• Please note that Audiokarma will be offline briefly for updates early on Monday morning, September 28th.

MC7300 One Channel On Fritz

Brief testing tonight. Good news / bad news. Good news, the MC7300 is fine (or at least, as fine as it was before last night's testing). Bad news, my Tek's Channel 2 is not OK.

As soon as I hook up Channel 2's probe & ground (either probe) across either channel's dummy load, it causes the respective channel of the MC7300 to go haywire like I showed last night. (Seems like maybe a short somewhere in the channel?)

If I use Channel 1, either probe, on either channel's dummy load, no hot mess, everything's fine.

Glad about the MC7300, sad about the scope. Maybe I need to shop for a new scope to use while I try to fix my 2236's Channel 2.
 
Last edited:
I can't fathom a reason why the scope itself would be an issue if you have the probes set to 10X. The probe tip should have a 10 megohm resistor in series with the input and should be adequate to provide a low-capacitance point suitable for probing RF circuits, let alone audio. When it does this, do you also have the Ch1 probe hooked up?

I'm thinking you have a ground loop problem of some sort. THe scope undoubtedly has a three-wire plug - not sure if the 7300 does also. You really only need one ground wire from the probe, try it with just the Ch1 ground probe connected. Might be some kind of capacitive coupling groing on with the wires you have on your dummy loads between your input device/wiring, scope/wiring, etc causing an oscillation - but then there wasn't an oscillation apparent on your screenshot when that was happening, was there?

All that said, this could be pointing towards an issue with the amp. The amp *should* be able to drive almost any sort of load at low levels without groaning and heating up like it is, even if the scope is somehow screwed up. I further note that the meters' power reading seems to be quite a bit higher than what we're seeing at the output. That could be because the meters are not correctly adjusted - OR it could be indicative of a phasing issue with the autoformers, feedback network, or something else. Note that the 7300 has a metering circuit that actually measures power - it has a current transformer which sends a voltage to the meter amp correlating to audio current draw through the autoformer, AND a audio voltage measurement point. You've had those boards disconnected as I recall - are the plugs that go to the autoformers in correctly?

Finally I note all kinds of intermittent issues you've had. I note a subtle but potentially important difference between the right and left channel autoformer wiring: the left channel COM lead is grounded to the chassis whereas the right channel autoformer is NOT grounded to the chassis. The only other chassis ground point is the input connectors and the bypass capacitor on the AC input. Note differences between right and left on the capacitor board - red and green circled points.

I further note discrepancies between the schematic, the chassis wiring diagram, and the capacitor board PCB layout with respect to the numbering of the four SP3 leads, going to the autoformer. There's obviously some stuff going on on the capacitor PCB that's not exactly clear on its diagram - probably flying leads. Somehow the COM lead from the right autoformer's gotta get to the ground point, and that's not shown on the PCB diagram, but it is shown on the schematic.

Now... what to do with this? I'd take a resistance reading between each autoformer COM to the chassis. Should be a dead short. Then go up each tap and read resistance to ground - should be very low but increasing as you go up the taps. Is there a discrepancy between the two channels?

The other thing I'd do is double check that each J3 on the capacitor PCB is oriented correctly, and not intermittent. THe respective J3s provide the chassis ground to the driver, as well as feedback to the driver. Note that in the capacitor PCB diagram (below, green circle), J3-1R doesn't have a connection to that pin - but it should I would expect have a flying lead to it - OR they're relying upon J3R for the ground. Anyhow, I hope this isn't a red herring but the behavior of the amp when connected a certian way with the scope is indication that something's not right with the amp, and a grounding issue associated with the output (thru the scope and/or power line) might be the problem here.
upload_2023-1-19_9-12-2.png
 
Thanks so much. What an amazing response and great help.

Tonight I tested the various combinations of scope hookups to try to re-confirm my conclusion above. I disproved it. (As you suggested.)

EDIT: I am an idiot. I had one of the probes reversed on the dummy load. Corrected it and we're back to normal.

Onwards ...
 
Last edited:
[EDIT: As noted above, my ground issue was from reversing the probe / ground on one of the dummy loads. I'll leave the below post intact, I now understand the ground issue was my fault from hooking up the probe...]

So, turning to the autoformer grounds and PCB board, first of all I immediately noticed evidence that looks like someone was in here before. Look at the scratches are the left channel's SP3L/4L/5L:

PXL_20230121_043522385.jpg

With that noted but nothing disconnected or broken as far as I could see, I measured the autoformer resistances to chassis ground ... using my Fluke 101 which I'm really starting to dislike because it's verrrryyy sllowww to find the actual resistance. The value starts high and sllllowwllllyyy falls until it stops falling. That's the number I will note below. Also, the leads that came with it have ~0.5ohms of internal resistance, so the numbers I actually measured are 0.5ohms higher than this:

Resistance to chassis ground in ohms:

upload_2023-1-20_21-23-58.png

I will check the J3 connectors next (measuring to chassis ground). The harness connections seem secure.

Out of pure curiosity (cough), what do you think would happen if the unit was powered on with the J3 connectors unplugged?
 
