Hello
Catalin I hope you do not get upset because.....
You know so many guy wrote "I'll build or "I'll test that amp", please send these or that etc and until now nothing..
To me over the 22 years or so the best sounding amp I built.
I stated many time it was my fault the thermal runaway. We can not set up a new amplifier bias high and fall sleep.. Also many darlington device from ISC BDW83/84 (high power garbage)!!! I powered up the amp and after 20-30 min. the darlington gone at 50mA low bias.
The lower power BDW93/94 from SC micro worked perfect even on 100mA bias for weeks!!!
If you don't find BDW83C/84C from SC micro (sometimes marked Morocco) please buy TIP142/147
I didn't meet a guy ever who heard the amp didn't like the sound.
One think keep in mind these is not a PURE Class A amplifier. Those darlingtons can't take the heat specially not at 40V rail.
padamiecki can you lead me to the thread where you took the circuit.
It would be great if I do not need to go over a few hundred page.
Wahab if you have at your hand the Japanese version of these amp (original) I would love to see it.
I'm interested to what type of transistor was used for VAS etc.
Greetings Gabor
Catalin I hope you do not get upset because.....
You know so many guy wrote "I'll build or "I'll test that amp", please send these or that etc and until now nothing..
To me over the 22 years or so the best sounding amp I built.
I stated many time it was my fault the thermal runaway. We can not set up a new amplifier bias high and fall sleep.. Also many darlington device from ISC BDW83/84 (high power garbage)!!! I powered up the amp and after 20-30 min. the darlington gone at 50mA low bias.
The lower power BDW93/94 from SC micro worked perfect even on 100mA bias for weeks!!!
If you don't find BDW83C/84C from SC micro (sometimes marked Morocco) please buy TIP142/147
I didn't meet a guy ever who heard the amp didn't like the sound.
One think keep in mind these is not a PURE Class A amplifier. Those darlingtons can't take the heat specially not at 40V rail.
padamiecki can you lead me to the thread where you took the circuit.
It would be great if I do not need to go over a few hundred page.
Wahab if you have at your hand the Japanese version of these amp (original) I would love to see it.
I'm interested to what type of transistor was used for VAS etc.
Greetings Gabor
was it here? http://www.diyaudio.com/forums/soli...diy-amplifier-20-years-go-20.html#post3264029 that you asked?
gaborbela;3274600 [B said:
It can be find on post 14...
http://www.diyaudio.com/forums/soli...-diy-amplifier-20-years-go-2.html#post1659097
Do you think that your transistor models accurately simulate BJT tempco ???
I havent ever had trouble with this circuit, build it first time about 22 years back and with some input from JLH himself, the original designer.
In this case one can get away to some extent with the vas bias stability as the darlingtons are biased high and no regard is taken to THD, its even better with the circuit in class A operation but what happens when you want accuracy.
Why souldnt it be the case ?....
I dont understand why you think that such a basic parameter
is not accounted.
In sims i checked the variation from 20°C to 80°C , and the numbers
correlate of course accurately with the theorical behaviour.
Member
Joined 2009
Paid Member
the solution is here: http://www.diyaudio.com/forums/soli...t-simple-symmetric-amplifier.html#post2722516
post 22.
Hello Guys
Let s stick with the given circuit to comment that!
I do not say not to write about similar amplifiers only just we do not lose the focus.
Here I had these circuit which I think wahab made it "more" stable..
How I understood all the similar amplifier design suffer from the same problem.
By the amp warms up the bias rise up to, (that is true about the Mosnster Hiraga to, I built it. Mostly the bias after total warm up almost was double like at the start up)
Now if these not compensated with a large enough heatsink it will lead to thermal runaway. That is 100% sure. Doesn't matter if the amp use power Darlington, normal BJT, or mosfet.
I read someone had similar issue with power mosfet!!!
Again that mean these amplifiers req. the adequate heatsink size.
What I ask here if we can slow done these bias movement a bit.
wahab came up with a solution which will be tested soon, thank God and thank to wahab for the help.
Also if I understood well if we take the feedback only from the input pair these drifting will be reduced by a lot. Now I'm not sure until I don't test how that effect the sound is anyway at all. We will see soon.
Here is nothing to do with anything else. We (I) do not want to upgrade the sound of the amp!
The amp sound better than someone can expect from!
I'm happy to see we have conversation about similar amplifiers. That is great , I want to learn.
But from a outsider (just came here now) it will look like we have a terrible unstable problem here.
I want to emphasise that is not true in these case!
