Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 29, 2024, 05:27:30 AM

Login with username, password and session length

Recent Posts

 

acoustic 150 zobel?

Started by ilyaa, April 15, 2015, 09:20:04 PM

Previous topic - Next topic

ilyaa


am i write to say that R404 and C404 make a zobel network across the output?

what would cause R404 to blow? my guess: amp plugged in without a load puts the output wattage across the puny 1W R404 and blows it up -

but then isnt that resistor just asking to be destroyed? ive seen one in another acoustic 150 that was blown literally in half...

J M Fahey

#1
Zobels turn amplifier crazy speaker load into a resistive one at very high frequencies, otherwise it's usually unstable and can oscillate to death, literally.

So resistor values must be in the area of typical speaker nominal impedances .

It was found that 10 ohms works almost everywhere, some use 22 as seen here, I have seen some chipamps which have tons of gain and bandwidth using as low as 1 ohm  xP or even just the cap to ground  :o

Will it die?

Well, it always has a small value cap in series (around .1uF) which normally sets the total impedance.

A .1uF cap shows 1600 ohms at 1 kHz  :o so no danger here.

It will show 160 ohms at 10 kHz , still safe

Now if the amp oscillates at, say, 100kHz, it will only be 16 ohms, and the 22 ohms resistor will fry in no time, thend the cap often cracks .

So under normal use , even abuse (but at audio frequencies), no big deal; at high ultrasonic frequencies it won't last much.

As usual, it being dead is more of a symptom than the cause of a problem.

Of course, if you don't replace a bad one, new power transistors may die for no apparent cause.

Enzo

Remember here that the output power is determined by the load, not the amp.  All the amp does is provide an output voltage, the load then determines teh current draw from Ohm's Law.  The amp doesn't push out power.

ilyaa

QuoteAs usual, itnbeing dead id more of a symptom than the cause of a problem.

what might be the cause?

QuoteAll the amp does is provide an output voltage, the load then determines teh current draw from Ohm's Law.  The amp doesn't push out power.

doh! i knew that....

phatt

If I recall, some Acoustic rigs have Hi freq horns? If so,
What would happen if the main speaker became disconnected?
I've had that zobel thingo fry when trying to drive only a horn. :duh
Just a guess,
Phil.

J M Fahey

#5
In that case the problem is ultrasonic oscillation in the amplifier.

Why would it oscillate?

Lots of possibilities: there might be a ground problem, the user may have connected a mess of badly set and matched pedals in the loop, thus injecting any potential problems straight into the power amp, I have *often* seen 2 distortion pedals cascaded  xP , flanger/chorus/digital pedals have high frequency clocks which may be poorly filtered and reach the output.

I was once called by some desperate customers who were playing our River Plate Stadium, 42000 people, and in the first half of the show the PA guy had to set the guitar real low and dull sounding because a cheesy FM station could be clearly heard mixed with it.
The cause of *that* problem? (which I had to solve in a darkened stage, with a flashlight held by my teeth, during a 15 minute break) was a loose jack nut in a Boss pedal which made that particular one groundless, touching the case created all kinds of synthesizer type noises.

Funny thing is that nobody suspected that particular pedal because it was "off" .
Big deal, it still was part of the grounding chain ... and failing at that.

I also remember many cases where an SS amp started getting real hot, unbearable to the touch and tripping the thermal protection, if any, with nobody playing it, or just lightly strumming some chords at low volume just to set it up.

Generally pulling pedals one by one or replacing cables solved the problem.

Another problem is that sometimes a cabinet with a piezo tweeter is poorly wired (for keyboards, acoustic guitar or voice) and the piezo is connected straight in parallel with the main speaker.

Problem is that the Piezo *is*  a capacitor, around .15uF , and many amps do not like that load.

The proper way is to wire a resistor , often 47 ohms, in series with it, so even at high frequency the impedance is never less than that.

Wrong:


wrong:


right (the 2kHz horn already has 47 ohms inside)


right, if somewhat of an overkill:


right, although still incomplete:


see that even some "official" diagrams show the piezo straight connected, no resistor, which *some*  amps can stand, but not all.

Roly

Quote from: ilyaaam i {right} to say that R404 and C404 make a zobel network across the output?

what would cause R404 to blow? my guess: amp plugged in without a load puts the output wattage across the puny 1W R404 and blows it up -

but then isnt that resistor just asking to be destroyed? ive seen one in another acoustic 150 that was blown literally in half...

