Diagnosing Weak Channel Problems

[thebes]

I considered putting this in the "Test Bench"  section of the tube forum, but since problems with tube amps on only one channel are very common, I thought I'd post a separate thread.

 

It seems to me that one channel being completly out is usually easier to fix figure out than a weak channel.  On this unit, a Marantz8b (which I picked up for only $50-yeah for me) has a weak right channel.

 

I've done the basic troubleshooting, swapping rca's speaker wire, power tubes, signal tubes etc. and hooked up to only a cd player (with volume control) and my two test speakers there is no question that the right channel is experiencing a decided lack of enthusiasm in carrying out it's duties. 

 

This unit has been modified. All caps have been replaced with Russian Pio's and the power supply has been doubled.  So values are different than the typical  8b schematic. 

 

The question that arises in my mind is where do you start, at the input end or the output end, at the power supply or someplace else?  I do believe there is a system used in diagnosing this type of trouble but I have no idea what it is. So before I stick a pair of probes in a box with several hundreds volts inside where a misstep can wake you up faster than the bottom dregs of a pot of coffee that's been brewing for a week, I though I'd ask you guys.  Not only for me, but I think the lessons that can be learned here will be informative for other hobbyists lacking a death wish.

[xxJPMxx]

Do you have a scope? You seem to know how to use a meter so check operating voltages of bad channel against the good channel. I would rule out power supply because both channels would be effected.

 

Possibly bad cathode bypass cap for the 6BH6 reducing gain.

Edited October 13, 2015 by xxJPMxx

[WMcD]

Well, I agree with the scope, and test signal of some sort.  A CD player with a test CD track of 1 kHz played in a repeat mode will do.

 

Then using a capacitively coupled probe, check for the signal starting at input thought output.  Of course a schematic and your formidable powers of deduction are necessary.   It is probably best to check the level at the various junctions in the known good channel versus the bad channel.  You don't have to know a "good" level.

 

Similarly, you might check the d.c. plate and bias voltage of the known good channel versus the bad channel. 

 

Play with high voltage with much caution.  The old rule was to keep your left hand in you pocket so that you don't get zapped through your heart.

 

WMcD

[Wolfbane]

Well, I agree with the scope, and test signal of some sort.  A CD player with a test CD track of 1 kHz played in a repeat mode will do.

 

Then using a capacitively coupled probe, check for the signal starting at input thought output.  Of course a schematic and your formidable powers of deduction are necessary.   It is probably best to check the level at the various junctions in the known good channel versus the bad channel.  You don't have to know a "good" level.

 

Similarly, you might check the d.c. plate and bias voltage of the known good channel versus the bad channel. 

 

Play with high voltage with much caution.  The old rule was to keep your left hand in you pocket so that you don't get zapped through your heart.

 

WMcD

 

Sound advice even when changing a light bulb.

[thebes]

I would rule out power supply because both channels would be effected.

check for the signal starting at input thought output

 

Ah, now we are starting to reveal a methodology to diagnosing the problem.  A great clue is that with one channel only affected, the power supply , which supplies both channel is not suspect.  One major part of the circuit has been eliminated as a possible cause of the problem.

 

Gil, is there some reason to go with the input signal first?

 

Again, while I want to try and solve my own specific problem, since this is a relatively common occurrence, I'm also trying to develop guidance for future readers who have the same type of problem.

 

Now, although I have a scope, I am far from competent in it's operation. That aside, most people dabbling in tube amps probably don't have one either but probably have a multi-meter so that would be the proper diagnostic tool.

[babadono]

Thebes,

 There are certain tasks which one must have a scope to troubleshoot. Analog AC waveform signal tracing is one of them very near the top of the list. Sure you might get lucky and fix it with just a meter, heck you might just shake a loose connection on one of the tube sockets and fix it.

 For noise and hum issues it is best to work from the output back towards the input. But for no signal or weak signal I prefer the "halfway" approach. Apply an input signal and there is no or little output. Pick the spot in your estimation (from schematics hopefully) that is "halfway" between input and output. Is there signal there?

1) No, then you know it is closer to the input so go "halfway" again towards the input.

Or 2)Yes there was signal "halfway" you know the problem is closer to the output so go "halfway" again towards the output.

etc......etc.......etc.......