If my M Hiraga did not had large enough heatsink the thermal runaway it would happened with that to. That is 100%
So I'm not sure all these SSA tend to act these way, but one think I'm sure these amplifier need large enough heatsink to stop that thermal and bias rise.
Also I mentioned these amplifier with 3 pair darlington from 18VAC transformer it was biased 3A. The amp was stable because I used large enough heatsink!
Another experience probably the darlington tend to produce more heat because that (small)cheap contain 2 BJT, diode, 2 resister com pare to the large TO 247 case simple BJT.
Only I wrote all these if someone read the topic not to have that feeling he must run from here because these amp (or what ever we talk) acting so weird better not to touch it.
Other wise please keep going on these topic, that is how we learn!
Today I cut done the clad-board for etching these PC boards...
Greetings Gabor
Let s stick with the given circuit to comment that!
I do not say not to write about similar amplifiers only just we do not lose the focus.
Here I had these circuit which I think wahab made it "more" stable..
How I understood all the similar amplifier design suffer from the same problem.
By the amp warms up the bias rise up to, (that is true about the Mosnster Hiraga to, I built it. Mostly the bias after total warm up almost was double like at the start up)
Now if these not compensated with a large enough heatsink it will lead to thermal runaway. That is 100% sure. Doesn't matter if the amp use power Darlington, normal BJT, or mosfet.
I read someone had similar issue with power mosfet!!!
Again that mean these amplifiers req. the adequate heatsink size.
What I ask here if we can slow done these bias movement a bit.
wahab came up with a solution which will be tested soon, thank God and thank to wahab for the help.
Also if I understood well if we take the feedback only from the input pair these drifting will be reduced by a lot. Now I'm not sure until I don't test how that effect the sound is anyway at all. We will see soon.
Here is nothing to do with anything else. We (I) do not want to upgrade the sound of the amp!
The amp sound better than someone can expect from!
I'm happy to see we have conversation about similar amplifiers. That is great , I want to learn.
But from a outsider (just came here now) it will look like we have a terrible unstable problem here.
I want to emphasise that is not true in these case!
If my M Hiraga did not had large enough heatsink the thermal runaway it would happened with that to. That is 100%
So I'm not sure all these SSA tend to act these way, but one think I'm sure these amplifier need large enough heatsink to stop that thermal and bias rise.
Also I mentioned these amplifier with 3 pair darlington from 18VAC transformer it was biased 3A. The amp was stable because I used large enough heatsink!
Another experience probably the darlington tend to produce more heat because that (small)cheap contain 2 BJT, diode, 2 resister com pare to the large TO 247 case simple BJT.
Only I wrote all these if someone read the topic not to have that feeling he must run from here because these amp (or what ever we talk) acting so weird better not to touch it.
Other wise please keep going on these topic, that is how we learn!
Today I cut done the clad-board for etching these PC boards...
Greetings Gabor
Last edited:
Hello
Did these amp was tested in real life?
I know sim a amp it help a lot, sometimes I wish I could has access to simulate to.
But the real test in real life how it act or more important sound.
Sim not always give 100% accuracy specially about the amplifier sound.
Sometimes our ear hear something different like the sim result...
I had a friend in Budapest, he build speakers, amps, TT etc for sale.
Taylor Acoustic- not the same Taylor from N America...
He always told me the best toll to really test a new product the ear!
Greetings Gabor
May i ask , the solution to what ?..
Seems that this topology s drawbacks are still not very well understood
by many people , not talking of you but of absolute beginners , think that
i ll soon make a little tutorial in this very thread to explain basicaly the
voltages and currents dependencies between the nodes , this will allow
non experimented people to yield an acceptably stable design.
Member
Joined 2009
Paid Member
It is the solution to an unstable Vas current influencing the output device bias point. The common Vbe multiplier does not hold a very constant voltage differential under fluctuating current conditions, but the Hagerman topology offers a significant improvement - good enough perhaps. This is what I used and it measured good.
Thanks for the precision Bigun.
The VAS current variation itself can be reduced by increasing the voltage
through its emitters resistors but this will yield lower output voltage swing
as well as lower OLG since the VAS emitters resistors must also be increased.
The other source of VAS current variation is the input stage varying current
if ever the emitters DC loading resistors are connected to unregulated supply
as in Gabor s original schematic.
On the DC output offset variation , the input stage is the main culprit
and to make things worse thermal coupling of the input transistors
is of no real help , moreover as the VAS and output stage have
an influence as the feedback path has a significant DC current flow.
The VAS current variation itself can be reduced by increasing the voltage
through its emitters resistors but this will yield lower output voltage swing
as well as lower OLG since the VAS emitters resistors must also be increased.