Last edited:
I wouldn't run it with J3 disconnected- the feedback and common point at the emitters of Q13 and Q14 on driver would be floating, and no ground connection between the output and the driver. Results could be spectacularly unpredictable.

Your meter's behavior is strange, I've only seen a Fluke do that if there's a voltage potential there. Turn the amp on, and measure voltage (try both AC and DC) to ground.... should be zero. If you short the meter's leads together (ignoring the amp for the moment) does it settle down immediately? It might be trying to autorange, if you have a Fluke with a button in the center of the range knob that will lock the range.

Finally, someone definitely has been in there trying to repair something associated with those pins. Could be they were tracing the same intermittent you've been tracing. That leftmost pins in particular look questionable. I'd pull J3 and see what the condition of the soldering is on its pins.... resolder everything that looks questionable and put a bit of tension on the pins while testing with an ohmmeter to ensure everything is well secured.
 
I turned the amp on, min gain on both channels (dummy loads still connected between C and 8ohm terminals, but scope off), no input signal (but RCA cable plugged in and not connected on 3.5mm side), and measured DC/AC voltage from each tap to chassis ground.

upload_2023-1-24_9-50-8.png

The multimeter ... here's a video of what happens when I short the leads together, measuring resistance.

I'll note that I had to hold the phone to record in one hand, so I smashed the red lead down with my other hand holding the black lead, and it got to a lower resistance more quickly. Normally when I'm just holding both leads in my hands, it's much much slower and doesn't get below 0.5ohms, and takes much longer to settle there ...

 
Last edited:
Those measurements look reasonable. No junk either AC nor DC.

Apart from the ambitious power meter readings, how's it working? Seems like the intermittents have settled down a bit.
 
Peak to peak is different from peak. Peak to peak is 2*peak, so you gotta divide by two again.

You were onto something ...

I hate to admit this one. I realized last night that my dummy loads are 4ohm, not 8ohm. I confirmed this by measuring the resistance across the dummy loads + speaker wire + banana plugs. My stupid multimeter eventually settled on ~4.0ohms for each.

This is one of the most embarrassing moments of my audio-tinkering journey. LOL!

So. I hooked up the dummy loads across the 4ohm taps, and ran the same 1kHz test tone through it from my phone.

Here's what I saw:

PXL_20230125_175951397.jpg

Meters about halfway between 0.3W and 3.0W. The scale is logarithmic so the geometric mean (VGMab = Sqrt (A * B)) between 0.3W and 3.0W is 0.95W.

On the scope (both channels on top of one another):

PXL_20230125_182630614.jpg

V peak to peak is almost exactly the full 8 divisions, at 0.5V/div = 4V peak to peak.

Where Vpp = Vrms * 2*Sqrt(2), Vrms = 1.41V.

Where V = I*R, I = 1.41V / 4.0ohm = 0.35A.

Where Power = V * I, Power = 1.41V * 0.35A = 0.5W.

Is my math right? If so, my meters are optimistic by a factor of 2. :(

... in parallel, here's a short video of the right meter twitching at idle:


... and finally, I note that the rightmost face bulb randomly decided to blow, so I'll add that to the shopping list (in addition to two of the right meter's backlight bulbs). I should rename my workbench "The Entropy Accelerator".

And to answer your other question, it sounds fantastic! Snapped a pic listening on my JBL C40s in the messy garage ...

PXL_20230125_051731270.jpg
 
Your math is correct. Note that the 7300's meter circuit uses both voltage and current sensors to calculate power. The current sensors are those resistors going thru the donut in one of your pictures. I would think that this should be well within the range of adjustment.

As for the twitching meter, it seems that you don't hear any transients in the audio, so that means the issue is in the meter sensor or circuit. Get at each sensor input on the meter board with your scope to see which input (if any) is causing the twitch. If you do't see anything at the board inputs, look at the output pin of each opamp in that chanel's meter circuit.

Those C40s look stunning! Bet they sound as good as they look.

This one's been (and continues to be) a tough one. I've chased my tail in similar ways so don't feel bad about the 4/8 ohm dummy load issue... the other day I spent 45 minutes troubleshooting a radio receiver that had no sensitivity until I figured that the RF OFF button on the sig gen had somehow gotten pushed. Yep, the radio worked fine - problem was between my ears.
 
Thanks. Yeah, the C40s sound awesome. I like them better in many ways than my L300s.

Last night, I adjusted the meters. (Now that I finally got the math right - thanks for helping me through that.)

~5.7ish divs at 1V/div = ~2ish Vrms

PXL_20230127_005816464.jpg

~2Vrms over 4ohm loads = ~1W (the geometric mean between 0.3W and 3.0W).

PXL_20230127_005820638.jpg

To get the scope connected to the meter board, with the meter board sandwiched between the faceplate and autoformers, I am thinking I can solder short wire leads to the test points and hook the probes up to the leads. A project for the weekend (hopefully).
 
Excellent! Yeah, that's the way to do it if the board isn't accessible. I've had to do the same thing with equipment that had cards that plugged into a motherboard and had no extender card to get it out of the cage to get access to the board.
 
