If you want to try some modern output transistors, OnSemi's 0302/0281 are a good choice. These can be used up to several Amps idle current. I have used up to 3A idle without any problems (proper heatsinking is mandatory of course).
OnSemi's transistors are widely available at low prices.
As cascode transistors you can find the original ones at some places or use Toshiba's 2SA2240/SA970BL or Sanyo's 2SC2362/SA1016.
The drivers can also be found. These are the hardest to replace, IMO. I guess Sanyo's video transistors would work here but I have not tested them as I am using the original driver transistors.
Lumba - what would you use for drivers?
Sigurd
OnSemi's transistors are widely available at low prices.
As cascode transistors you can find the original ones at some places or use Toshiba's 2SA2240/SA970BL or Sanyo's 2SC2362/SA1016.
The drivers can also be found. These are the hardest to replace, IMO. I guess Sanyo's video transistors would work here but I have not tested them as I am using the original driver transistors.
Lumba - what would you use for drivers?
Sigurd
lineup said:
The 6 Watt level, as being the limit, when using 0.80 A bias, is not surprise to me.
I have setup many such Class A amplfiers. Of very different circuits.
But most of the time such Class A have the same critical level, at where distortion begins to rise:
.. this is when Peak Output Current is between 75 - 85 % of theoretical Max
In this case was 1.2 A out of 1.6 A ( 2x 0.8A ).
Conclusion:
-------------------------------
When designing Class A amplifier, a good bias target value is:
1.33-1.25 x The max output current the amplifier will deliver
Regarding Lumba Ogir comment.
I don't take it too much serious and I am not very bothered. I have seen it looks like Mr. Ogir likes to be ciritical. In fact, I can not remember any post where Lumba actually has been positive to some other peoples works. There is always some detail he find to bang down into the head of other posters
And I am quite sure Jean Hiraga had been much more positive, if see my circuit. At least if he would have had any points, I think he would have been able to explain himself better.
Lumba is right in one thing.
My version is not an original Hiraga Le Monstre.
In this respect, it is probably like most of those 'Le Monstres' built today.
But my version is an attempt to try to improve on Hiraga's good idea. Using some of today's common and affordable transistors.
This makes my circuit a Lineup Class A, a la Hiraga Monstre.
If this is good or not, is much up to the many readers here to decide.
Lineup
Lineup and Sigurd,
The well-known Toshiba 2SA2240/SA970 have of course excellent linearity sound quality.
Lineup`s input stage is OK, BC550/560 will do a good job in common base configuration but the driver and the output device will not.
Essentially, BD139/BD140 are switching transistors, I`ve never seen their Cob value specified in any datasheet, it`s probably pretty high, high Imax, they perform best in current stages.
Using inappropriate or inferior types means not only poor sonics. Regarding the VAS, Cob MUST be low otherwise under dynamic conditions it will strongly increase, decreasing the bandwidth causing distortions, phase errors and instability. Large fast moving voltages require rapid changes in transconductance, you need a cheetah not a sloth here.
I made up this (incomplete) list for a guy when he needed a PNP device (TO-126), most of them have a complementary pair. I`ll try to do something similar for this amp within a couple of days.
2SA1381
2SA1353
2SA1406
2SA0914
2SA1538
2SA1360
2SA1209
2SA1538
2SA1539
2SA1540
2SB1109
2SB1110
Sanyo is a kinda specialist. Video transistors are optimized for low current, Imax 50-100mA, in practice 10-20mA, have been used in CRT monitors so availability can be a problem in the future.
The well-known Toshiba 2SA2240/SA970 have of course excellent linearity sound quality.
Lineup`s input stage is OK, BC550/560 will do a good job in common base configuration but the driver and the output device will not.
Essentially, BD139/BD140 are switching transistors, I`ve never seen their Cob value specified in any datasheet, it`s probably pretty high, high Imax, they perform best in current stages.
Using inappropriate or inferior types means not only poor sonics. Regarding the VAS, Cob MUST be low otherwise under dynamic conditions it will strongly increase, decreasing the bandwidth causing distortions, phase errors and instability. Large fast moving voltages require rapid changes in transconductance, you need a cheetah not a sloth here.
I made up this (incomplete) list for a guy when he needed a PNP device (TO-126), most of them have a complementary pair. I`ll try to do something similar for this amp within a couple of days.
