View Full Version : Question about Fault Finding with a signal & CRO


50mm
08-24-2011, 02:38 PM
Hi, I'm new to this and trying to bring a Kenwood KR-9400 back to life. (Progress thread (http://audiokarma.org/forums/showthread.php?t=387427)).

I have a nice scope (Tek 465), a function generator, an adjustable bench power supply, a couple of multimeters, and a dummy load.

Here's my question… It seems to me that it should be possible to run a signal into a board (like the amp board under test in the aforementioned thread), then, using the schematic and scope, trace the signal through the circuit to identify problem components. Right or wrong?

It seems that this would be a faster and easier method than having to desolder diode leads and pull transistors, etc. and meter them individually.

Any knowledgeable techs want to set me straight? Links to tutorials on using a signal & scope to find faults would be fantastic (believe me, I've looked - my electronics bookmarks have grown exponentially, but I'm not getting much relevance).

Thanks in advance.

BinaryMike
08-24-2011, 04:05 PM
Yes and no. Signal tracing with a scope can often zero in on the defective stage of an amplifier chain, but in multi-stage global feedback circuits like an SS power amp, the whole circuit is brought to its knees by a defect in any one stage. In that case, you're better off troubleshooting with DC voltage measurements and component testing or substitution.

DCinDC
08-24-2011, 04:56 PM
Although it's "fun" to use test equipment, don't get lost in it.

I have a little MP3 player I use as a source on the bench. Music is often better for finding common audio problems. A strong single tone from a FG can mask scratchy controls for example.

When I have no output, I look for obvious visual problems. Remember the FIRST step in ANY troubleshooting is a thorough visual inspection! Fuses fall in this step.

Then I check the power supplies unless it is obvious that they are working, like in a receiver that is tuning in a station, it is pretty likely that all power supplies are working. the separate power supplies for output sections are rarely regulated and usual work unless a fuse is blown.

Then I half split by getting in the pre/main loop. If there isn't one touch the volume control terminals with your finger. If you get a buzz, then everything past that prolly works.

Once you've had thousands in test equipment at your disposal, it is a challenge and fun to try and fix things with a 2.99 Harbor Freight multimeter. I do it all the time. Then when I fix it, I verify with the expensive toys.

50mm
08-25-2011, 07:46 AM
Thanks, this is good information and I think I've been intuitively doing some of this, at least when it comes to semi-working units.

I'm still interested in how to use a scope and signal, and maybe I need to clarify the question. Specifically, I'm asking about testing individual stage boards that have been pulled from a non-working unit.

Let me give you an example of what I'm doing now:

I have pulled both of the main amp boards pulled from a Kenwood KR-9400 - one working, one with issues. I went through and fully recapped both and made some upgrade mods from a service bulletin.

Now that I have two boards that should be identical I hook up a sine wave via alligator clips to the input of the boards and pull up traces along the circuit, checking the scope to be sure that they are identical.

I could've thrown them both back in and powered up to test, but I wanted some reassurance that I'd done the job before igniting more resistors or worse.

nashvillebill
08-25-2011, 09:04 AM
Given what you've done, I think that's a pretty logical approach as a double check for your work.

I use signal injection and tracing through an amp when I'm trying to resolve issues, but as BinaryMike says, there are circumstances where it's not clear where the issues are coming from. (Of course, I'm no expert so there's many times where I'm not clear where the issues are coming from! Tracing the signal through on my scope has been a great learning method for me)

Have you made a dim bulb tester? If not...do so before powering up the amp!! When I finally made mine, I wondered why I put off making one for so long.

50mm
08-25-2011, 09:20 AM
I did make a dim bulb tester and it has already saved this particular receiver - it's how I found out that two output transistors had shorted. :)

So, as you trace the signal on your scope, what in particular stands out as "wow, this doesn't look right"? In this case (two supposedly identical boards), it's pretty obvious, but what about a single board, e.g. if I were to pull the sketchy phono 1 input from my sx1250?

I'm learning as I go here, but I've only been doing this a short time. I've been putting together simple circuits on a breadboard, running a signal through it and seeing how that signal changes, but I've only done this with a few standard components and all of those components were working, so I have no reference for how the signal changes when those components go bad.

Also, I'd like to take this opportunity to thank AK once again - all of the books and sites I've put together don't approach the wisdom I've gotten here, whether from asking questions or just reading archived threads.

DCinDC
08-25-2011, 09:27 AM
You don't say, so I have to ask.

You have the boards out with a sig gen connected, do you have power connected to the boards?

50mm
08-25-2011, 09:58 AM
:) - I'm hoping I'm about to learn something.

No DCinDC, I don't, well, nothing outside of the signal coming in from the function generator itself.

I do have a variable bench supply and thought about hooking that up to the power pins on those amp boards, but didn't since I was getting (what I thought were) usable traces with just the signal.

//edit: silly DC.

nashvillebill
08-25-2011, 10:01 AM
You will need to power those boards to obtain usable results. Don't forget a dummy load.

Incidentally, it's AC coming from the sig gen.

50mm
08-25-2011, 10:11 AM
Incidentally, it's AC coming from the sig gen.

