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

March 28, 2024, 03:17:07 PM

Login with username, password and session length

Recent Posts

 

Vexing Acoustic 230 repair, output gets weaker over time

Started by nsubulysses, August 17, 2016, 04:17:31 PM

Previous topic - Next topic

nsubulysses

Ok I have hit the wall and need assistance from the real pros   :dbtu:

This Acoustic 230 has weaker and weaker output over time.

When you turn the amp on output looks roughly like this


After about 10 minutes it will be something like this


It seems to be some sort of thermal drift problem because when the output is weak I can literally blow it back up by blowing on the PCB with output section on it

Q8 and Q10 have been replaced because I thought they were not able to drive the output transistors. Still the same. With new output transistors the result is the same too.  I have clipped caps in parallel over every cap in the output section and the result does not vary.

Here's a video of the output. It bounces all over and is not stable, let alone losing power over time. See how when I turn it up it has some sort of settling effect as it deals with trying to put out a big signal?
https://www.youtube.com/watch?v=Jq_UwCRaG6w&feature=youtu.be





Enzo

In the vid, are you turning it up and down?  Or is the signal steady and the amp is cycling off and on?

In the photos, the lower one looks like the top half of the waveform is missing.

SO start at the start.  Your power rail is about 80-85vDC, right?  Does it stay up there or does it slowly drop?  Likewise the output bus sits at half that, so 40-45vDC.  I am talking the output bus on the amp side of the big speaker cap.  Does that stay up at that 40 some volts or does it drift off.

If you were to run the amp without signal for say 10 minutes or something and THEN apply signal, is the output already collapsed?  Or does it start good and drift down from there?  Or does the signal start good and then collapse no matter when you start?  I am loking to see if there is anything we can monitor without bothering to apply signal.

Something clearly changes, so once we know if power rails stays or not, and that the amp balances at half rail or not, we move on to looking for what does change.  The schematic has a number of DC voltages on it, how do yours compare.  I don;t care if they say 43 and you have 39 or 46, what I care about is if you start at 43 and drift down to 23 or up to 63, that sort of thing.

Let us see what we can eliminate. have you tried input the signal at the power amp ion jack?  And with a normal input out front, have you monitored the preamp out to see that the signal remains strong there?  I am assuming a power amp problem because you led me there, but are we sure it is in the power amp?  For now that is what I discuss.

Input at power amp in jack, now monitor the signal right at the point above Q2, at the left end of C6.  Does the signal remain stable there or does it collapse even there?  My concern is that maybe Q1 and/or Q2 is a problem.  You could remove them during test, and if that helps, then there we are.

Oh, you said blowing on it makes a difference.  Great.  Do you have a can of freeze spray?  Or even a can of computer blow off aerosol dust remover?  That is what freeze spray is for, to isolate and identify sensitive parts.  Even just blowing, try using a straw to direct it to smaller areas.  Also a 3x5 card or piece of cardboard can be held down against the board to shield parts from the cooling breeze.  I have even rolled a piece of paper or card into a cylinder and slipped it over a component to cool only that one part.  This allows us to more selectively cool stuff to identify what is sensitive.

You replaced Q8,10, but did you check the 33 ohm resistors?   Does the voltage drop across  the two ballast resistors, R29,35, change or is it stable?  Q7,9 are limiters, disable them by lifting D5, D6.

nsubulysses

#2
Hi Enzo thanks for your reply. I will try going through some of this and report back.

In the video I am turning the volume amp up and down to show how it responds

I don't have freeze spray but a straw is a great idea. Let me see what I can find.

Power rails seem to sag. the main cap has 85V but when you turn up volume it goes down to about 76V, then settles about 79V with signal still applied. isolating cap at the speaker does the same, about 40V at idle, dropping to about 35 when amp is cranked then settling around 38 with signal applied.

Using the power amp in jack I get the same signal after the amp has been on for a bit, about 10 minutes.

When I just turn the amp on the power amp in jack produces a very symmetrical and powerful signal that gets weaker over time.  The same thing happens when I use the regular input jack, but the output signal has much more variance, basically modulating shape as it's turned up and then settling in some sort of stable wave after about a second or two.  If this is hard to understand I can make a quick vid.

The amp performs good when it's first turned on. If you turn it on and apply no signal for about 10 minutes it will already have a weak output when you apply the signal.

Enzo

OK, so it is not responding to signal, it is just going out of kilter on its own.  The signal is just how we notice it.  SO I an back to looking at DC voltages etc.  SO Q1,2 and stuff probably OK.  Check rest of those things.

J M Fahey

1) it´s clearly thermal.

2) please repeat but do NOT turn it on/off repeatedly , it´s maddening.  :duh

3) the amp does have an input mute, which also acts like a very crude limiter/compressor, disable it while testing by desoldering and lifting Q1 and Q2 Drain legs and bending them up outside the holes.
I mean the legs touching each end of R2.

4) repeat and let us see the slow signal deterioration in just one take.

5) if you don´t have freeze spray and blowing is too wide and inconclusive, get a glass of cold water, let an ice cube float on it, get a QTip, wet it in cold water, shake it and touch Q3 body.
If no change, cool it again in the glass and now touch Q4 body ... and so on until you find the culprit.

6) there *might* be a leaky electrolytic or something messing with voltages, but let´s walk the thermal road first.

nsubulysses

#5
Ok the wizards have helped sort the problem once again and I am truly appreciative. Q5 has a thermal issue. I used an ice pack and a Q-tip. I assume I am homeward bound now. I will replace RCA 40408 with 2N5320 tomorrow and let it run for a while to make sure, but I bet it will maintain full power.

