Why i always have those losses in power using lower impedance loads?

[destroyer X]

I made this question to many folks, directly, using mail.... none understood my question.

Some said this is "crazy" (was to 1 ohm example i have made)
Other said how to tweak to reduce the problem.

My question is:

What factors, were are the reasons, what part to be blamed.... who is the guilty? when i face those things... example:

- Let's imagine an amplifier that is using a special supply that has not losses of voltage...no voltage drop when feeding 1 to million amperes..some special mythic supply.

- Now let's make some amplifier (every amplifier.... any thopologie..standard or sophisticated unit..anyone!)... and injecting audio we gonna find 20 volts over 8 ohms (also special load, non inductive, non reactive...purely resistive)....

- Now let's replace the load by a smaller load...this time a 4 ohms load.... i will be waiting the same 20 volts over this 4 ohms load...let's imagine that i will adjust my input signal to obtain those 20 volts.... now, things make sense, and the power increase to 100 watts....no problems...this makes sense to me.

- Now...let's replace the load by a even smaller load....now it will be 2 ohms load..... i will be waiting to see 20 volts as my supply had not losses in voltage.... but i have checked into real world and into simulators that this do not happens..... you can adjust your input signal and negative...you do not reach 20 volts anymore...so...the possible 200 watts is unreachable!!!!...why?...this is the question..... what component is the responsable?..... where is the losses?.... is saturation voltage that will variate depending the current?...were is the guilty?

Yes...of course i tried several tricks...i have increase the output transistor till 6 pairs...using emitter resistances.

Of course i have increase my current into the VAS...i have increase my current into the drivers, also into the output...also i have increase my current into the differential..i have replaced VAS with darlington...i have tried tripple darlington...i have tried the blameless circuit.....but NEGATIVE!...always a loss when you reduce your load impedance...even beeing the input signal adjusted and modifications made into the circuit trying to compensate..... remotion of stop resistances...reduction of emitter resistances....increasing in the output transistors number..including output coil, excluding output coil.... changing frequencies...well...someone.... a guilty part of a technical thing i do not know is "eating" some output voltage.

WHY?

My worries are because i have perceived that good amplifier are the ones have the capacity to manage "difficult loads".... related the speaker crossover impedance and deep in impedance that we can find in some speakers...not to have losses, i am thinking that one good reason to be good may be this current capacity.

I have tweaked my Dx amplifier and i could find that after tweak.

50 watts over 8 ohms undistorted
100 watts over 4 ohms undistorted
153 watts over 2 ohms undistorted.

Why not 200 hundred whats over 2 ohms into simulator....there, in simulator we have not voltage losses when the amplifier ask for current.

I hope someone can help me.

I am not really searching for solutions only...i will be happy to "understand" the reasons...why i have those losses?

Yes...i have to be blamed...i am the guilty and the responsable...so...already knowing that.... i am answering, humble behavior, trying to scape from the deep ignorance that is bothering me.

I have asked 5 guys.... 3 of them answered with a new question...one tryed to suggest bootstrapp increasing of current (but this is something that is happening in every topologie).... i hope someone can really help me in this problem.

Please...translate to simple words...do not post mathematic formulas...i need a simple answer, not to have sure you are a good Mathematic man.

regards,

Carlos

 

[richie00boy]

I found that when simulating I got similar results, I put it down to the models not being perfect. Don't forget that lowering load impedance means more current, so more voltage will be dropped across the output device. Also in real world more losses will be in the wires and other components.

 

[destroyer X]

So...you are telling me that the saturating voltage..from colector and emitter will

be variable?

Depending the current i will have different saturating voltages?... this can explain losses when driving lower impedance loads.

This may be the answer i want to have.... are you sure my dear Richie?

Interesting that.... experienced people that i have asked could not even understand my question..so simple it was.

regards,

Carlos

 

[destroyer X]

Yeah!..thank you Ricchie.... so simple the answer... already answered in the

data sheets i have.

I could see differences in the saturating voltage depending form the current intensity....so..this is the answer.