Yes that is the Zobel network.  Doug Self says it's for high and very high frequency stability (but I personally wonder if it's Elephant-Proof Paint, insurance against something that is very unlikely to happen in a guitar amp not driving piezo tweeters or a crossover?).

{
Quote from: phattIf I recall, some Acoustic rigs have Hi freq horns? If so,
What would happen if the main speaker became disconnected?

Good observation.  Most likely the load would then look fairly capacitive due to the LF blocking cap in series with the pressure driver.

Perhaps you have in mind the 271;


The 150 seems to have a choice of 2x12;


4x12;


or 6x12 cabs;

}


C404 going short (or full power supersonic signal from somewhere).


No.  This is a solid-state amp so it looks like a battery, a voltage source with a very low internal impedance.  Unlike an amp with an output transformer (e.g. valves) it makes very little difference if a load is connected or not.  The supply rails dictate the highest voltage you will get on the output, say around 35-45Vpk depending on the amp rated power and load impedance.  The output can even go squarewave but there is no way the voltage can go higher than the rails.


Exercise:
Taking the peak output voltage you can apply it to the Zobel at say 10kHz (a reasonable maximum for a stage amp), work out the reactance of the cap, Xc = 1/(2 Pi f C), and thus the highest current through the Zobel.  You can then work out the power due to the current in the resistor.

At 0.1uF and 22 ohms/1 watt this is a pretty typical Zobel, they don't change much, maybe 0.22uF, maybe 47 ohms, but that's about it. {and as JMF says about chipamps.}

I cheated and used LTSpice.  The current in the Zobel at 10kHz/35Vpk is a bit over 200mApk.

200/root(2) = Irms

200/1.414 = 141.4427157

P = I2R

(.141^2)*22 = 0.437W  Well under its 1 watt rating.

Even with a lot of higher harmonics it still wouldn't be too uncomfortable at full power ('tho the power transistors might be).


Now let us assume that the cap has shorted and we just have the 22 ohm across the output.

35Vpk = 35/1.414 = 24.75247525Vrms, call it 25V

{intuitively we should estimate that a 22 ohm resistor across 24 volts is less than an ohm per volt and so is going to pass more than an amp, and an amp in 22 ohms is 22 watts.}

P = E2/R

25^2 / 22 = 28.40909091 watts.  One watt resistor --> POOF!.


There are a couple of possible ways here; either your experience of these amps is a statistical outlier, or perhaps the manufacturer had a generic problem with these caps (miss-specified, faulty batch, whatever).  You could "HiRel" this by using a mains-rated "X" or "Y" series yellow block cap.  These are available in 0.1uF, 240VAC (630VDC test) rating, and are self-healing if they punch through due to a mains spike, just a small loss of capacitance and they don't catch fire or short.

{I might use a scrounged one in my own gear, but a new one in a client's amp.}

Perhaps this amp encountered a supersonic signal somewhere, or possibly these have a generic supersonic instability problem, possibly as the caps age.  You could try putting a low level squarewave through it and see what it comes out like.  If there is any stability weakness you will see it as overshoot on the edges and possibly ringing oscillations.




I also had cheesy FM radio breakthrough on a Tandy PA used at a Remembrance dawn service(!) due to speaker lead pickup and no Zobel or output inductor to stop the RF getting into the NFB loop.  0.5 ohm 5W/ww in series with the output resistor fixed.
If you say theory and practice don't agree you haven't applied enough theory.

gbono

#7
I always look for simple things that can cause big problems.  ;)  Acoustic put a note in their 150 service manual warning that the user/musician must not use a shielded cable (especially the coiled type you don't see anymore) to connect amp to speaker caninet, because you were in jeopardy of creating a very high frequency "tuned resonant" circuit that would destroy the Z network. Check you cables.

phatt

Quote from: gbono on April 16, 2015, 01:00:47 PM
I always look for simple things that can cause big problems.  ;)  Acoustic put a note in their 150 service manual warning that the user/musician must not use a shielded cable (especially the coiled type you don't see anymore) to connect amp to speaker caninet, because you were in jeopardy of creating a very high frequency "tuned resonant" circuit that would destroy the Z network. Check you cables.