Well, the meter was behaving and I had it playing at a pretty low volume all morning.

I wasn't in the garage but all of a sudden I heard the music get way louder ... went into the garage, one channel's meter is WAY higher than the other (just like the original problem!)...

This time it does audibly clip - sounds horrible - at any volume above a whisper.

Hooked it back up to the dummy loads (4ohm taps) & scope ... hooked up 1kHz test tone ... here's what I'm looking at. Left channel in protection, audibly buzzing again, and heat sinks getting HOT fast.

Picture taken with both channels' gain at max (the smaller truncated wave being the L channel):

PXL_20230205_010107925.jpg

PXL_20230205_010111741.jpg
 
Last edited:
Be ready with your DVOM to start to take voltage readings. Tap around with a chopstick to see if you can prod it into misbehaving. You need to catch it in the act, obviously.
It almost seems like AC from the light bulbs is getting into it, or it's loading down the power supply so much it's causing the buzz. What frequency is the buzz - 120 hz or 60 hz? If 120 hz it's probably off the big power supply since the bridge rectifier output would be 120 hz DC pulses. OTOH the +/-12V supplies are half wave filtered so you'd be getting 60Hz buzz.

I'd hit the +/-12V supplies first with your DVOM when it acts up. Take both DC and AC readings. Any appreciable AC means that power supply section is suspect. Then proceed to take voltage readings on the driver board and compare with the manual. Sooner or later you'll find something off.

Looks like both channels were distorted. And at least we know it's not oscillating. If it were, there'd be fuzz on the scope traces.
 
Thanks ... let me add some info:

If the R gain was 100% with the L gain 0%, the R channel would produce a clean sin wave.

But as soon as the L gain started to get turned up, I could see the R channel wave form starting to deform, and when L gain reached 100% I saw what I pictured above.

In other words, it's as if the R channel is being distorted by the L channel.

I am not good at identifying the buzz by ear so the next time it does it, I'll try to use the spectrum analyzer app on my phone. It seemed like it was coming from the autoformers rather than the power supply, but I can't be 100% sure. The L channel heat sink (but not the dummy load) heats up FAST when this happens so I'm nervous to leave it at 100% gain for very long.

Will poke around from here.
 
Last edited:
Something has a bad / weak ground path.

Follow the ground traces around, ground connections, heatsink mounting, all of the common points and I’ll bet you find your culprit.
 
Yes, in one of the previous pictures there was evidence of repair work done on the capacitor PC board, where the common point of the autoformers actually goes to ground. If there's voltage at that point compared to chassis ground then we have a ground problem, certainly.

My point made above is this - if the +/-12V supplies aren't right, funky things might happen in the op amp preamp stages. If one or both sag appreciably, the input signal is going to be clipped, and to make matters worse the PG detection also uses the 12V supplies.

The +/- 90V supplies could sag a bit due to high loading, but a lot of sag on either the + or - will upset the bias and current detection operation. It's always best to make sure the power supplies are correct.
 
Thanks all. Sitting there with the test tone for an hour+, no luck, normal operation.

I started with a wooden chopstick, poking every single harness connector, prying between components to shift them around, prying under circuit boards to get them to bend, nothing. Normal operation the whole time.

I remembered the earlier mode of causing the issue: I started picking it up from the right front foot area (grabbing the corner under the right meter and lifting), in the same way that created the same buzzing & protection in the beginning of the thread.

It was successful. I have to manhandle it a bit - lifting aggressively up and down - causing quite a bit of physical movement of the components - but doing so does create the same buzzing situation. Now I can see on the scope that the buzzing is accompanied by the deformed waveforms showed earlier.

Some observations:

-If both channels are off (either min gain or no signal), no buzzing.

-With an input signal, if R channel at 100% gain, L channel at 0% gain, no buzzing.

-With an input signal, if L channel at 100% gain, R channel agnostic, buzzing.

-The deformed waveforms are correlated with the buzzing.

-Buzzing causes immediate heating of the L channel heat sinks.

-Buzzing does not cause heating of the dummy loads.

-The buzzing is coming from the power supply area, seems almost like from the protection relay, but not sure.

I haven't been able to record it with my spectrum analyzer since my phone is already being used to generate the test tone. I may switch to my laptop for the test tone so I can measure the buzz with my phone.

So, I can turn to ground chasing now. We did some ground-checking before and came out clean, but I will look again.
 
No success finding the intermittent ground, but I DID figure out what's buzzing.

(I didn't get any buzzing on the test tone, but after a bunch of jiiggling by lifting up under the right autoformer, I got it to short & buzz pretty easily, along with the left meter going crazy high and left protection kicking in.)

So what's buzzing - It's the Right autoformer. The buzzing comes from somewhere inside of it.

So I will focus on following where it connects to ground and go from there ...

I also found another clue that someone has been in here before.

Look at the zip tie ... what's that mettalic thing? It's a blob of solder!

PXL_20230217_233603864.jpg

That zip tie is directly below the diode bride rectifier ... hmm ... someone has de-soldered that before I guess?

PXL_20230217_233659184.jpg
 
Back
Top Bottom