Thank you Richie00boy.... you were fast, efficient and clever...and have beated a lot of older and experienced guys.... thank you by this help...i will remember you because of that...always.

I could not discover that alone!

Ahahahhaa...now i know how to behave related that into circuits.

In my point of view, this thread can be left to die...as the question was answered with competence.... no need of this thread anymore, related my personal needs...of course others can use this information to ligth their minds too.

regards,

Carlos

 

[sreten]

Hi,

In most amplifiers "standard" amplifiers a 2 ohm load means you
will get significant losses in the output devices emitter resistors.

Even without emitter resistors you still have Re losses.

If you can build an amplifier that goes from 50w to 100w into 4 and
8 ohms, then by double paralleling the output stage it should go from
100w to 200w into 4 and 2 ohms by definition.

(Assuming your output inductor has a low enough DCR)

/sreten.

 

[Christer]

While the losses mentioned will be a limiting factor eventually, I would think that most amplifiers will rather be limited by the current capability of the output devices. Remember, for instance, that BJTs suffer from so called beta droop. That means that beta (or hfe) starts to decrease rapidly over a certain current,

 

[sparkle]

Carlos - the question is very interesting - i was thinking about it too..... the problem is that we have to have quality that will be able to withstand difficult speakers that sometimes drop down to 2 ohm...
so - as Carlos said - we need 50W/8ohm, 100W/4ohm and almost 200W/2ohm....

o.k. for what i know is - i should double the output devices, have few volts more on VAS than on output devices, have lower emitter resistors in output or no resistors at all, also have very fast filtering caps and big transformer with fast rectifying diodes....also thick wiring....

but - what more ?!

 

[richie00boy]

Glad to have helped Carlos.

 

[destroyer X]

Yes Sreten and Christer.... this make sense.

Thank you folks.

The gain is reduced when we drive big currents.

also the emitter resistances play a part in this game.

Maybe i will have to search for reduced saturating voltages units...if they exists!

This is not so good... a very annoying limiting factor, as there are enormous deep in impedances in some speakers, caused by themselves or by the passive crossover.

Well..at least i already know the cause...and i am happy to have this information now.

A very clever Consultant Mr. Richie00boy.... he is the one have the competence to understand my strange "englanderishen" (not a real english) language....

thank you all...in special by Mr. Doctor Richie00boy.

I would be happy to know your personal family name Mr.Richie...my respects to you...as better than know things is the competence to understand questions and answer them with precision.... also not using this as a chance to show yourself as and "illuminated"....thank you once more.

Also, even more important, the cooperative mind.... ready to help folks.

regards,

Carlos

 

[sreten]

I found that when simulating I got similar results,
I put it down to the models not being perfect.

Hi,

D.Selfs work on the subject shows the SPICE models are very accurate.

Low impedance loads cause a rapid rise in 3rd harmonic distortion.

Beta loss at high current in the output devices induces the
distortion in the driver stage due to Early effect in the drivers.
Turn off the SPICE parameters related to the Early effect in
the drivers and the distortion disappears.

I suspect a triple output stage would be good, if its stable.
Take a normal double emitter follower output stage and
rejig the driver stage as a CFP (complementary follower pair)
by adding small signal transistors to the driver transistors.

/sreten.

 

[richie00boy]

SPICE models can be accurate, but it depends on the model and it's complexity/completeness. Not all models are good.

Thanks for your kind words Carlos.

 

[sreten]

Hi,

True. But the point is accurate SPICE predicts these losses.

/sreten.

 

[destroyer X]

Michael Bittner informed something that completed everything into my brain.

Now i have my old 5 neuroniuns working harmonically.... all them slowly processing data.

He informed that this saturating voltage losses can be compensated using 4 output pairs.

Also introducing the needed current, into the drivers and output...from the VAS that current.

So...a matter of a good supply, a good design, enougth currents and good output pairs and the lost will be smaller...of course emitter and some losses will be accepted, as finally i know the reasons and this loss do not surprises me anymore.

I am extremelly satisfied with all folks contribution, informations and kind assistance.

regards,

Carlos