Good point :dbtu:
why yes I still often run into well seasoned players (old buggers like me ;)) who still don't realize the danger of using coax for speaker cable.
Phil

ilyaa

whether or not it was something simple is not, unfortunately, for me to tell - i bought this amp at a swap meet as a fixer-upper -

replaced the zobel -

looks like things are working OK but there is definitely an oscillation - a square wave in the input looks bad at the output - sharp overshoots on the edges. and a sine wave looks wobbly and double-vision.

any strategies for tracking down this kind of problem, aside from trying to isolate whether its pre- or power-amp?

J M Fahey

Quote from: ilyaa on April 20, 2015, 01:54:49 PM
looks like things are working OK but there is definitely an oscillation - a square wave in the input looks bad at the output - sharp overshoots on the edges. and a sine wave looks wobbly and double-vision.......... any strategies for tracking down this kind of problem, aside from trying to isolate whether its pre- or power-amp?
None of that shows an oscillation ... at all .

Scope screen reading 101:

* square wave with sharp overshoot?: treble/bright boost in the amp .... very very normal in a Guitar amp.

* wobbly and double-vision sinewave?: the sinewave is perfect, you are having a sync problem in your scope.

What frequencies are we talking about anyway?

Roly

Square wave test - caution: (I should have said) if fed through the preamp the EQ will have a profound effect on waveshape, therefore this test is really only meaningful when applied to the power amp alone (e.g. via Fx Return).


Quote from: J M Fahey* wobbly and double-vision sinewave?: the sinewave is perfect, you are having a sync problem in your scope.

My first thought also.


Your enemy;


The lower trace is showing bursts of high frequency parasitic oscillation on a signal. Unlike straight supersonic oscillation parasitic oscillations only occur under certain circuit conditions of drive and load, the circuit is said to be only conditionally stable, and this is the implication of the Acoustic caution mentioned by gbono above.  Ideally all amplifiers should be UN-conditionally stable under all load and signal conditions.

Using a square rather than sine wave is poking a stick down its hole, a serious provocation for an instability to show itself.  The sharp edge of the square wave is an electronic version of giving it a whack on the case.  If anything is going to ping, ring, or burst into HF oscillation, that's when it is going to be "shocked" into doing it.

I've encountered the very occasional roasted Zobel, but apart from obvious stuff like the cap shorted, it's hard to tell what conditions may have caused the failure since they always seem to be situational, unreported to me (e.g. using a curly cord for the speaker), and no other fault can later be found in the amp.  I've had amps in for other repairs where the Zobel must have been burned out years before (e.g. fluff accumulating on the burnt resistor), the cause lost in the mists of time, and perhaps multiple ownership.

Most times a bit of detective work will develop a reasonable explanation for a fault, (e.g. one of the speaker leads has an intermittent short), but there are also a small percentage of repairs where is is not obvious from the equipment and there is a lack of other information.

{I'm reminded of a multi-headphone amp from a studio with one output seriously fried.  The only thing I could say for sure is that there wasn't enough power available within the unit to cause that much damage, so the power had apparently come in via the headphone socket.  I can only guess that somebody accidentally plugged an amp speaker lead into the headphone output (both 6.5mm) and gave their amp a thrash.  It certainly isn't the first time I've seen damage from that mistake.  For a conscientious repairer a guess, even a good one, isn't very satisfactory because it's nice to be certain, but without somebody putting their hand up it had to remain a mystery.}


If you say theory and practice don't agree you haven't applied enough theory.

ilyaa

hmmmm ill do some more snooping -

the zobel cap was only rated for 100VDC, so maybe a mains spike shorted it?

Enzo

It may be shorted but that is not likely why.   Anything on the mains that could do that would have damaged a bunch of other things too.  To get to that cap, a mains spike would have had to make it all the way through the power supply, and then down through the + side outputs to get to the amp output bus, THEN through the resistor to zap that cap.

Roly

Yep.

Maybe it just decided that it was mad as hell and wasn't gonna take it any more.    :lmao:


{'course if you are feeling lucky you could try to re-create the fault conditions.  Repair the Zobel, hook your CRO on the output, and connect the amp to its speaker via a long curly cord, and hold on to your hat.   8| }
If you say theory and practice don't agree you haven't applied enough theory.