Needing opinions, the "return of someone that had never departure"

[Upupa Epops]

Carlos, by my personal experience isn't good to use in multiband system different amps for each band.... Amps must be the same ( they must have the same connection ), they can have different RMS power, but the same peak power.

[Samuel Jayaraj]

Carlos,

Thanks for all the help that you rendered in the GEM amplifier thread that both motivated and enabled me to complete the GEM amps. They are indeed good but just check my comments on the sonic qualities of the GEM thread.

Ofcourse, you have the right to your subjective appreciation of the sound of all the amps that you have right now rated as the best. I would still suggest that you try out a Stochino amplifier which is very close to an Alexender amp. Once you try out the Stochino you will know how all the others fade away in performance. I have not yet constructed MikeB's amp, but the GEM, Aksa clone and LM3886 from a host of others have been built and tested. The Stochino is among the 'very best' possible with Class AB amps and will give the best of the Class As a run for their money.

Since you have already built over 4k amps, another one will not hurt. Come on, do it.

[destroyer X]

The same peak power..... as music, sometimes can have trebles with levels over 10 percent...and this may distort!.... All rigth, i will use same voltage to all amplifier, and will take care to have this capacity of holding 50 watts peaks each one of them.....well, subwoofer i will make it 4 times bigger, as i perceive that this proportion works fine.

Interesting Epupa....and i think you are completely rigth...if i made the way i was thinking, i could have treble clipping....rigth!...will follow your idea...now my idea too!

Samuel, can you show me the Stochino circuit?...i already had it in my amplifier files....but not anymore...they may use Fet!...and as i do not have easier FETs here...normally i jump over those amplifiers, and i delete them on my files....my experience is only about a couple of tube amplifiers and thousands of BJT amplifiers only, and many of them were repeated in thopology, i used to go constructing what was appearing in magazines..... i cannot say a word about FET amplifiers!...as i am deeply, and completelly, ignorant about that subject.

Samuel, i have already printed your GEM text (are they only one...a big one...have other?)...and i will answer you using direct mail...send me please a mail, this way i will have your personal mail adress....i had it!...but i have lost a lot of adresses...had problems with my computer...it is still a little bit crazy!

nanabrother@yahoo.com

regards,

Carlos

[Upupa Epops]

Yes Carlos, the same connection of amps, the same rails, but for treble " softer ".... and you will be satisfied and you will be king in you " harem " ....

[destroyer X]

And i can see that you do not shave your face...ahahahahah!

regards,

Carlos

[Samuel Jayaraj]

Carlos, by my personal experience isn't good to use in multiband system different amps for each band.... Amps must be the same ( they must have the same connection ), they can have different RMS power, but the same peak power.

Upupa, I don't quite agree with your comment.

The more important criterion is 'phase' characteristics of each of the individual amplifiers; if the phase is significantly different, there will be no coherence in the sound.

Think of an active system where the mid and treble bands have to be delayed in order to make the system phase and time linear -this can be done either with a digital crossover, or by including All Pass Filters in an analogue crossover and the mounting baffle of the drivers will also have to be staggered so as to have all the point sources (woofer, midrange and tweeter) on a single vertical plane.

The next important criterion is the tonal characterestics of all the amplifiers used. Ofcourse, you can make a judicious combination, like an amplifier with relatively more F3 for the bass and another with predominant F2 for the mids and whatever pleases your particular tweeter, in order to have a system that sounds wonderful. Read up on the Rushmore active system from Pass Labs and you will better understand.

As long as you have adequate power for the bass region (and nominally sufficient power for the mids and highs) and depending on speaker sensitivity, listening level and room dimensions, POWER is the least criterion. And it is certainly not necessary for all amplifiers to be powered off the same PSU.