The other source of VAS current variation is the input stage varying current
if ever the emitters DC loading resistors are connected to unregulated supply
as in Gabor s original schematic.
On the DC output offset variation , the input stage is the main culprit
and to make things worse thermal coupling of the input transistors
is of no real help , moreover as the VAS and output stage have
an influence as the feedback path has a significant DC current flow.
Hello
Wahab at these amplifier can I use those Hitachi driver transistors or better to use the Toshiba here to.
I would like to test BD829/830 transistors for VAS, I read those are great drivers all do hard to find them.
Today I received 10 pair TIP142/147, I have 10 pair BDW83/84C from ST also some MJ11015/15 & 2N6284/87
The last two pair are in metal case but I will test those to. I need those metal case transistor for Pass A40 amp.
I borrowed from that old circuit some idea, I'm not sure will help to improve the sound.
I added to you circuit wahab.
These circuit based on a Hungarian popular amplifier called HEED amplifier.. Very similar to my orig circuit accept the feedback was taken from the front and the 33p cap was added to the VAS.
If you think it make any sense to use those 100K resisters with the 33p capacitors please let me know. I will test that way to.
I have some clad board left over to make one more set PC boards.
For these amplifier I need the resisters also, I made the PC boards which include your mode.
Greetings Gabor
Wahab at these amplifier can I use those Hitachi driver transistors or better to use the Toshiba here to.
I would like to test BD829/830 transistors for VAS, I read those are great drivers all do hard to find them.
Today I received 10 pair TIP142/147, I have 10 pair BDW83/84C from ST also some MJ11015/15 & 2N6284/87
The last two pair are in metal case but I will test those to. I need those metal case transistor for Pass A40 amp.
I borrowed from that old circuit some idea, I'm not sure will help to improve the sound.
I added to you circuit wahab.
These circuit based on a Hungarian popular amplifier called HEED amplifier.. Very similar to my orig circuit accept the feedback was taken from the front and the 33p cap was added to the VAS.
If you think it make any sense to use those 100K resisters with the 33p capacitors please let me know. I will test that way to.
I have some clad board left over to make one more set PC boards.
For these amplifier I need the resisters also, I made the PC boards which include your mode.
Greetings Gabor
Attachments
Hi Gabor ,
Both pairs will work of course , with the BD829/830 being in principle
better due to their higher gain at low current but both have way
more input capacitance than the 2SA1360/SC3423 wich are undoubtly
superior to both these drivers.
As said in previous posts the BDW are the best choice in this list ,
i wouldnt recommend the TIP for anything other than as capacitance
multiplier where they would suit perfectly.
Theses 100K resistors should really be eliminated for good , they only
help creating thermal run down as well as more distorsion , higher output impedance and so on , they degrade the amps perfs by any possible metric.
It is likely that this schematic has gone "borrowed" from a designer to another
with this original "feature" being unquestionned or even thought as possibly
improving the sound , but for sure not only it is the opposite but it has been
undoubtly a source of never ending hassle for non experimented builders,
i just imagine all the overheated/exploded devices due to this misdesign.....
The 33pF compensation capacitors are better left connected to the input stage emitters , as in your original schematic as well as in the modded version , since the amp is more stable with better phase response this way.
Edit : A possible mod is to add a 10nf capacitor from collector to emitter of the VBE multiplier ,
this wont change the whole picture , though , it s almost an aesthetical mod.....
Last edited:
Hello
Thank you for the information.
OK I'll stick 100% to the circuit was modded by you. Of course I'll compare that with the old circuit.
Those TIP transistors were ordered before you wrote better to use the BDW.They just arrived today after several weeks.
I will buy the Toshiba devices for VAS, I just don't want to get them from ebay or other unsecured source. Even in the local store here in Toronto we can get lot of counterfeit semis.
I do not mind the investment in the TIP ($20 for the 10 pair) including shipping.
It didn't sounded bad at last test , on that time I had only the cheap quality BDW from ISC the TIP outperform that easily. The ISC is garbage it blew up 50 mA bias after 25-30 min twice.
I see people work with these type topology a while but they use Lateral mosfet.
I would be happy to see these topology using BJT power transistors but not not something over complicated. Or someone would implement the Sanken (the thermal track darlington type)
Greetings Gabor
Thank you for the information.
OK I'll stick 100% to the circuit was modded by you. Of course I'll compare that with the old circuit.
Those TIP transistors were ordered before you wrote better to use the BDW.They just arrived today after several weeks.
I will buy the Toshiba devices for VAS, I just don't want to get them from ebay or other unsecured source. Even in the local store here in Toronto we can get lot of counterfeit semis.
