for the input. Another aspect is that the input transistor are obsolete. These transistors and the topology used where giving the amp a special sound quality that made it so special.
Hiraga said in an article that kaneda and him waited a year for publishing his schematic because the original input transistor that have been chosen was out of production at that time and couldnt make it anymore.

After reading this I resign about the thought of kaneda and stick to the Hiraga.. Everyone can adjust these 20/30w amps if the pcb's are made correctly and transistors matched proprely.

From what I remember the tantal resistor for the amps where chosen from 20 different types of resistor including some Vishay series. I have chosen from six types of resistors from different producers for the input, from carbon, metal film, power metal, metal oxide just the metal fim. Vishay Dale RN Series to be more specific, wich sounded a bit more cleaner than a regular 100 ppm and 1% metal film resistor wich I had in stock.
The problem is that I have just one pcs for one channel and if I wanted one in plus I have ro pay 45 bucs just for the transport from the USA form Farnell so I sticked to the Royal ohm regular one wich I had in stock.
Its pitty that you have to pay 45 dollars just for one resistor...

Cheers
Sergiu

 

[Eldam]

Hello,
If 8 watt will do just build it. Because of the lower current and voltages compared to the Hiraga amps the French did experiment a lot with car batteries. Most present diy people will try to save money on the power supply. In the eighties there were some serious capacitor manufacturers in France and some were even asked to make special caps like the ones used in the Hiraga le classe A. There was no ebay with cheap copies lol
Greetings, Eduard
p.s where are the choke users?

I just can agree with this ! And bis about the chock/coil (CLC) which is at the center of the philosophy : no active regulation; one cap for smoothering one other cap only as power reserve for the load (Hiraga was just interested by the demand of current for the multiple dynamic peaks which are making the music and that the speaker need !). Ok, simple, there is a logic here...

But WTF ? Being french and having access to la Revue de l'Audiophile without missunderstanding, I believe everyone who want to dive in the construction of such amp needs one important information : an amp but for what ?

Hiraga liked the high eficienty speaker. But one thing is missing : what is the limits about the good combo with speakers ?

- impedance curve : limit of the lowest ohms peaks and where ? (in the low? In the highs like with an ESL ?) ; (limit of the sensivity of hte speaker ?) ; (complexity of the passive filter ? Hey can it work with a Kef 104/2 ref ??? despite 92 dB efficienty (92 dB/1W)... so :

-Sensivity of the speakers : limit ? E.G. On the paper a 85 dB 13.5 cm driver needs 200 W to output a peak at 105 dB ! Can a 20 W able to drive 5 drivers with very complex crossover like the one of the Kef 104/2 ref ! (I have more than a big doubt here ! And the same if I putt an Hypex on it ! Less with a Quad 909 or a big Bryston amp with a big reserve of curent - > version 7 and above !)
9
- Which is the best sounding rank between all Hiraga designs ?

- What about the outputt trafo today (the sota ones are lessier and lessier !).

Seing some 20 W or 30 W with so little 15 000 uF caps is a non sense . But again what are the limits of each true Hiraga ? A Le Monstre for A 15" efficienty speaker ??? Really ?

Sorry for the off topic !

 

[thoglette]

I have more than a big doubt here !

Naturellement! I run a JLH 69 and it's fine with my Ariels - but the venerable NAD-3020 will sound better into "difficult" loads which the JLH has not enough current to drive.

Matching the constraints of the amplifier to the vagaries of the speaker is essential! One must match the amplifer to the load.


(Mon Français est merde donc(?) excusez-moi pour mon utilization d'anglais, s'il vous plaît)

 

[Eldam]

, ok good info ! I have a friend with not our disease about audio devices and he has the little NAD in E revision since 1993 with some 94 dB/1W Cabasse : he is happy with it, very casual sound !

Your french is very good enough and perfectly understandable

I wish I have a little JLH with a high efficienty speaker (lurking now at the M2 clone construction thread with genuine JBL parts at Lansing Forum ( the other bar, saloon )or will go with some Beyma stuffs as now I have a 85 dB efficienty speaker only (and a dip at 2 ohms at 125 hZ... hey !) and want to die completly silly !

I had the hope than the body builded caped Hiragas could match with it... but as you say, I think it's a dream without the appropriate speaker !

cheers thoglette,

Eldam

 

[wahab]

Hello,
A real pity that the people who dont have or cannot read all the publications about these French amps like the 8 watt LE MONSTRE, 20/30 watt Hiraga, Nemesis ( one mosfer power amp) and the 50 watt Kaneda and all its variations think they can improve these designs in a few hours. The 50 watt Kaneda was a tricky one. Some did work, some didn't. If you dont need the power i would go for the Hiraga 20 or 30 watts. The big difference between these 2 is the power supply. The 30 watt did have a few extra volts Dc in the supply, a higher current but most of all expensive power caps.
Greetings, Eduard

The Kaneda can be improved, Hiraga s designs not so much because they are intrinsicaly limited as there s not enough OLG to get better results..

As for the articles, and related to what is bolded in your post, well, i made all my scolarship in France from pre elementary school to university, so i have no doubt that i m as qualified as you, if not more, to fully understand what it is about....