2SA1381
2SA1353
2SA1406
2SA0914
2SA1538
2SA1360
2SA1209
2SA1538
2SA1539
2SA1540
2SB1109
2SB1110
Sanyo is a kinda specialist. Video transistors are optimized for low current, Imax 50-100mA, in practice 10-20mA, have been used in CRT monitors so availability can be a problem in the future.
Dan,
I don`t have any, just get irritated because I know what that VAS arrangement does to the sound. For higher bias current I`d basically increase the value of R6/R7. I agree on the lowered closed-loop gain. Made a small suggestion earlier in this thread, which could be tried out, however meaningfully in real life only.
I don`t have any, just get irritated because I know what that VAS arrangement does to the sound. For higher bias current I`d basically increase the value of R6/R7. I agree on the lowered closed-loop gain. Made a small suggestion earlier in this thread, which could be tried out, however meaningfully in real life only.
Lumba -
that is a nice list you have made!
I use the 2SC3600/SA1406 and 2SC3601/SA1407 complementary pairs. I have never seen any BJTs as linear as these! As you say they cannot handle much current. But for up to 30mA resp 50m they are fine. Still available, and there is a SPICE model for one of the pairs.
I also use the 2SC2911 / SA1209 from your list. These can handle more current compared to the above BJTs. Up to a few 100mA.
There is one Toshiba BJT that I find very good up to 5-10mA and that is the pair 2SC3423 / 2SA1360.
Sigurd
that is a nice list you have made!
I use the 2SC3600/SA1406 and 2SC3601/SA1407 complementary pairs. I have never seen any BJTs as linear as these! As you say they cannot handle much current. But for up to 30mA resp 50m they are fine. Still available, and there is a SPICE model for one of the pairs.
I also use the 2SC2911 / SA1209 from your list. These can handle more current compared to the above BJTs. Up to a few 100mA.
There is one Toshiba BJT that I find very good up to 5-10mA and that is the pair 2SC3423 / 2SA1360.
Sigurd
Lumba Ogir said:
>>Sigurd If you are using the original drivers, running 2,5A through the outputs will bias them at around 25-30mA, well into steep rolloff part of their beta curve.
With the original outputs with the specified grading from the article the current through the drivers should be around 7.5mA. (Wich is right where the beta curves for 2SD756 is at its flattest and 2SB716 resonably so.)
Also the 2sk170/2sj74 are specifed to be of the -Y grade wich is bellow GR in idss and currently unobtainable. Using GR or BL jfets will increase the bias currents and increase the gain. In my le monstre (wich use BL 8mA) I have compensated for both theese effects by lowering the value of the jfets drain resistors.
With the original outputs with the specified grading from the article the current through the drivers should be around 7.5mA. (Wich is right where the beta curves for 2SD756 is at its flattest and 2SB716 resonably so.)
Also the 2sk170/2sj74 are specifed to be of the -Y grade wich is bellow GR in idss and currently unobtainable. Using GR or BL jfets will increase the bias currents and increase the gain. In my le monstre (wich use BL 8mA) I have compensated for both theese effects by lowering the value of the jfets drain resistors.
danieljw said:very nice lineup,
interestingly my le monstre puts out approx 6 watts RMS just like the sim
it is surprising how far 6w goes, however i am building the higher power Le class A (upgraded output stage) at the moment.
-Dan.....
danieljw said:
I have found there are 2 kinds of posters.
Constructive & destructive. Of different degreees.
------------------
Destructive are those that take the spirit away from others contributions.
They can find the bad detail in whatever posted work
Some do not critizise those with 'a name'. like John Curl / Pass or Wurcer circuits.
But can do whatever with a 'newbie' circuit, even if it has a new or interesting idea.
Something valuable to try to explore
But some would even never see or say anything good about about a published Nelson Pass amplifier.
Constructive & Creative.
These often post their own ideas, their own circuits.
They can see both the good and what needs to be corrected in a 'newbie' circuit.
They do say a Nelson Pass is bad.
They can ask him why he did this or that. But they recognize a good idea.
Lumba Ogir is not the first to be like he is.
He obviously have great knowledge. But what does he do with this?
To quote Dan: Can we see your version also please ?
And I promise I will just post what I think is BAD with such a circuit.
Because I am not such a person.