[xxJPMxx]

I would download a free signal generator. Not sure what OS you have but I run linux and I have Siggen. I am sure there are free ones for Windows users. If not then you could probably find a 60Hz test tone on youtube.  Why 60Hz? Because at that frequency your multimeter will have the best resolution. Or use whatever frequency your heart desires.

 

 

Always start at the input that way there you can see where it is getting lost or attenuated. Since you do have one channel working you don't even have to calculate each stages gain just compare it to the working channel. You want to compare input voltage to AC voltage at the plate (pin 3 I beleive) through a capacitor, make sure the capacitor is 400v rated, any small value cap should be okay because the input impedance of most meters are pretty high.

 

Have you measured DC operating points yet?

[WMcD]

Let me first disclaim that I'm experienced with this. But I had experience as a youth with Heathkit amateur radio receivers. The troubleshooting manuals were excellent.

With the receivers, tuning and thus troubleshooting started at the front end of the receiver (the antenna input) because farther down the line, you're relying on having a good input from the previous stage to peak the performance of the next stage.

But I think it is a logical way for you to do things too. You can say, "Yup, it is good at the input." Then move along to the next stage with some confidence it is getting a good signal.

I'm not sure what is going on with you. It seems that you got this used at a discount after it had be modified by someone. Maybe not done to well and that is why the seller got rid of it?. So the problem might be a botched update?

I looked at the service manual and schematic. I see that after the first tube as sort of a pre-amp, there is a second tube which is used as a phase splitter to give a "plus" audio signal and a "minus: audio signal to drive the split in amplifiers to work the push-pull design. Therefore, my experience with receivers is simplistic and difficult to implement.

You do have the service manual . . . Yahh? http://www.bmamps.com/Schematics/Marantz/Marantz_8B_Service_Manual.pdf

It does have a listing of what the d.c. voltages should be at each pin of each tube (input shorted). You should check those first (edit, after bias described below). Easy with a Vacuum Tube Volt Meter but a modern transistor multimeter should work too.

There is also a table of simple resistance to ground for each pin. Of course you must make sure all power is removed and the filter capacitors aredischarged. If I'm reading the schematic correctly the power supply caps don't have bleeder resistors.

Also, the manual has instructions for setting bias, which should probably be done first. Did you say you've done that?

Just a guess: If there was a defect in the push-pull stages, you might get distortion rather than low level. So I'd suspect the first two tube stages.

Others here may be more helpful.

WMcD

Go Cubs!

Edited October 14, 2015 by William F. Gil McDermott

[tube fanatic]

Marty, pull out the copy of Practical Radio Servicing which I recall you bought a couple of years ago and follow the section on "Servicing the audio amplifier."  This problem should be very easy to find using voltage analysis (no 'scope needed), and/or a simple signal "generator" such as a CD player, tuner, or any other device with a couple of hundred or more millivolts output.  An audio amplifier is an audio amplifier whether it's in a radio, tv, or any other piece of equipment.  Also, in this case, since you have a working channel for reference I'd work backwards from the speaker terminals.  Doing so should allow a very rapid location of the problem area.

 

Maynard

[WMcD]

Good Grief.  I woke up this morning and recognized that I may have insulted Thebes by alluding to a botched update.

 

Looking back I see Thebes made a use of the passive case by saying the caps have been replaced. My initial thought was that the replacement was made by the previous owner.  But it could also mean that Thebes did the work, and I inadvertently insulted him. Sorry. 

 

I thought a bit about checking for signal starting at the front and working from the back.  Let me say based on vague epistemology the best way is to start at the front end and work to the back.  Here is why.  It is based, in part, on the notion that there may be more than one stage at fault, and it is best to gain knowledge with certainty as quickly as possible.

 

If you start at the back end (stage 5), you may find the signal is not getting through to there so the problem could be there, or also in one or more of the stages toward the front.  You can go to the next earlier stage (stage 4) and find no signal.  You are still not really sure either of those two stages are working and don't know whether there is a problem in stage 1, 2 or 3.  Now eventually you may to stage 2 and find it does have a signal and is working.  You hope that fixing stage 3 will fix the entire problem.  But if it doesn't, you have to go back to stage 4 to check that.  And maybe you have to check stage 5 again.

