Hi Pavel,
I read Hugh's (AKSA) comments to read "the amplifier is technically sound". In other words, it is a good solid design. Sheldon picked up on this also. I do not believe he meant the amplifier was "sterile" or technical sounding.
Hugh (AKSA) did say that that he believes the amp can be made to sound better still. I will agree with this since two pcb modifications have been mentioned already. One, thicker supply traces. Two, ground plane on the top side.
Anyway, I am rather excited to try and get a working pair made up. I think it shows promise as a good amplifier. There is no one doubting the excellent work that has already gone into this amplifier design. Both you and Michael are to be applauded. I am ...
-Chris
I read Hugh's (AKSA) comments to read "the amplifier is technically sound". In other words, it is a good solid design. Sheldon picked up on this also. I do not believe he meant the amplifier was "sterile" or technical sounding.
Hugh (AKSA) did say that that he believes the amp can be made to sound better still. I will agree with this since two pcb modifications have been mentioned already. One, thicker supply traces. Two, ground plane on the top side.
Anyway, I am rather excited to try and get a working pair made up. I think it shows promise as a good amplifier. There is no one doubting the excellent work that has already gone into this amplifier design. Both you and Michael are to be applauded. I am ...
-Chris
mikeks, where do you take all your knowledge ? 
My bodeplot tells me a phasemargin >90°, the problem was that CL-gain kept only ~4db above OL-gain... You can't make the bodeplot openloop. BTW, Pavel lives just 500km away...
12 cents, did you at least see some traces when putting the boards into the NaOH ? They should immediately get visible when putting the board into the NaOH. And after 5 minutes or more even the photoresist not having got light should peel of. If not, your NaOH needs more concentration.
About your lamp, sorry, i had troubles not to laugh ! For comparison, i use a several 100watts mercurylamp for 2minutes, and that is barely enough. Consider several hours with that lamp... Or get a daylight lamp, much stronger.
Mike
My bodeplot tells me a phasemargin >90°, the problem was that CL-gain kept only ~4db above OL-gain... You can't make the bodeplot openloop. BTW, Pavel lives just 500km away...
12 cents, did you at least see some traces when putting the boards into the NaOH ? They should immediately get visible when putting the board into the NaOH. And after 5 minutes or more even the photoresist not having got light should peel of. If not, your NaOH needs more concentration.
About your lamp, sorry, i had troubles not to laugh ! For comparison, i use a several 100watts mercurylamp for 2minutes, and that is barely enough. Consider several hours with that lamp... Or get a daylight lamp, much stronger.
Mike
Thank you MikeB, I now have tried one hour with a 100W light bulb and am walking strongly in the right direction 
. Due to the paper structure I have not a very clean free areas now, I think it needs still more light. Where do they have mercury lamps? I suppose the free areas should get violet in the NaOH (with most board materials) quick and uniformly?
. Due to the paper structure I have not a very clean free areas now, I think it needs still more light. Where do they have mercury lamps? I suppose the free areas should get violet in the NaOH (with most board materials) quick and uniformly?
Yes, 12 cents (Do you have a name ?), these parts get some kind of violet within 1-2 seks. If it did not completely peel of, you might have to make the NaOH slightly stronger.
The kind of mercury lamp i use is not selled nowadays, way to strong !
It's about 35 years old... And i have to leave room when it is turned on, because the visible part of the light is too much and it stinks because of the O3... But there are other UV-lamps.
Mike
The kind of mercury lamp i use is not selled nowadays, way to strong !
It's about 35 years old... And i have to leave room when it is turned on, because the visible part of the light is too much and it stinks because of the O3... But there are other UV-lamps.
Mike
Sheldon, Samuel, and Anatech had 100 percent understanding
Some others loose just a little of comprehension.
But it is all rigth to me.
Hugh seems not interested too....he did not enter to comment as he has some packages to send for foreign countries.
Also,He is not answering my mails with the normal "instantaneous way" those days...may be busy.
To me it is all rigth, 6 or half dozen means the same to me...and those guys, and theirs amplifiers are all good, and much more than enougth for my exigences, my needs and my tastes.
Enjoy them and be happy.
regards,
Carlos
Some others loose just a little of comprehension.
But it is all rigth to me.
Hugh seems not interested too....he did not enter to comment as he has some packages to send for foreign countries.
Also,He is not answering my mails with the normal "instantaneous way" those days...may be busy.
To me it is all rigth, 6 or half dozen means the same to me...and those guys, and theirs amplifiers are all good, and much more than enougth for my exigences, my needs and my tastes.
Enjoy them and be happy.
regards,
Carlos
MikeB said:
Already ~1.5%. Do you reuse the NaOH?
So, which devices should be matched against which others? The output pairs, PNP vs NPN? Or left channel NPN vs right channel NPN? With hFE or something other?
Lukas
MikeB said:mikeks, where do you take all your knowledge ?
My bodeplot tells me a phasemargin >90°, the problem was that CL-gain kept only ~4db above OL-gain... You can't make the bodeplot openloop. BTW, Pavel lives just 500km away...
Mike
Perhaps we need to be clear about which Bode plot we are refering to here....
You appear indicate plots of the closed cloop response....
...from which you've obtained a margin of 90°............
On the other hand, i am refering to the loop-transmission phase margin....
.....which is one of two quantities determining a linear feedback system's stability....
The other is gain-margin...........
Neither of these quantities is directly related to the foward-path's response at unity gain, unless of course, the amp. is used with 100% feedback...
