Shure M67 Self Oscillation

[heatwalk]

Hi all,

I have a Shure M67 I was intending to do the direct out mod to and give to my brother. I've already recapped the PSU, but I have encountered a strange problem and not knowing what causes it, I’m not sure how to resolve it and don’t want to move forward until I understand what is happening or if I messed something up.

Using it in a chain of Bass / DX7 → Passive DI → Shure M67, the unit functions and sounds as expected 98% of the time, but unplugging the instrument cable on my bass or turning off my synth will sometimes send the unit into a bizarre low frequency oscillation. With the bass, it is every once in a while, but with powering down the synth (or unplugging the cable from the synth to test once it started happening) the issue is far more common.

Upon initiating the issue, the needle starts pinning and rattling like crazy as if something is sustaining at max volume, yet there is no longer a source, and nothing coming out of my headphones (listening at normal levels). Input and output volume pots don’t reduce the amplitude. Pulling up a frequency analyzer in my DAW shows that the frequency is an inaudible sine wave at around 10Hz, so really I really hear the needle tweaking out more than anything else. Turning up my headphones I can hear fluttering too, presumably the headphones moving air and not the tone?

The ways I can resolve this are -
-Flipping it off and on again/ flipping the power to "battery check" - always works
-Touching the tip of the recently unplugged instrument cable - works most of the time
-Plugging the cable back in to the instrument - sometimes works, but seems to not be an immediate fix but a slow release from the oscillation

Once I get the oscillation to stop, I can start playing again, but unplug a cable again or turn my synth off and all bets are off.

Since turning the volume to 0 on input or output doesn’t do anything once the problem begins, it makes me think it is something to do with the master amplification circuit, post output volume pot. If the input or output is set to 0, I can’t initiate the problem.

Is there a chance that overloading the circuit via a disconnect “pop” can send it into self oscillation? I don't know what else would be happening. With unplugging the bass, most of the time it just sounds like when someone plugs or unplugs a guitar from an amp with the volume on and you get high volume hum, but once it has a mind of its own it doesn’t sound like that at all. I should note that I believe the output relay in my DX7 is starting to go… but as this is happening with the bass too I don’t think this is the only culprit.

I know the easy solution is to turn the input gain to 0 before unplugging since there isn’t a mute input option, but I’d like to understand what is happening from a technical perspective and actually fix the problem. I've already worked on one of these units previously and did the same thing I was intending to do to this one and am not able to recreate this problem on the first unit, so it is localized to the second one.

Here is few second snippet of the oscillation if that helps at all. Be warned it has cable disconnect noise followed by the silence (the 10Hz oscillation) as that is how I initiate this.
https://drive.google.com/file/d/1d9jSIUu2iGHGZi0CfO-wOH4qwhg0IXS0/view?usp=sharing

Thanks.

 

[Monte McGuire]

You say that the PSU has been recapped, but how about the rest of it? Sometimes when electrolytics go "high impedance", nodes that would normally be stabilized can bounce around, perhaps causing this oscillation?

Also, in my experience, the smaller caps in a circuit can dry out faster than the large caps, so any small electrolytic bypasses in the rest of the circuit should also be replaced with a working cap, not just the larger PSU filter caps.

Honestly, in lots of equipment, I find that the larger caps are still fine and only the smaller ones need replacement, which is quite convenient, since some of the larger ones are hard to replace because of their unusual physical size.

 

[heatwalk]

Well I went through with replacing the smaller electrolytics in hope that would be the quick fix as I had ordered those as well but doing so didn't seem to help at all. Probing around I found that shorting the circuit to ground just after the coupling cap C21 stops the oscillation, while shorting after C23 to ground stops it momentarily but it can slowly oscillates back up so I imagine it has something to do with around C21.

Good to know in general about smaller caps drying out for future reference, thanks.

 

[Tubetec]

Suspect some of the higher value resistors to have drifted also , cold check anything above 470 kohms with the multimeter , 10-20 % drift wont matter much but you might have a few components way beyond that ,especially with carbon comp resistors more than 1 meg and 50-70 years old. If you have to replace anything never mind getting the legs of the old components out of the tag strips ,clip the component out with a snips and solder the new component to what remains of the old lead wires .

 

[heatwalk]

Alright went through with measuring today looking at all the large value resistors around where I was guessing something is off as recommended. Took measurements of the resistors on the functioning unit I have in addition to the faulty one (even though the resistors would be just as old in the functioning one, figured it might point to where the problem is). Results are... inconclusive and confusing for my novice brain but feels like this a good learning experience.

Some of the values are within the 10-20% drift, some seem way too far off. Some resistors installed aren't even the value the schematic calls for (but are the same installed value in both builds). I'm also getting different values depending on the range I set my multimeter for (20M vs 200M), and which orientation the probes are on the resistor. It is a cheap multimeter if that explains the inconsistencies, maybe I'll switch a new battery in as well, but more likely I'd assume I'm just lacking a basic understanding for taking measurements? It's been unplugged for a couple days now but I suppose I need to go through and discharge all the caps before any measurements are taken?  I did keep the power switch at on while it has been unplugged, figured that would do the trick

 

[Bo Deadly]

You can't really measure resistors in-circuit necessarily. A meter just applies a tiny current and measures the voltage drop. But if any current gets diverted due to other components, it will throw off the measurement.

Turning the power switch on after unplugging it should drain everything although with this circuit it might take a few minutes. Measure the voltage with your meter at the 1000uF filter cap, unplug it (leaving the power switch is on) and then watch the meter to see how long it takes to get to 0V.

If grounding the base of Q11 stops it like you say, then I do agree that it's probably downstream from there.

It could be that you installed an electrolytic capacitor backwards. When that happens it can get fried and leak current which could be the source of oscillation.

It could be a bad solder joint. In particular a bad ground solder joint could definitely cause oscillation. Look closely at the solder joints using a magnifier or take a picture with a good phone in bright sunlight and study it on your computer. Look for cracks. When a solder joint goes bad it's just the cracked pieces touching each other. After a while oxidation builds up and the joint resistance goes up.

It could be a busted silicon part like a bad rectifier diode or transistor or one of the rectifier diodes in the meter. You might consider changing the rectifier diodes on the power transformer secondary. Look closely for anything that looks like it's been stressed by heat. Maybe the board looks like a little discolored around a part. Power related silicon parts like rectifier diodes tend to get burned up inside and then not turn off or on completely or too slowly.

It could be a broken trace. This is common on old pcbs. When removing electrolytics there is a compulsion to lever on the end of the cap while desoldering. That is bad because the other lead will be pushed down through the bottom of the board which is a good way to lift that trace and break it.

It could be the wrong value of capacitor. If the cap is supposed to be 1000uF but you use 2200uF instead thinking more is better, it's not.

Another possibility is that the thing was unstable all along because of the design. The design has a couple of 2.2M downstream of the master gain which is dubious. Shure designs are more infamous than famous but this M67 thing is actually not a disaster. Its all transformer I/O which has a lot going for it. Just keep the levels going into it low if you want vaguely good fidelity.

The biggest clue is probably that the oscillation is at such a low frequency. An oscillator requires a current source and a capacitance. Low frequency means that the impedance of the current source and capacitance have to be large. At 10Hz, that means it almost certainly has to involve the power supply filter caps. As for the impedance, that could be from a bad solder joint, reversed cap or bad silicon part.