I really do appreciate it. Thank you

https://www.youtube.com/watch?v=ajAP4Ff1HrI&feature=youtu.be

J M Fahey

Try to get a small light TO39 "star"  heatsink


I insist on light because a heavy one will quickly crack legs with vibrtion.

I´ve seen some crazy ones, how can they expect legs will survive?   :duh :duh :duh
barely acceptable:


WTF?


You-must-be-kidding:



The Southern Way (tm) (meaning Argentina/SA/Australia/NZ):

:lmao:

nsubulysses

Hmmm, well with original transistor being 90v 700mA and 2N5320 being 75V 2A wouldnt this be pretty over spec'd as is?  Transistor in circuit only has 40V on it so no worries there.

I am running the amp now with 2N5320 now. So good so far

Enzo

You also have to be concerned with dissipation.  The current rating and wattage rating are dependent upon temperature.  When the temperature of the part is allowed to rise above the ambient 25C or so, then it is rated lower.  The small heat sink helps keep the temperature down.  Look up "safe operating area", or SOA as it is on the data sheet for most power semiconductors.  This part may be too small to rate an SOA graph.  But it still has an SOA.

The 2N5320 is a good part, I stocked it and its complement (2N5322 maybe?) for years, but it is not good for both 75v and 2A at the same time.

By the way, the transistor has 40v across it NOW.  sitting at idle, but look at the circuit.  The emitter is tied to ground.  When the base is fully driven, its collector comes close to ground.  But what about when the base is turned off?  Now we have pretty much the entire 80-90v of the power supply across it.  SO no, the 75v part is not overspec'd.  During large signal situations it cycles up to close to full rail voltage across the part.

If you have no little heat sink, I might help.  If you decide to pick up my 250w Dales, remind me and I will throw a couple heat sinks in the box.

nsubulysses

Of course I believe you but I must be misunderstanding something. Monitoring collector voltage of Q5 with an analog meter I only can get it to vary from about 35-50VDC when I turn the amp up and down between full volume and no volume.

2N5320 is 10W power dissipation and RCA4048 is 1W, so it seems ok. Has put out full power without a hassle since my last post at 4:05PM, so about 2hrs and 15 min of full power.

I still want to buy the Dale dummy load resistors thanks Enzo. Getting 3 parts orders this week and trying to finish many repairs. Should have some $$$ for you soon :)

Enzo

Maybe the heftier part is Ok running naked, but I still think it is a good idea to sink it.  The wattage ratings are all based upon not letting the temp rise above X.

You are measuring DC voltage?  or AC?  If DC voltage, your meter will average whatever AC is on it.  So if it sits at 40vDC and has 25v of signal , the average voltage there is still 40vDC.

If you are measuring 40vAC, That is RMS, meaning the peak is 56v.  And in fact it also goes the other way, so 40vAC is 112v peak to peak.  In any case, when that transistor is turned off, the collector voltage rises to B+.  And that is what is then across the part.  80v.  In my 40v example we'd be clipping, but ignoring that, if my reference point was +40vDC, then the signal peak would add 56v to that for 96v.  We can't go that high, so we'd clip before we got there, but since I made up the 40v, we could do the same exercise at 25v.

25vAC RMS is 35v peak, so 40v plus 35v is 75v.  And that would be across the part.


I think I still have a few 40408 in my Acoustic drawer.  Along with some of their house numbered TO3s.

J M Fahey

Quote from: nsubulysses on August 22, 2016, 04:05:13 PM
Hmmm, well with original transistor being 90v 700mA and 2N5320 being 75V 2A wouldnt this be pretty over spec'd as is?  Transistor in circuit only has 40V on it so no worries there.

I am running the amp now with 2N5320 now. So good so far
Sorry but it does not work that way  :o

Read the datasheet properly:

they say transistors can stand *up to* 75 or 90V or *up to* 700mA or 2A but NOT max voltage AND max current at the same time  :loco
If so one would be able to dissipate .7x90=63W and the other 75x2=150W which CLEARLY is not the case.

Even datasheet dissipation, which I guess is a couple Watts at best, is a "textbook spec" because if you read the full spec it states: "at 25C case temperature"  which can only be kept if transistor is bolted to a huge aluminum cube kept a few degrees *under*  25C .

The realistic spec , which is also in the datasheet, is "free air dissipation" which means the unaided transistor mounted on a PCB and which probably amounts to 1W or less ... and even so you shouldn´t reach even half that, to protect it against internal chassis temperatures higher than 25C .

So I bet the transistor was dissipating a few hundred mW which put it dangerously close to max temp, so it entered thermal runaway.

Try to add a small heatsink, worst case improvise one yourself out of a small strip of aluminum, bend it like this:


or even clip a fuse holder clip around it:


nsubulysses

You guys are great thanks for explaining further. Really need to sharpen my skills in SS realm. Slowly getting there.

Enzo I'll be putting in a parts order in the next few days. Heat sinks, dummy load resistors, 40408s, the whole deal. Thanks.



gbono

Don't forget to derate the transistor depending on power dissipated  (IcxVcc in watts) AND the ambient temperature - the chassis is rarely at room temperature for long (assume room temperature =Ta=25C = 77F). The transistor date sheet will usually give a thermal "resistance" based on case to ambient, junction to case, etc. Units are degrees/watt or something similar.The goal is to calculate the junction temperature of the transistor.

Once you calculate the junction temperature of the transistor using the Ta, Pd and appropriate thermal resistance(s) you can add a heatsink to get an even lower junction temperature. You are doing the same thing when you rebias a tube/valve based on plate voltage and cathode to plate current - note that tubes run at 250C while a transistor should not exceed a junction temperature of 125C - better to keep Tj at 80% of 125C.

You can guesstimate what's going on with your substituted part buy measuring the case temperature of the part with a thermocouple or (worse) IR scanner.