I´m very happy to found and read this thread. Especially when I found out that I can buy advert´s Toshibas for only 1,7eur in local shop where I usually goes shopping. I´m very hearty with results of good and hard work of Mike and Pavel. I´m impatient for building of this amp.
12 Cents: I used the top of the street lighting lamp for luminous exposure. It was fine but it is hard to find tool like this. 😉
regards,
Karel
12 Cents: I used the top of the street lighting lamp for luminous exposure. It was fine but it is hard to find tool like this. 😉
regards,
Karel
Michael (Kiwanuka, to prevent misunderstandings),
It is absolutely necessary to measure this in a real amplifier. Simulation always fails in HF region when trying to draw and assess the Bode plots.
For once you should take a soldering iron in your hand and build a working amplifier.
Regards,
Pavel
It is absolutely necessary to measure this in a real amplifier. Simulation always fails in HF region when trying to draw and assess the Bode plots.
For once you should take a soldering iron in your hand and build a working amplifier.
Regards,
Pavel
Mikeks, i made bodeplot closedloop because of the effects caused by the 22pf in the feedbacknetwork. It does 2 things, it reduces phasehifts seen by the feedbacknode and lowers CL-gain at higher freqs.
I checked phaseshifts at the inputnodes of the input-ltp. At the point where OL-gain dropped below CL-gain i checked a phasehift of ~72°, giving a phase margin of ~108°. The OL-gain keeps only 4-8db below CL-gain, so gain margin is not very big and caused problems when outputdevices were replaced by the Toshibas.
Maybe you forgot to subtract the phaseshift introduced by the inputfilter ? At these freqs it is ~90°.
The point of 180° phasehift is at ~7-10mhz, which confirmed with reading from the scope while amp was oscillating.
Symasym is not unitygain stable, so would not survive 100% feedback.
Once symasym is stabilized, there is no further instability detectable, and it does a good job driving difficult speakers.
Mike
I checked phaseshifts at the inputnodes of the input-ltp. At the point where OL-gain dropped below CL-gain i checked a phasehift of ~72°, giving a phase margin of ~108°. The OL-gain keeps only 4-8db below CL-gain, so gain margin is not very big and caused problems when outputdevices were replaced by the Toshibas.
Maybe you forgot to subtract the phaseshift introduced by the inputfilter ? At these freqs it is ~90°.
The point of 180° phasehift is at ~7-10mhz, which confirmed with reading from the scope while amp was oscillating.
Symasym is not unitygain stable, so would not survive 100% feedback.
Once symasym is stabilized, there is no further instability detectable, and it does a good job driving difficult speakers.
Mike
Originally posted by Sheldon
..."You have put together an amplifier that sounds very good and is technically sound. You have the basis maybe for a top-notch performer. But to get there you will have to do the final tuning by ear, ...
Sheldon
Originally posted by PMA
But he would have no right to speak about "technical sound", if he did not build the amp, and the exactly same amp (difference counts).
.... Symasym is not "technically sounding" at all, and it is not "the amp like every other 'why me' design"....
"Sound" sometimes means just "OK" "correct", like "save and sound" and I guess Sheldon meant this
..."You have put together an amplifier that sounds very good and is technically sound. You have the basis maybe for a top-notch performer. But to get there you will have to do the final tuning by ear, ...
Sheldon
Originally posted by PMA
But he would have no right to speak about "technical sound", if he did not build the amp, and the exactly same amp (difference counts).
.... Symasym is not "technically sounding" at all, and it is not "the amp like every other 'why me' design"....
"Sound" sometimes means just "OK" "correct", like "save and sound" and I guess Sheldon meant this
darkfenriz said:
Thank you, you are right, I got it. This makes sense. Thanks for explanation and I apologize to Sheldon.
Re: can i use other output bjts like Sanken 2SA1695 / 2SC4468 ?
Hello Michael (MikeB),
my 4 pices PCB sysasym v2.0a have arrived, so i can start to build the amps !
Yesterday I order 2x 2SA1695 / 2SC4468 because the MJL are soldout from reicheld.de.
Later more ...
Hello Michael (MikeB),
my 4 pices PCB sysasym v2.0a have arrived, so i can start to build the amps !
Yesterday I order 2x 2SA1695 / 2SC4468 because the MJL are soldout from reicheld.de.
dx.master said:
Later more ...
2SA1695/2SC4468
Uwe,
these Sanken devices are little bit slower, so there probably should not be the stability problem (probably, should be tested).
But hFE (current gain) of these transistors falls quite quickly with Ic, IMHO the possibilities (parameters) of the symasym will not be fully utilized with these devices.
Pavel
Uwe,
these Sanken devices are little bit slower, so there probably should not be the stability problem (probably, should be tested).
But hFE (current gain) of these transistors falls quite quickly with Ic, IMHO the possibilities (parameters) of the symasym will not be fully utilized with these devices.
Pavel
So I went into my local favourite 2nd hand thrift store and got me a "Kurzzeit-Heimsonne 3000" (nice german words, huh 😉 ?). It has one UV lamp and three Infrared heat emitters. Unfortunately only IR or IR + UV can be used. So I thought of simply disconnecting the IR emitters, but they are wired a little bit complicated. Please have short look, if I could do the shown modification with
a) a wire
b) a resistor of 80 Ohm and 5..10W rating
I measured 80 Ohms resistance for the two outer IR emitters in parallel plus the third in series (disconnected). Forgive me if this is just too off topic.
a) a wire
b) a resistor of 80 Ohm and 5..10W rating
I measured 80 Ohms resistance for the two outer IR emitters in parallel plus the third in series (disconnected). Forgive me if this is just too off topic.