I do not mind the investment in the TIP ($20 for the 10 pair) including shipping.
It didn't sounded bad at last test , on that time I had only the cheap quality BDW from ISC the TIP outperform that easily. The ISC is garbage it blew up 50 mA bias after 25-30 min twice.
I see people work with these type topology a while but they use Lateral mosfet.
I would be happy to see these topology using BJT power transistors but not not something over complicated. Or someone would implement the Sanken (the thermal track darlington type)
Greetings Gabor
In TSSA thread you can find the simplest TSSA using one pair of input stage and one pair of output stage, so only 4 transistors. The darlington output is more suitable.
Bigun's TGM5 is not too complicated. The input stage is very good. The output topology is similar to the one used by Roender's Fc100.
Well , laterals are quite forgiving and have the big advantage of a negative temperature coefficient , that is , the hotter , the less the iddle current , i also like those components but , contrary to an unexpectedly high number of members , i m a big proponent of darlingtons as well , i just dont understand why they are vilified , surely because there s a lot of misconceptions about them , the most recurrent being a very bad understanding of their beta vs current curves....
A straight forward replacement of the darlingtons by lateral fets
would work , that makes no doubt , but then the circuit has been
optimized for BJT output devices in respect of the input stage
to yield an acceptable DC output offset variation.
You can try directly this version using Lfets but if ever the output
DC offset variations are more than 20mV for 50°C temperature
variation of the WHOLE circuit , a few components values are to be
changed , see below , the feedback resistors should be reduced and
the compensation capacitors increased to 39 to 47pF to cope with the
resistors lower value influence in stability.
Attachments
Last edited:
Hello Jay
Above mentioned amplifiers both has the same weakness.
How I read from Rudy and Bigun the bass of that topology not it strength.
Otherwise the sound can be beautiful - I never heard neither of them.
But I asked Rudy's opinion over the FC100 and he wrote com pare to the Symasym it is a great amp. I hate the Symasym,one of the worst amp I ever built!
Last time I read his test he only gave an 7-8 out of 10 to the bass. Clearly he wrote the weakest point of the amp the low frequency.
Now these Darlington has incredible bass, very deep, well controlled, the best I ever heard out of my DIY amps.
The reason I wrote I would like these topology with the new Sanken (thermal track) darlingtons or some good power BJT.
Again bass not everything to me (do not misunderstand)
I have a lot of project half ready or ready to test them and 75-80% of those mosfet amps.
Top of that when I'll gave up my DIY my last build will be SE tube amp (s).
A GM70 sure but before I go to that high voltage I have to get some experience (more than 10 years I didn't built nothing with tubes) may be I'll put together a 300B..
Bot will be SE, and SE strength not the bass also.
I hope you understand me.
Greetings Gabor
Well , the Darlington version values would work as well , this variant
is only if it doesnt work good enough DC offset variation/temp wise
when switching to laterals.
Of course source resistors can be added usefully if the devices
are not matched but anyway two pairs are needed to provide
enough output current in safe conditions.
That said i think that laterals would find a better use in the FFet
Hitachi amp project as higher feedback allow to compensate for
their lower conductance.
I m surprised that some people hear difference in basses from an
amp to another one , wich is possible if an amp has really poor
design and power supply or dried supply caps, but with reasonably
designed amps there should be no difference.
For the record , i and a friend once compared his 2 x 200W Bjt commercial amp
that he s using as electric bass amplifier with my venerable 30 years old 2 x 120W
lateral fet amp and the laterals litteraly destroyed the Bjt at the first notes of his bass,
so much that i thought that the PSU caps where not good , while in fact their value
was half of the lateral amp ones , wich could have been the main issue , who knows...
is only if it doesnt work good enough DC offset variation/temp wise
when switching to laterals.
Of course source resistors can be added usefully if the devices
are not matched but anyway two pairs are needed to provide
enough output current in safe conditions.
That said i think that laterals would find a better use in the FFet
Hitachi amp project as higher feedback allow to compensate for
their lower conductance.
I m surprised that some people hear difference in basses from an
amp to another one , wich is possible if an amp has really poor
design and power supply or dried supply caps, but with reasonably
designed amps there should be no difference.
For the record , i and a friend once compared his 2 x 200W Bjt commercial amp
that he s using as electric bass amplifier with my venerable 30 years old 2 x 120W
lateral fet amp and the laterals litteraly destroyed the Bjt at the first notes of his bass,
so much that i thought that the PSU caps where not good , while in fact their value
was half of the lateral amp ones , wich could have been the main issue , who knows...