 

It is more efficient, arguably, to start at stage 1.  You will find that it is working and thus cross that off the list of problems as a known good .  You've obtained right away knowledge of a working stage.  Therefore, you've obtained more knowledge more efficiently. Then if there is no output, you move to stage 2.

 

So, I dunno.  It seems like I'm on to something really brilliant.  But maybe not.  Maybe I'm too stupid to realize it is sophomoric.  Smile.

 

WMcD

Edited October 15, 2015 by William F. Gil McDermott

[tube fanatic]

Gil, as a former service shop owner in which diagnosis of problems in a timely fashion was critical, I have to disagree with you.  In an amp like the 8B there are many places between the input and output in which signal can be attenuated resulting in weak output from one channel (remember also that this amp has had work done to it so there's the possibility that the issue has more than one cause).  Marty has already ruled out a tube issue by swapping the good tubes to the weak side with no improvement.  By starting at the input, the voltages and gain of each stage has to be checked and verified against the "good" channel.  This is too time consuming.  On the other hand, by reversing this and working backwards we are testing in a more efficient manner.  Say Marty injects a signal into the grids of the output tubes and discovers that the signal output of the weak channel is identical to that of the normal channel (this can be done by putting a load resistor of appropriate wattage across the speaker terminals and measuring the AC voltage appearing across it with the multimeter, or using a test speaker and simply listening).  Now we've ruled out the output stage and output xfmrs as the cause of the problem.  So, now we move back to the phase inverter and do the same thing.  If we now find that the weak channel is still equal to that of the normal channel, we have ruled out that stage as the cause and know that the problem is between the input and phase inverter.  On the other hand, if we now find that the weak channel remains weak, we know that the problem is between the phase inverter and output stage.  Simple voltage and resistance checks, and checking for cold solder joints,  should allow quick isolation of the problem.  When that's repaired we move back to the front end and when we inject the signal into the grid of that tube say we find a big difference between the 2 channels once again.  That problem is now isolated to being between the front end tube and the phase inverter.  Once that is known some simple voltage/resistance checks and a check of solder connections for cold joints should isolate that problem quickly as well.  Once that is repaired, the signal can be applied to the input to verify proper operation.  This is a much faster methodology for resolving issues such as a weak or dead channel. 

 

Maynard

[xxJPMxx]

I do both depending on what the symptoms are. If I think it's in the power amp section I start there, if I think it's in the front end I start there. I have two issues with Maynards methodology and that's one; how is the OP going to get two equal signals 180 degrees out of phase to drive the power tubes? Second; How is the OP going to get enough drive voltage to the output tube grids? You will need something like 100V peak to peak. So for a beginners sake I think it is much more practical to use a line level source like a computer or CD player to drive the input and then see where the signal gets lost. I still have my money on the first stages cathode bypass cap

Edited October 15, 2015 by xxJPMxx

[tube fanatic]

I do both depending on what the symptoms are. If I think it's in the power amp section I start there, if I think it's in the front end I start there. I have two issues with Maynards methodology and that's one; how is the OP going to get two equal signals 180 degrees out of phase to drive the power tubes? Second; How is the OP going to get enough drive voltage to the output tube grids? You will need something like 100V peak to peak. So for a beginners sake I think it is much more practical to use a line level source like a computer or CD player to drive the input and then see where the signal gets lost. I still have my money on the first stages cathode bypass cap

Hi xx and welcome to the forum.  An extremely knowledgeable member ("thesloth"), who made many fabulous contributions to this section and was instrumental in its success, is no longer participating so please hang around long term!!!  We need as many knowledgeable tube folks as possible in the interest of promoting the use of that type of equipment with Klipsch speakers.  Of course, an unspoken rule is that newcomers should never argue with "seasoned" veterans (just joking, of course ).

With the method I suggested it's not necessary to provide much drive.  All we're looking for is a similarity between the 2 channels.  So, when testing the output stage, if we're measuring the AC voltage across the resistive dummy load, and find say 1V for the good channel, but only 200 millivolts for the weak channel, we know that there's an output stage issue.  We're going to perform that evaluation with both sections of the phase inverter which will tell us what each output tube and half of the xfmr is doing.  But, I concede that Marty may find it easier to use the method suggested by Gil and yourself.  The bottom line is finding the problem and fixing it.  Either way the amp will then provide excellent sound.

