Have a roland cube 40 amp DOA. Both internal 4amp fuses blown. My first check was the pair of output transistors Q11/Q12 NEC D586. Both where dead. After that, I got the schematic to trace what the source cause failure might be but I'm not seeing it. It's almost like one must have died and killed the other. If you look at the attached schematic of the power section, with the two output transistors out of circuit, all the circuit voltage pretty much match what is expected, with the exception in 12 which I have 0v (I'm guessing expected given the transistor are out for circuit) but have two strange (as in completely different) readings coming their respective bases 0.58 and -33 shown in red on the schematic. Checked all the major resistors,caps and diodes around the area, most everything fine. R54 was reading 300ohms instead of 330 so I replaced it, C43 reading 120 instead of 100uf so I replaced it but everything else looked pretty good. I even pulled Q9/Q10 and tested, they are good. So is it just a matter of replacing those two transistors or is there a deeper culprit? Any help trouble shoot advice is most welcome!
Ad Widget
Collapse
Announcement
Collapse
No announcement yet.
Roland Cube 40 Repair advice needed
Collapse
X
-
Originally posted by rocknroller View PostHave a roland cube 40 amp DOA. Both internal 4amp fuses blown. My first check was the pair of output transistors Q11/Q12 NEC D586. Both where dead.
After that, I got the schematic to trace what the source cause failure might be but I'm not seeing it. It's almost like one must have died and killed the other.
Point is they are dead and must be replaced, thereīs many possible reasons why an unmolested well treated amp still dies, call it old age if you wish.
That amp must be over 30 years old
You check to find if you find something abnormal.
If you look at the attached schematic of the power section, with the two output transistors out of circuit, all the circuit voltage pretty much match what is expected, with the exception in 12 which I have 0v (I'm guessing expected given the transistor are out for circuit)
The 18V inside a rectangle you see by it is Audio/AC voltage, what the amp will put out at maximum power on an 8 ohms speaker.
Notice other voltages shown are not inside a rectangular box but plain written.
but have two strange (as in completely different) readings coming their respective bases 0.58 and -33 shown in red on the schematic.
Base voltage is not read relative to ground but from base to emitter.
On the top one emitter sits at about ground so your measurement roughly matches, but on the bottom one, emitter sits at the negative rail , nominally -35VDC, so -33VDC is close (and makes me think negative rail is actually -33.6V )
Please remeasure from terminals 8 to 10 (top base emitter) and 11 to 13 (bottom base emitter)
If you measure around 0.55V each, amp is fine (you got lucky ) and replacing output transistors is the cure.
Please measure resistance of R55 and R56, they might have blown open.
Please donīt replace them if your meter shows a low but different value, regular meters have great difficulty reading very low values with precision, all we need is to check they are not open.
Checked all the major resistors,caps and diodes around the area, most everything fine. R54 was reading 300ohms instead of 330 so I replaced it, C43 reading 120 instead of 100uf so I replaced it
but everything else looked pretty good. I even pulled Q9/Q10 and tested, they are good. So is it just a matter of replacing those two transistors or is there a deeper culprit? Any help trouble shoot advice is most welcome!
1) build a lamp bulb limiter (search this Forum) for safetyn when you turn it on for the first time
2) turn it on without a speaker connected.
Then measure :
* +/-V rails
expect somewhat less than indicated because the lampbulb (use a 40/60W one) will"eat" some mains voltage but , say, +/- 20 to 28V look fine to me
* DC voltage at speaker out rail (pin 12) , 0V DC or within 200mV
* BE voltage on each output transistor
* the cherry on the cake: measure voltage across (from end to end) R55/R56 , I expect a few mV there, say 10 to30 mV
Post reading here for reviewing, if normalthe amp should be fine.
Then you can connect the speaker (without turning ampoff) and it should pass sound.Juan Manuel Fahey
-
See below
Originally posted by J M Fahey View Post
0V DC is normal and whatīs to be expected there.
The 18V inside a rectangle you see by it is Audio/AC voltage, what the amp will put out at maximum power on an 8 ohms speaker.
Notice other voltages shown are not inside a rectangular box but plain written.
AH - I should have caught that but thanks for the reminder.
You are not reading them the right way.
Base voltage is not read relative to ground but from base to emitter.
On the top one emitter sits at about ground so your measurement roughly matches, but on the bottom one, emitter sits at the negative rail , nominally -35VDC, so -33VDC is close (and makes me think negative rail is actually -33.6V )
Currently, with the more accurate 330 ohms resistor in place, I'm seeing -34.87 on the negative rail
Please remeasure from terminals 8 to 10 (top base emitter) and 11 to 13 (bottom base emitter)
If you measure around 0.55V each, amp is fine (you got lucky ) and replacing output transistors is the cure.