[johnnyx]

I had a conversation with a bloke in a hi fi shop about this. I told him I was thinking of using a low power class A amp for the tweeters, and maybe a higher power class B (maybe a valve) amp for midrange, and using an active crossover. The idea being to take advantage of the lower distortion of the class A at low power for the tweeter. He said it would be like having a car with two different engines.
Yet the speakers he was demonstrating at the time had electrostatic panels for most of the frequency range, and a cone bass driver. I thought the difference between cone and panel drivers would be far greater than the difference between class A and class B, or even between 20 watts and 100 watts.
If speakers can have differing designs optimised for the frequency range they cover, then why not optimise the amplifiers?

[Samuel Jayaraj]

Carlos,

Here is the link to the Giovanni Stochino 'Ultra-fast amplifier' http://www.soton.ac.uk/~apm3/diyaudio/Stochino_amp.html

You need to work on the PCB design.

Here is the circuit design: http://www.soton.ac.uk/~apm3/diyaudi...no_circuit.jpg

A few changes are required to what is shown in the schematic diagram.

There is a 47K resistor between signal and ground, at the very front end.

This is followed by 100uf 25V x 2, connected with the + terminals in series and the - terminals to the outer ends.

The emmiter load resistors of the front end 2N2222A and 2N2907A transistors have to be changed to 5E, not 4E7.

The current sink resistors of the input pairs are 50E, not 60E.

The DC blocking caps at the feedback junction ought to be 330uf x 2, not 220uf. Move the gain determining resistor, 330E from ground side to base of inverting input and caps/diodes to ground side.

The cross-coupling electrolytic cap in the next stage is shown as 47uf 50V; this should be changed to 22uf, even 25V is sufficient.

The bias trimpot need not be 10K as shown, 5K will do.

All diodes shown as BAX16 types were specified as IN4448 by the author. He tested these to withstand reverse voltage of 120V and used only the ones that passed this test.

The position of cap and resistor in the zobel network need to interchange places.

Decoupling caps for the output side are 470uf 63V x 3.

All of the above are as per Stochino's original circuit as published in Electronics World of August 1988.

The changes I made are as follows:

Separated quiet and noisy grounds with a 10E resistor.

Used onboard decoupling of 220uf, instead of 470uf for the front end.

Increased onboard decoupling to 8200uf 80V.

By passed input and DC blocking caps with 1uf film caps.

Used BAW62 diodes in place of IN4448.

Caveats:

Unit gain frequency before output filter = 22MHz.
SS bandwidth before output filter(-3db) = 20Hz(-0.1db) - 1.3MHz.
DC open loop gain = 110db
Open loop gain at 20kHz = 66db
Closed loop amplifier phase margin before output filter = +76 deg.
Output offset voltage = 30mV (approx.)
Output noise (BW=80kHz, Input terminated with 50E) = 42uV rms.
Positive slew rate = 320V/uS
Negative slew rate = 300V/uS

Now watch out with your PCB design philosophy.


jonnyx,

Please note that not many tweeters can handle about 12watts of RMS power and very few would max out at 15watts. Elektor did a study not very long ago and reached the same conclusions.

[destroyer X]

me to produce, or to not produce, or to arrange values and power, and phases...hehe.... things a little bit more complicated than i hope for.

But the main question is:

Will be better to use one powerfull transformer with many windings, or to use many transformers separatelly.

Sorry Samuel, but the amplifier is too much complicated to me, and have FETs... those are hard to obtain here...this one i will not construct, as the better he can be, will be not so good to change it by my health...to construct it, i may have a heart attack...i prefer more simple designs, even having some problems here or there....but i thank you very much by the link that i am reading with interest, to learn, but not to construct.

regards,

Carlos

[darkfenriz]

The more important criterion is 'phase' characteristics of each of the individual amplifiers; if the phase is significantly different, there will be no coherence in the sound.

If one speaker is closer to your ear by only 3cm, than for 3kHz crossover freq. it makes phase shift of 15 degrees. How much do you expect a solid state amplifier to shift phase at 3kHz?