Sorry, of course, sine wave :) (was thinking about the DC offset knob on the fg). Ok, so let me see if I've got this…

1) Function generator to Input.
2) Bench supply to Power In.
3) Dummy load to output.
4) Rescope from there.

So the traces I was getting from the FG alone (which where certainly changing and seemed to be somewhat informative), were useless?

Jon_Logan
08-25-2011, 10:22 AM
I disagree that music is a better signal source in tracking down dirty switches, etc. A sine wave generator delivers a steady tone, when observed on a scope, shows intermittent connections clearly.

It is possible that too large of an input signal can make an intermittent connection connect, but that's why you'd pump in 50~100mV as your signal source. You can definitely see "scratchy" controls with a steady tone as the source.

I'd skip the bench supply. Unless you have a board yanked out of a set, just use the supply in the set. An external bench power supply is definitely a luxury. Even if you get one, what kind do you get? One capable of powering an output stage to several watts, or the typical one that's good for preamp type of power?

For signal tracing, you only need dummy loads if there's an output stage failure.

You said the FG output is "changing". Make sure you aren't using a Function generator that is set to some modulation mode (AM,FM, Sweep).

For audio, I'd recommend a low distortion oscillator instead of a function generator.

50mm
08-25-2011, 10:36 AM
Hi Jon - I think that what you are saying is close to what I'm doing, except that I do have the boards pulled and I'm not (yet) using the bench supply to power.

I'm sending a low distortion sine wave through the FG (no sweep) with high enough amplitude to see the sine as it moves through the circuit. What I meant by the traces changing is that I see it start as a sine, and watch it change on the scope as I probe through the circuit, as seen below:

https://lh6.googleusercontent.com/-y2kmvJUnTSk/TlXDYXsuokI/AAAAAAAAAls/3Djath3mE50/s1152/IMG_20110824_233615.jpg

50mm
08-25-2011, 10:41 AM
And for the record, I have a BK Precision 1735 power supply, continuously adjustable between 0-30V 0-3A (super cheap on craigslist).

Jon_Logan
08-25-2011, 10:45 AM
Yes, that scope waveform is no good.

So the generator and scope have gotten you to a faulty location. Time to start checking stuff with a voltmeter. If you're unsure about what voltages there should be, you can hopefully compare with the (hopefully) working channel. If both channels are hosed, look for something common to both (power supply, etc.) I've seen dirty switches produce waveforms that ugly.

50mm
08-25-2011, 10:56 AM
OK, good. I think I'm formulating a plan here.

I will try powering up without the amp boards and outputs and take a voltmeter to the test points outlined in the service manual. If all looks good there, I'll put the (hopefully) good board back in and check it.

Thanks for all the good info.

DCinDC
08-25-2011, 12:50 PM
That scope waveform is of course, no good. The boards are unpowered. It's self biasing.

Put the boards back, and use the dim bulb tester.

rBuckner
08-25-2011, 01:18 PM
And for the record, I have a BK Precision 1735 power supply, continuously adjustable between 0-30V 0-3A (super cheap on craigslist).

If that's the older non-A version it's a single supply it likely won't work to power the amp. Most amps run on a bipolar supply, a positive supply and a negative one. The new 1735A supply is dual and would work though.

If that scope picture is of the non-powered amp it doesn't mean a thing. I suggest you check the diodes and transistors with the diode function of a meter first as mentioned earlier. There's a thread on AK as to how to do that somewhere.

50mm
08-25-2011, 03:42 PM
All good info. Know what else is good?

I just heard the words "I'm Lester the Nightfly, hello Baton Rouge" come through a Baby Advent powered by a Kenwood KR-9400 that hasn't made a sound in god-knows how long. Right channel only, but hey. Plugged in the amp board, tested the dim bulb, plugged in the right output and dim-bulbed again, etc, being once bitten by this before… anyway, it ends with music.

But back to the main topic - with my current setup (non-A 1735), it is not possible to pull boards and test them out of their powered boxes?

rBuckner
08-26-2011, 12:17 AM
All good info. Know what else is good?

I just heard the words "I'm Lester the Nightfly, hello Baton Rouge" come through a Baby Advent powered by a Kenwood KR-9400 that hasn't made a sound in god-knows how long. Right channel only, but hey. Plugged in the amp board, tested the dim bulb, plugged in the right output and dim-bulbed again, etc, being once bitten by this before… anyway, it ends with music.

But back to the main topic - with my current setup (non-A 1735), it is not possible to pull boards and test them out of their powered boxes?

Glad some progress is made!

My premise is that probably it can't be tested based on the dual supply rails most common in gear. I'm too lazy to find a schematic to verify this though!

DCinDC
08-26-2011, 08:33 AM
You could mock up a test fixture with outboard power supplies and run it out of the chassis, but why?

Glad you got it going. So the right works what about the left?

50mm
08-27-2011, 09:26 PM
Why... Not sure really. Again, very new to this, but I was thinking it would be easier to isolate the issues by testing a board outside of the rest of the system.

The right channel is waiting on some resistors from mouser.