8 to 10 is 0.58v as before so that's good, however 11 to 13 shows 1.3v. So what is suspected for driving that second one high? also see following re: Emitter resistor value
Please measure resistance of R55 and R56, they might have blown open.
Please donīt replace them if your meter shows a low but different value, regular meters have great difficulty reading very low values with precision, all we need is to check they are not open.
Had done those on the first pass, settling around .6 ohms. Accounting for my lead resistance of about .3 ohms I thought I was good there. However, remeasuring now R56 is bouncing between .5 and .4, which leads me to be concerned the value has dropped off a bit. Also perhaps contributing the the 1.3V volt reading? Have to check if I have a replacement for that one in my parts stock, that may be an order item as well.
1) build a lamp bulb limiter (search this Forum) for safety when you turn it on for the first time
Already have one! But will have to wait for the slow boat to arrive with the replacement transistors. On a budget a too cheap to pay the marked up US prices a plus 8.95 shipping!
2) turn it on without a speaker connected.
Then measure :
* +/-V rails
expect somewhat less than indicated because the lampbulb (use a 40/60W one) will"eat" some mains voltage but , say, +/- 20 to 28V look fine to me
* DC voltage at speaker out rail (pin 12) , 0V DC or within 200mV
* BE voltage on each output transistor
* the cherry on the cake: measure voltage across (from end to end) R55/R56 , I expect a few mV there, say 10 to30 mV
Post reading here for reviewing, if normalthe amp should be fine.
Then you can connect the speaker (without turning ampoff) and it should pass sound.
Comment
-
They are not repeat not bouncing.
An unaviodable defect of digital meters dealing with the Real World which of course is analog is what value to *display* for the last digit.
Suppose that resistor is 0.34 ohms (well within rating) but meter can only display 0.3 OR 0.4 ohms, pick one.
So it , say, displays 0.4 ohms.
A fraction of a second later it rechechs and finds hey!!! itīs actually *less* than 0.4 ... Iīll display 0.3 .
A fraction of second same happens but now value displayed is again wrong.
It will bounce up and down trying to hit the right value.
Is the actual resistance value bouncing up and down??? NOT AT ALL.
But you shouldnīt be worrying about that anyway, I TOLD you the measurement is poor and all we want to know is whether said resistor is open or not
As of the about 1.2V across the bottom BE junction, Iīd wait to have the transistor there and remeasure.Juan Manuel Fahey
Comment
-
To be accurate, you do not have a B-E voltage at 11 and 13, you have a voltage where the B-E would belong. Once the transistor is in, this might well fall into place.Education is what you're left with after you have forgotten what you have learned.
Comment
-
Yes in this particular case.
For a crude explanation:
* Top BE voltage is an actual *voltage* served from a "voltage source" which is Q10 emitter, and does not depend much on load (unless actual load is a gross short) so I see 0.58V there and can guess: "hmmmm, when I connect a BE junction there voltage wonīt change much, maybe it will go down to 0.55 or 0.53V , still fine"
In fact you can vary R44 from, say, 100 to 500 or 1000 ohms and BE voltage will vary very little (a few mV).
* bottom BE voltage comes actually from a "current source", collector of Q8, and will directly vary with R54 value.
It might easily go from 50mV to 5 V depending on resistance value alone, leaving everything else unchanged .
Ok you might say, now I get it, but then why didnīt Roland set its value or the current through it so we have also 0.58V as above?
Well, transistors do take base current, in this case said current will substract from the one actually passing through R44 ..... I bet voltage will then basically be same as the other one once you install said transistor
So instal the missing transistors, remeasure, I bet everything will be fine.
And if not (although I doubt it) , it can be corrected.
* there is also another reason why even on a fully working amp both BE drops might be different: this is a non symmetrical output power amp (it would need NPN/PNP transistors yet uses both NPN) so itīs common that the internal circuitry drives one of them more than the other to keep output at 0V DC, which is the main point and only thing which must worry us.
Just curious: if at all possible please post a couple open chassis and open cabinet pictures, this amp is a marvel of compact packaging, I want to make a similar compact combo (with my own circuits and speakers of course) and could always pick a couple ideas.
Specially interested in the "niche" where the L shaped main chassis sits and how/where is it bolted to the main cabinet, also how the inside of the speaker section is separated from chassis inside, so air pressure does not vibrate that chassis/backpanel to death in a couple years.
And some speaker picture too
Of course, only if you find the free time and have a camera available; if not, donīt worry.
I googled tons of Roland Cube pictures, but all are outside onesJuan Manuel Fahey
Comment
-
Thanks a lot, it was very kind of you and sure clears many doubts.
I was amazed by quit a few details:
* quality of construction and parts used.
* all wires go to riveted and soldered terminals (lots of hand wiring) , not crimped or straight through PCB holes
* all soldered connections marked/protected by "nail enamel" or similar.