We do not have to be professors in electronic. Like Matti Otala & Marshall Leach.
To do something 'nice' or get a new useful idea.
I have seen newbies come to our forum.
Having more interesting and unique ideas than the best designers around.
Something NEW to try and explore.
youtube: Little Boxes - all the same
Sigge, Sigurd R. is a constructive person of great knowledge.
He can question things, but in a constructive way.
A way that Does Not Take our DIY spirit away.
Thanks, Sigge
And so he contributes to keep our forum a positive brainstorm.
Where new creative babies can see the light of day.
And he has some great audio hifi ideas. And he publish them.
Regards, Lineup
-----------------------------------
PS.
Right now my spirit is gone. I have other more interesting circuits to explore.
But I may post my full version along with Fourier Analys.
And Bandwidth analyse.
If it was not showing good data. Or a harmonics spectrum I could not live with.
Then I would not have posted it.
As noted, my version runs these Sziklai drivers at 60 mA +.
So 2C2240/2SA970 or BC550C maynot be my best choice.
Would they?
DS.
Good observation!
Yes, the drivers are working pretty hard. But not as hard as you say. The NJW0302/0281 have a max beta of 150 but at the high temperatures I am using them at, beta is higher. I have not measured the base currents into the NJW:s but I estimate around 10-15mA.
Simulations with beta = 100 for the output devices, and at a die temp at 100 deg C, I get 12 mA base current. This might not be the ideal current, but me and another person have a total of 8 of these Hiragas up and running and they are sounding wonderful. Especially with the TX2352 resistors and the extensive raw PS with chokes and RIFA PEH200/160 caps together with extensive filtering of the input stages and cascodes, and at other places.
PCB material is Teflon.
You are correct about the Y-grade - it was available for a short time only. I use the GR grade matched for +-2.5% and with an Idss of 4mA. The pot is 200 Ohms. The 1k resistors are still 1k, but the power resistors (Caddock MP930 in T0-220 case) are from 0.5 - 2 Ohms depending on what currents we want to use.
We run the "Hiraga le Monstres" at 20-25 VDC and at a bias current of 2A I use 3 fans to cool the HS.
As with Lineup's "Hiraga le Monstre" the one I have designed is not really a Hiraga le Monstre anymore. The other guy has an original Hiraga le Monstre with a huge battery powered PS, but this new Hiraga sounds even better.
Sigurd
Yes, the drivers are working pretty hard. But not as hard as you say. The NJW0302/0281 have a max beta of 150 but at the high temperatures I am using them at, beta is higher. I have not measured the base currents into the NJW:s but I estimate around 10-15mA.
Simulations with beta = 100 for the output devices, and at a die temp at 100 deg C, I get 12 mA base current. This might not be the ideal current, but me and another person have a total of 8 of these Hiragas up and running and they are sounding wonderful. Especially with the TX2352 resistors and the extensive raw PS with chokes and RIFA PEH200/160 caps together with extensive filtering of the input stages and cascodes, and at other places.
PCB material is Teflon.
You are correct about the Y-grade - it was available for a short time only. I use the GR grade matched for +-2.5% and with an Idss of 4mA. The pot is 200 Ohms. The 1k resistors are still 1k, but the power resistors (Caddock MP930 in T0-220 case) are from 0.5 - 2 Ohms depending on what currents we want to use.
We run the "Hiraga le Monstres" at 20-25 VDC and at a bias current of 2A I use 3 fans to cool the HS.
As with Lineup's "Hiraga le Monstre" the one I have designed is not really a Hiraga le Monstre anymore. The other guy has an original Hiraga le Monstre with a huge battery powered PS, but this new Hiraga sounds even better.
Sigurd
Tazzz said:
Lineup - thanks for the compliment!
Keep building those amplifiers!
Sigurd
Keep building those amplifiers!
Sigurd
lineup said:
I have found there are 2 kinds of posters.
Constructive & destructive. Of different degreees.
------------------
Destructive are those that take the spirit away from others contributions.
They can find the bad detail in whatever posted work
Some do not critizise those with 'a name'. like John Curl / Pass or Wurcer circuits.
But can do whatever with a 'newbie' circuit, even if it has a new or interesting idea.
Something valuable to try to explore
But some would even never see or say anything good about about a published Nelson Pass amplifier.
