Hi ! symasym5 has a low distortion because
-beta of mpsa18 is so high ,>900
-no use degeneration reisistor
If input trsansistors is replaced by 2n5551 ,or a low beta transistor such as c1815 , symasym5 will be defeated by Blameless .
Iq for VAs is only 1.5mA .This is a normal thing . Because VAs use current mirror load . Output impedance of current mirror is high extremely ->gain of VAS is also high teriblely ---> 1.5mA is high enough to fed output satge
-beta of mpsa18 is so high ,>900
-no use degeneration reisistor
If input trsansistors is replaced by 2n5551 ,or a low beta transistor such as c1815 , symasym5 will be defeated by Blameless .
Iq for VAs is only 1.5mA .This is a normal thing . Because VAs use current mirror load . Output impedance of current mirror is high extremely ->gain of VAS is also high teriblely ---> 1.5mA is high enough to fed output satge
Flamebless!!!! oh!...No!!!...No!!!...i am starting to be sick again!
Where's my pils!
Call the Doctor!.....no!.... this doctor will kill me more fast!
Air!!..... water!...oxigen to me!!!
Blood pressure increasing now..... 33 by 29!!!
Flamebless is too much for me..... good by evil world!
ahahahahaha!
regards,
Carlos
Where's my pils!
Call the Doctor!.....no!.... this doctor will kill me more fast!
Air!!..... water!...oxigen to me!!!
Blood pressure increasing now..... 33 by 29!!!
Flamebless is too much for me..... good by evil world!
ahahahahaha!
regards,
Carlos
Attachments
Oh my dear and beloved old friend Thanh!
That was very strong to me.... really excessive impact to my 54 years off existence.
The day the Gamestress win Symassym, i will be in paradise!
Hô Hô Hô..... wanna know?...... really wanna know?.... i will tell ya!!!.... your avatar is not pretty anymore, will not give you more Ice cream and will not borrow you my Bike anymore!
regards,
Carlos
That was very strong to me.... really excessive impact to my 54 years off existence.
The day the Gamestress win Symassym, i will be in paradise!
Hô Hô Hô..... wanna know?...... really wanna know?.... i will tell ya!!!.... your avatar is not pretty anymore, will not give you more Ice cream and will not borrow you my Bike anymore!
regards,
Carlos
Attachments
Regarding the 10 Ohm resistor
Michael,
this 10 Ohm resistor really helps to reduce hum/buzz in some instruments arrangement, as it inserts impedance into groundloop and allows input pair to work in CMV rejecting manner.
The background of my concern is this: the same aid was used in another amplifier, produced in hundreds of pieces. The 10 Ohm resistor worked and reduced groundloop hum/buzz problems. In case that 0V power GND was connected to mains protective earth, and this was the case, and second instrument like preamp, CD ar another one somewhere in the arrangement had also ANA GND connected to PE, and mains cord in several meters distant wall outlet, even a potential difference between the 2 PE pins of the wall outlet plugs - or a simple mistake - created current high enough to burn the 10 Ohm resistor. This had really happened in several cases, resulting in 10 Ohm resistor burned and unfortunatelly together with output protection relay not working well, the disater was here - burned speakers. That's why I am always a bit concerned when I see this solution. 1N4148 is not enough, I would guess 1N5408 at least.
I have already added the resistor into Eagle layout, just for my personal evaluation or for DIY purposes, but would be careful if it was industrial product to be sold. You might consider output DC protection?
Regards,
Pavel
Michael,
this 10 Ohm resistor really helps to reduce hum/buzz in some instruments arrangement, as it inserts impedance into groundloop and allows input pair to work in CMV rejecting manner.
The background of my concern is this: the same aid was used in another amplifier, produced in hundreds of pieces. The 10 Ohm resistor worked and reduced groundloop hum/buzz problems. In case that 0V power GND was connected to mains protective earth, and this was the case, and second instrument like preamp, CD ar another one somewhere in the arrangement had also ANA GND connected to PE, and mains cord in several meters distant wall outlet, even a potential difference between the 2 PE pins of the wall outlet plugs - or a simple mistake - created current high enough to burn the 10 Ohm resistor. This had really happened in several cases, resulting in 10 Ohm resistor burned and unfortunatelly together with output protection relay not working well, the disater was here - burned speakers. That's why I am always a bit concerned when I see this solution. 1N4148 is not enough, I would guess 1N5408 at least.
I have already added the resistor into Eagle layout, just for my personal evaluation or for DIY purposes, but would be careful if it was industrial product to be sold. You might consider output DC protection?
Regards,
Pavel
Michael,
yes, I have already sent the files to production, the PCBs will be ready in 2 weeks. It is not reasonable to push on earlier delivery, as price accelerates then..
Regarding upper layer, I have increased drill diameters to 1.12mm, and pad diameter to 2.1mm. This will assure possibility of wire jumper soldering, if needed. These modifications are not made in the files I posted here.
Regarding limitation - I mean clipping. I am getting nasty HF oscillations in the internal structure of the amp, but still only simulated, the real circuit check will come in some 3 weeks. A part of those oscillations can be seen at the output after clipping.
I am very interested in results on a real circuit. I have decreased closed loop gain to 26 dB, as 40 dB is too much for me. And I am getting into troubles with feedback capacitor.
Yes, I was speaking about input differential pair. Q1 and Q2 are names according to your eagle schematic file.
Regards,
Pavel
P.S. Oh, there's a Trabi fan club in your city
yes, I have already sent the files to production, the PCBs will be ready in 2 weeks. It is not reasonable to push on earlier delivery, as price accelerates then..