* the Japanese madereverb tank. Guess Accutronics sued them because later Rolands and Yamahas used US made Accutronics tanks.
* the funny reverb shelf on top, and the tank supported by "bed mattress" type springs. Weird
* the power transformer bolted directly to cabinet wood, no metal chassis or subchassis involved.
FWIW I always do the same, but everybody calls me cheap for that.
I guess Iīll tape this picture to the shop wall
Not so impressed by the abbundant use of chipboard.
Same for non-separation between chasis and speaker cabinet.
That said these amps last years, so .....
The speaker looks very good, certainly has a beefy magnet.
Is it a 10" ?
Now to:
what is the summary synopsis here? Q11 up and died on it's own and took out Q12? Or Q12 up and died on it's own and took out Q11? (or the third more remote possibility statistically speaking - they just both up and died on their own.)
So much so that some datasheets offer MTBF data (Mean Time Between Failures).
Can we make stuff that never fails
Doubt so.
Can we make stuff that fails in , say, 20 years instead of 10?
Maaaayyyybeeeee.
If so, whatīs the cost involved?
If we make a transistor for $1 which fails in 10 years and a 20 year one costs $40, I guess general people will use the first one and NASA/Military will take the 20 years one.
Which one failed first?
In this case, insufficient data to know.
IF, say, a filter cap had shorted and that took out the rectifiers, well, we can find a sequence with reasonable certainty.
In the case of 2 identical transistors, subject to identical duty it becomes impossible.Juan Manuel Fahey
Comment
-
This is like asking which end of a rubber band broke. Why did my E string break now instead of yesterday or tomorrow?
It may be intellectually interesting, but really irrelevant to repair and use. A random component failure may have caused this, or some outside influence caused their failure. A shorted or intermittent load or speaker cord may have stressed this. A static electrical charge could have discharged into the speaker wires and killed one or both. Such an event could have weakened one or both months ago and it just now finally failed. You will never know unless you discover something else that is still wrong.
If both outputs are burnt out and being replaced, I HIGHLY recommend not being cheap, replace the drivers as well, that is Q9, Q10. You tested them as good, but you did it with a hand meter, I'd wager. That does not put the full 70 volts across them that this circuit does, plus it doesn;t draw any current through them to speak of. SO they function under zero stress, but who knows what they will do when working hard. For a couple dollars we will be safe.
By the way, with Q11, Q12 out, R55, R56 are no longer in a complete circuit, and thus cannot have any effect on the voltages you are concerned about.
There is a tendency to want to blame something, as if that accomplishes anything, but it doesn't. You have two transistors working together, the current flows through both. Sure, I imagine one failed before the other, so you could say one took the other down, but it doesn't matter which one went first, as it is common for outputs to fail in pairs when they fail. When one shorts the other is seriously stressed by that, and soon fails. Knowing which one changes nothing, we still have two dead ones, and it tells us nothing useful in terms of prevention.
There is another tendency to see any failure as a sign of a shortcoming. Often accompanied by the desire to use beefier parts and increase voltage and current ratings on parts. But an amp that was made by the tens of thousands, most of which have been running for decades, is not likely guilty of a weak design when it ultimately fails. Fix it right and it should run another 30 years.
One other cosmic possibility. Do you know what a soft error is in computers? A memory bit in RAM is a voltage level stored in a transistor. Cosmic rays are tiny atomic nuclei whizzing through space. They shower the earth all the time. They are zipping through you as you read this. Some of them go completely through the earth without hitting anything. But occasionally one does hit the atom in a molecule of something. if it happens to be an atom in a transistor in a RAM IC, it can actually change the logic 1 to a 0 or vice versa. That changes the memory. It didn;t hurt anything, but now memory is corrupt. REset and all is well, and no way to know how it happened. But sometimes a cosmic ray hit can physically damage something. Knocks and atom or two out of a semiconductor. So it is possible some errant cosmic ray roared through one of your transistors at some point and took a nick out of the silicon there. Then at some time, that was just enough to allow the part to croak. Sounds far fetched, but it could happen.
In the industry, unless we have a specific reason to point to a specific cause, "random component failure" is an accepted explanation. Essentially: now and then, parts fail.Education is what you're left with after you have forgotten what you have learned.
Comment
-
I totally hear what you are saying. My experience is quite limited and still learning everyday. To date, most of the failures I've seen, especially in regards to transistors, have clear source causes - the emitter resistor failed causing the transistor fail, A cap shorted causing the resistor to fail etc. etc. so pulling from that (limited) experience, I was searching for another (in this case non existent) root cause. The points you made are valid - the root cause is not always sourced inside the circuit, or outside the component itself. Thanks for that info.
BTW - does anyone know any good online tutorials for troubleshooting/debugging transistor amp circuits? I'm sure I would find that useful.
Comment
Comment