ra500 channel blown

[kepeb]

hello.
i just got an alesis amp with a channel down.
would anyone like to explain to me the cause and effects of this particular failure?

here is the information i have...
as i am to understand, it is believed the channel went down after the speaker was disconnected while it was on.<p>
<p>
<p>
<p>
the amp powers up and the output for the second channel works fine.
this channels board is receiving power, and on initial test would still pass minimal signal to the speaker with the pot at about 50% (although the clip light came on but not the signal level light)

the components i have circled on the schem either looked slightly or even severely burned as pictured. or they were in parallel (r146-148 & r157-159) so i have clipped them all out to replace. but i dont assume this will fix it and would like to better understand the cause.

from what i can tell the transistors check out well according to the simple multimeter tests i have performed after reading this site..
Basic Amplifier Repair

two resistors burned away completely, r146 and r158 as pictured. would this of happened instantly in the event of a load change? or was it possibly running for too long with a larger current going through?

any thoughts?

[Enzo]

Load change? NO. Havi9ng the tip of a speaker cord shorting to its sleeve as it pulls out of a jack, sure.

The thing is, don;t focus on how it might or might not have caused the damage. The amp is as it is, and we need to fix it.

D111,112 are probably OK, it is more likely the transistors in parallel with them are shorted. You have 6 output transistors, Q118-121 tucked under the board. Are not at least some of them shorted emitter to collector? And the 6 white power resistors standing on end, check them to see if they are now open.

Look at your schematic. Those 10 ohm resistors going to the output transistor bases are not in parallel, they are connected together at only one end. If they read as in parallel it would be because the transistors are shorted internally.

Now look at the larger picture. I mean look at the schematic overall. You have +V at the top and -V at the bottom. Look at the circled parts. The three 10 ohm R146 et al, then the 220 ohm R140, anf then the three other 10 ohm R157 et al. (Please correct me if I get a part number wrong) The only way a resistor burns is if excess current flows through it. Look at the only current path that blows up those resistors. It is through the output transistors. That is why I suspect they are shot. Further, looking at those 0.22 ohm 5w emitter ballast resistors, I imagine they haver opened, leaving those smaller resistors as the only other current path.

Not only that, but the 100 and 200 ohm resistors just to the left are also burnt, which makes me suspect that driver transistors Q116 and Q117 are probably also blown.

[kepeb]

woohoo a reply! thanks man.

ok. i still have it in front of me all the white (0.22 5w) resistors measure ok (more like 0.60-0.70 but all very close).

they aren't in parallel no :S, and they didn't read as that. but as they looked slightly suspect and i had the same values at hand so i replaced them. my bad wording/knowledge.
i figured they possibly may have had more current than needed through them already, but the one next door was weaker and took the brunt of it and burned completely. oh well no harm there so far just swapping them out.

i must admit i still have limited knowledge on this but i thought that may be the case, and that's really as far as i got but i just checked all the output transistors again both ways on each c-e junction and there is no short... and it appears the same with the driver transistor too.

please let me just confirm my test:
i found the data sheets for the concerned transistor online, found the pinouts, and checked both B-e and B-c BOTH ways with my dmms diode check function.

i also checked q112 as the diode at its base was looking a bit brown and that seems ok.

thanks for taking the time to look this over for me

[jrfrond]

With a direct-coupled circuit like that, which is typical of power amps, ALL of the semiconductors are potential game for replacement with a catastrophic failure like this. A simple failure would be single shorted power transistor.

Just a wild guess looking at the failure mode, you probably have one or more bad power N devices (2SC5200). However, I would desolder all the power transistors, make sure the leads aren't touching their traces, and do the standard 5-way test (B-C/B-C reverse/B-E/B-E reverse/C-E). ANY leakage is unacceptable. Replace the defective one(s) and their associated emitter resistors. I'd replace all four diodes (2 x 1N4004, 2 x 1N4148). And yes, it's likely the drivers are blown too.

[kepeb]

really?
just replace everything?
ok. well i guess i can still learn from that, in this case.

just so i dont misunderstand

"do the standard 5-way test (B-C/B-C reverse/B-E/B-E reverse/C-E). ANY leakage is unacceptable".

you do mean 'any continuity from any point to another' is not acceptable with transistors?

so. is this damage likely to have occurred from the speaker being unplugged while on?
what would of caused this if not?

if it is, what exactly happens when the out is shorted? how does a shorted output force more current through these components if its not a change in the load?
whats this 'failure mode' ?

and i guess more importantly, what other bits should i be de-soldering and ordering replacements of?
thankyou

[jrfrond]

Actually, I didn't say that you SHOULD replace all of the semiconductors, but ALL should be tested.

The 5-way test simply checks junction in forward and reverse polarities. Do you have a DVM with diode test? Or a good analog VOM? Either one will work. In fact, man old warhorses like myself (and I am guessing Enzo too), actually prefer the analog over digital meters.

If you don't know how to properly check semiconductors with a meter, then for sure you might be in over your head with this amp. You also need a variac to bring up the voltage slowly to watch for current draw and "soft" shorts. If you replace components and miss one, and power it up blindly, you can potentially ruin all of the work you've done. Happens to experienced techs too. Just some fair warnings when working on SS power amps.