Again, welcome, and participate often!

Maynard

[xxJPMxx]

Thanks for the kind greetings. It's actually funny, I have been a member on here for years but never participated. I noticed there is this tube section which really interests me so I plan to participate more now. I am not sure if I can fill anyones shoes but I will share what I do know. I shall never question a seasoned member such as yourself ever again

 

Maybe we will get an update on thebes progress this weekend. For all we know he's got it fixed and jammin out as we type LOL!

[WMcD]

My start at the front end arose a bit from the experience I described with receivers.  I never thought that Thebes would be injecting a signal into a given stage. 

 

Of course I defer to the experts here.  Interesting reading.

 

Now, where the Dickens is Thebes to give us an update?

 

WMcD

[thebes]

Sorry guys.  I should have never started this thread with company staying at my house.  Between my hosting demands and work I haven't done a thing.  Hopefully I should get a chance tomorrow to start the testing. I've also pulled the electronics book and will be reading  that tonight. 

 

Again, I do envision this post of use to others, so I'm in no rush since I'll have to do a proper write-up.  Plus I don't want to wreck a perfectly good $50 amplifier by hurrying.

 

After all I want to be able to say with perfect honesty that no magic smoke escaped during the course of this thread.  Plus I've been told that if I bring down the mid-Atlantic power grid one more time I'm in serious trouble.

[mike stehr]

It seems you have a schematic and a multi-meter. You got dummy load resistors? If not, use junk speakers for a load. Ground the ground probe on a good ground point, (first cap ground will work) set your meter for high voltage DC, (I usually use the highest setting) and power up the amplifier.

Now you have the schematic. On that schematic, there should be several voltage indications on certain areas, (plates, cathodes, caps, etc..) check these voltage points and see if they coincide with the schematic.

 

If there are points not referenced on the schematic, grab a pencil and paper and write down the voltages. For instance, side one: plate voltage, cathode voltage, grid voltage, screen voltage. (for EL34 output tubes) Same with side two. Check the grid voltage at the output tubes, to make sure there is no DC leaking from coupling/blocking cap...it shouldn't be more than a volt. 

 

When probing around checking voltages on the powered up amplifier, use only one hand. You want the ground probe fixed to the ground point with a alligator clip, so you don't have to touch it. Then just use one hand to check the voltages with the hot lead of your multi-meter.

 

A couple things you should do before even powering it up. Flip it over, put on your best cheaters and look around very carefully for any sort of little bits of wire or a piece of solder that may be wedged against a terminal point, or a socket pin or some such.

Check the resistances between each output transformer. Put you meter on resistance of course, (you may have put it on the lower settings) and measure and write down the resistance of the primaries of each OPT (with ground probe on common of the OPT) and the same with the secondary taps.

 

This is to check if possibly a output transformer may have a problem, by comparing resistance between each other. They won't exact, but should be close...not off by a mile.

You may have to unsolder a place or two to get proper measurements, but if you have the tubes pulled, you should be okay. Make sure the capacitors are drained first!

 

Have fun, take your time and be careful.

[WMcD]

Plus I've been told that if I bring down the mid-Atlantic power grid one more time I'm in serious trouble.

 

That's funny, Marty.

 

WMcD

[thebes]

By the way Gil, you weren't insulting me at all. I have done about every screwup possible in repairing or building tube electronics and I am still very  much a not-so-talented rookie. Craig did the initial recap of the signal caps, and under his supervision I did the power supply recap and upgrade.  This was over a year ago and the 8b had played well until recently.

 

Well this damn thing has fixed itself.  Started the initial testing using a test signal through speakers and everything sounded normal. Did some basic measurements and both sides are virtually the same.  Put it back into the system it sounds fine.  Can only figure, like others have mentioned, that I dislodged something when doing  my physical  inspection, since all the solder joints looked fine.  Only time will tell.

 

 

One caveat.  This unit is very quiet as a line stage, but very noisy with phone using a LOMC. It also has a "whooshing" sound every few minutes which can often be heard over the music.  So I'm wondering if there is any way to quiet this down further and would using a scope be the way to track down noise, and if so, how do I do that.

Edited October 26, 2015 by thebes

[JJkizak]

One thing about electronics is you never 100% rule out anything, not even the full moon.

JJK