Regarding upper layer, I have increased drill diameters to 1.12mm, and pad diameter to 2.1mm. This will assure possibility of wire jumper soldering, if needed. These modifications are not made in the files I posted here.
Regarding limitation - I mean clipping. I am getting nasty HF oscillations in the internal structure of the amp, but still only simulated, the real circuit check will come in some 3 weeks. A part of those oscillations can be seen at the output after clipping.
I am very interested in results on a real circuit. I have decreased closed loop gain to 26 dB, as 40 dB is too much for me. And I am getting into troubles with feedback capacitor.
Yes, I was speaking about input differential pair. Q1 and Q2 are names according to your eagle schematic file.
Regards,
Pavel
P.S. Oh, there's a Trabi fan club in your city
Ah, that's a problem ! If you reduce closedloopgain you increase nfb
the same amount, you should have added degeneration to compensate
and reduce openloopgain to match again the amount of nfb.
It's possible that this change made the amp unstable.
The combination of feedbacknetwork, feedbackcap, amount of nfb
is very delicate and sets the characteristic of the amp.
I use that much closedloopgain as i work without preamp and connect
the symasym directly to cd-player.
The feedbacknetwork is the "heart" of the amp, changing it is like a heart-operation...
I am quite sure that symasym did not show internal ringing, but i will
recheck this evening. The files you uploaded, do they contain the changes
for lower gain ? Or can you just tell me ?
Symasym is very sensitive to such changes because of its high GBW,
at 10khz is still has openloopgain of ~30.000...
I always observe the currents in input-LTP, as these are the key to a really stable amp.
If these show HF-artefacts, the amp is unstable and likely to sound bad, like broken glass.
Looking at output only tells not much...
A Trabbi-club in Gütersloh ? I must have missed that...
Mike
the same amount, you should have added degeneration to compensate
and reduce openloopgain to match again the amount of nfb.
It's possible that this change made the amp unstable.
The combination of feedbacknetwork, feedbackcap, amount of nfb
is very delicate and sets the characteristic of the amp.
I use that much closedloopgain as i work without preamp and connect
the symasym directly to cd-player.
The feedbacknetwork is the "heart" of the amp, changing it is like a heart-operation...
I am quite sure that symasym did not show internal ringing, but i will
recheck this evening. The files you uploaded, do they contain the changes
for lower gain ? Or can you just tell me ?
Symasym is very sensitive to such changes because of its high GBW,
at 10khz is still has openloopgain of ~30.000...
I always observe the currents in input-LTP, as these are the key to a really stable amp.
If these show HF-artefacts, the amp is unstable and likely to sound bad, like broken glass.
Looking at output only tells not much...
A Trabbi-club in Gütersloh ? I must have missed that...
Mike
The files I posted are same as yours, only the 10 Ohm resistor was added and board size behind power transistors reduced, for the reason of mechanical construction (you will see after you open them). The closed loop gain was decreased only by value of feedback resistor, so it makes no difference for using the files with original gain.
Hi thanhthanh said:
I am planning to start a prototype of Symasym these days and I am collecting parts. But unfortunately I couldn't find MPSA18 and I will make it with BC546B instead (I haven't measured their hfe). Do you think that will kill the amplifier, making it uninteresting? I also have 2n5551/5401 but perhaps I could order MPSA18 from abroad (or search some more for them!).
Have a nice day
Dimitris.
Hi Dimitris !
The bc546b is not really bad, it typically has hfe of 300.
Maybe you find some bc550c ? or 2sc2240 ? I would not recommend
the 2n5551, it is not bad, but has only hfe of 150 and can create
turn on/off thump.
Yes, mpsa18 is not too easy to get, normal shops don't have them...
PMA, yes, the original value is ~1:45 ~33db, this gives saturation at ~0.7v input.
This seemed the perfect value for use with normal linelevel.
Mike
The bc546b is not really bad, it typically has hfe of 300.
Maybe you find some bc550c ? or 2sc2240 ? I would not recommend
the 2n5551, it is not bad, but has only hfe of 150 and can create
turn on/off thump.
Yes, mpsa18 is not too easy to get, normal shops don't have them...
PMA, yes, the original value is ~1:45 ~33db, this gives saturation at ~0.7v input.
This seemed the perfect value for use with normal linelevel.
Mike
Hi edl,
because of the much higher inputoffsets with low hfe ? Until the feedback dc-blocking
cap is charged, these inputoffsets are not identical on both sides
of the input-diffamp. So, the lower these offsets are, the less DC-offset
is generated when turning the amp on.
In the case of symasym it drops to a level that is not recognizable.
But, use of high-beta devices is not the cure for turnon-thump, it does only help...
Mike
because of the much higher inputoffsets with low hfe ? Until the feedback dc-blocking
cap is charged, these inputoffsets are not identical on both sides
of the input-diffamp. So, the lower these offsets are, the less DC-offset
is generated when turning the amp on.
In the case of symasym it drops to a level that is not recognizable.
But, use of high-beta devices is not the cure for turnon-thump, it does only help...
Mike
Hi Pavel,
I am interested in the boards. I would need artwork to make my own by hand. I do have MPSA18 to try also. I will match hFE on each set. The difference in gain in these parts will only cause an incremental DC offset change. The "thump" on turn on depends on how fast the front end settles down to servo the offset close to zero.
-Chris
I am interested in the boards. I would need artwork to make my own by hand. I do have MPSA18 to try also. I will match hFE on each set. The difference in gain in these parts will only cause an incremental DC offset change. The "thump" on turn on depends on how fast the front end settles down to servo the offset close to zero.
-Chris