[Enzo]

Failure mode just means a part or circuit that no longer functions like it should. Nothing more spedific than that in this context. Just like I am in my "at work mode" right now, and I will be in my "off duty mode" in a while. So then looking at how it burnt up...

let's talk about the meter test. A transistor checks like it was a couple diodes connected to the base. You check from base to collector as one diode, and base to emitter as another diode. presumable if both "diodes" are OK, then so is the transistor, right? No, not right. What if both those "diodes" are OK individually, but the collector is shorted to emitter? The two "good" diodes would now be in parallel, and the two individual tests would not reveal that. SO you need to test emitter to collector in addition to the base to either one testing. And since transistors are one-way devices, you have to also do each check in reverse to make sure there is no reverse leakage.

Look at the first photo, Q121 is easy to see. Upside down from how I usually look at them , but that doesn't matter. Three legs across the top. The one on the right is the base, see R157 going to it? The center is hte collector and the left the emitter. Go down teh row and verify none are shorted center leg to left leg as we look at them here.

While in circuit, it is difficult to take real radings of transistors because of all the parts around them. In circuit readins are to help determine that something is PROBABLY bad. But if in doubt, we pull the part and test it all by itself, then nothing can confuse. And in that condition we don;t expect to see any reverse leakage.

A shorted output means that when an output transistor turns on, instead of it sending current through a speaker as it expects, it winds up sending current to a dead short. Imagine what a set of jumoer cables would do if you connected them to yo9ur car battery and then just shorted the other ends together. They catch fire.

SO the problem is not a "change in impedance," the problem is simply the short across the output.

There is enough wrong with this amp that we can see to know that the parts in it suffered some serious stress. That is why John is suggesting just replacing those parts. Even if they test just fine, can we be confident they are not weakened and just waiting for the first real exercise?

ANother thing about a meter test. The meter is a very basic instrument, it cannot test these parts at the full operating voltage, nor can it run remotely realistic amounts of current through them.

[kepeb]

That's great thanks for the the detailed replies
sorry John, I didn't mean to correct, I just wanted to be sure I was understanding properly.
I was using the 5 check process enzo just described again with my dmm on 'diode test' as I confirmed in my second post. I see now that you meant no leakage JUST at the junctions that aren't supposed to pass signal.. for a second i thought my understanding of transistors was right out so. i did the checks again, same thing.


I get that i cant test these parts at proper operating conditions either, and i can see the parts have suffered a great deal of stress, hence my replacing all the parts i had replacements for which I suspected MIGHT have been overworked. I will replace the transistors too of course then.

i have heard before that there is a preference for accurate analogue meters over digital ones. i think the one pointed out to me was 'IVO'?

anyway. i only have a portable dmm right now but my interest is still high in this kind of thing so i'll probably invest in a varic shortly if it is so commonly used in amplifier testing. i was put off by reading how they can damage my tube amps(i'll post address if i can find it again)


outside of the driver/output transistors and the components circled earlier are there any others i should be looking at or would the damaging high current have gone straight to the common just through the transistors discussed? and just to clarify then enzo, its not that the load was changed, its the fact there was NO load ie. a short/zero resistance - so the maximum passable current was exceeded straight to the common/ground?...

[Enzo]

Nothing wrong with digital meters, I have been using digital meters professionally since the 1970s. I can read 5.06vDC on my Fluke DMM. I defy anyone to get that reading on an analog meter they are likely to own.


And to be technical, a dead short is not NO load. Open circuit is no load. A short is the maximum load possible. And unless there is current limiting circuitry, that is likely to damage the outputs.

The outputs and probably drivers ought to replace. That leaves just five small transistors in this basic amp circuit. And looking at them, they appear to cost less than a quarter, so why not just boot the whole cast? Verify all the resistors.

[kepeb]

thanks, i will replace all in this section.
should the damage be isolated to this part of the circuit... there is more to it:
http://electronics-diy.com/pdf/Alesi..._Amplifier.pdf
i am assuming the rest would be ok?

thanks for explaining too

[Enzo]

You never know until you try it, but the rest of the amp is probably OK.

[J M Fahey]

I'd like to add to the above excellent posts a couple minor details.
Just in case:
1) scratch all carbon, it is conductive and will throw off your measurements. Brown and still hard is OK (sigh!) but black and crumbly is not.
2) pamper the PCB, use a good solder pump, don't overheat the pads nor wiggle parts when pulling them.
Solder wick can be an extra detail, it sure cleans pads.
Parts "should" to be able to be pulled as if just inserted yet not soldered.
Well, at least as close to that as possible.
Good luck.
PS: if you haven't got your variac yet, build a Lamp Bulb Current Limiter, search around this forum.

[kepeb]

hey, cheers

you mean the pcb when you talking about carbon? where the resistors burnt through etc?
i brushed it down with a wire brush to get off all i could, thankfully the majority of damage was on top away from the tracks.

yea and desoldering.... this is one of those things. i have a solder wick and i have upgraded my pump and bought an aluminium one with better construction. and sometimes i just still cant get it right. especially when there are other components in the way. it seems to take an age to get a part out clean. maybe i'm missing something.

and i dont have a variac yet no. but since i am replacing everything in this section now anyway. i thought that could wait.

[J M Fahey]

Yes, it's basically that.
If no variac available, build a simple and cheap Bulb Lamp current limiter.
Don't worry , you are missing nothing, sometimes removing solder is pesky.
Anyway, "practice makes perfect", as in anything else.
Just do *not* plug it straight into the wall, even if you think you replaced everything suspect, you may have missed the tiniest detail.
Always play it safe and use some form of current limiting.
Good luck.