Constructive & Creative.
These often post their own ideas, their own circuits.
They can see both the good and what needs to be corrected in a 'newbie' circuit.
They do say a Nelson Pass is bad.
They can ask him why he did this or that. But they recognize a good idea.
Lumba Ogir is not the first to be like he is.
He obviously have great knowledge. But what does he do with this?
To quote Dan: Can we see your version also please ?
And I promise I will just post what I think is BAD with such a circuit.
Because I am not such a person.
We do not have to be professors in electronic. Like Matti Otala & Marshall Leach.
To do something 'nice' or get a new useful idea.
I have seen newbies come to our forum.
Having more interesting and unique ideas than the best designers around.
Something NEW to try and explore.
youtube: Little Boxes - all the same
Sigge, Sigurd R. is a constructive person of great knowledge.
He can question things, but in a constructive way.
A way that Does Not Take our DIY spirit away.
Thanks, Sigge
And so he contributes to keep our forum a positive brainstorm.
Where new creative babies can see the light of day.
And he has some great audio hifi ideas. And he publish them.
Regards, Lineup
-----------------------------------
PS.
Right now my spirit is gone. I have other more interesting circuits to explore.
But I may post my full version along with Fourier Analys.
And Bandwidth analyse.
If it was not showing good data. Or a harmonics spectrum I could not live with.
Then I would not have posted it.
As noted, my version runs these Sziklai drivers at 60 mA +.
So 2C2240/2SA970 or BC550C maynot be my best choice.
Would they?
DS.
lineup,
The right way to go is improving the linearity of the stages, not increasing the open-loop gain and GNF.
Hiraga has showed that stability can be maintained without a single dirty part. It`s an extremely clean design.
Would this work?
was I destructively just? Would polite agreeing constructively promote your creative thinking?
The right way to go is improving the linearity of the stages, not increasing the open-loop gain and GNF.
Hiraga has showed that stability can be maintained without a single dirty part. It`s an extremely clean design.
Would this work?
Attachments
Hello All,
If you wanted to see the completed Le monstre build i did, it is here:
http://diyaudioprojects.com/Solid/Jean-Hiraga-Le-Monstre/
Lumba, the use of tl 431 is insteresting.... i cannot comment on how it would work in this circuit perhaps give it a try ?
Lineup, it would be a shame to see your diy spirit disappear you have contributed alot!!!
Audiomanics, that layout is great !!!!!
If you are interested in the other hiraga amp find it here.
it is completed and i have listened to it.
http://www.diyaudio.com/forums/show...193&perpage=25&highlight=hiraga&pagenumber=13
-Dan
If you wanted to see the completed Le monstre build i did, it is here:
http://diyaudioprojects.com/Solid/Jean-Hiraga-Le-Monstre/
Lumba, the use of tl 431 is insteresting.... i cannot comment on how it would work in this circuit perhaps give it a try ?
Lineup, it would be a shame to see your diy spirit disappear you have contributed alot!!!
Audiomanics, that layout is great !!!!!
If you are interested in the other hiraga amp find it here.
it is completed and i have listened to it.
http://www.diyaudio.com/forums/show...193&perpage=25&highlight=hiraga&pagenumber=13
-Dan
For the Fet's (2SK170 and 2SJ74):
These are grouped for their Idss: Without: 2,6-20mA
GR = 2,6 - 6,5 mA BL=6,0-12mA and V=10-20mA
Y is probably the same as V
For the normal transistors (2SC1775 and 2SA782):
These are grouped for their hfe: no ext.: 400-1200
E = 400-800 and F= 600-1200
As you have to match the transistors you have to measure the parameters of each transistor individual anyway..
For the fets you can change the resistance of the trimmer to get the correct bias and the current through the endstage.
It has been described somewhere in this tread before..
Kees
These are grouped for their Idss: Without: 2,6-20mA
GR = 2,6 - 6,5 mA BL=6,0-12mA and V=10-20mA
Y is probably the same as V
For the normal transistors (2SC1775 and 2SA782):
These are grouped for their hfe: no ext.: 400-1200
E = 400-800 and F= 600-1200
As you have to match the transistors you have to measure the parameters of each transistor individual anyway..
For the fets you can change the resistance of the trimmer to get the correct bias and the current through the endstage.
It has been described somewhere in this tread before..
Kees