An interesting repair

[Mark Hammer]

A local buddy brought me his Diamond Memory Lane 2 pedal for "curing".  Everything about the pedal worked just fine, but it had this annoying audible whine, that synced in pitch with the LFO.  He had a second one that was dead quiet, so clearly it wasn't the basic design of the pedal.  At first I thought that maybe the balance trimmer/s on the one or both of the BBDs were not set optimally on the offending pedal, and we were just hearing whatever clock whine was still leaking through.

I wrote to Diamond, describing the problem and asked for information to identify the relevant trimmer/s.  They provided a pic with the parts circled, and advice for setting it properly.  I followed the advice and dickered with the trimmers, but no dice.  The whine was still there.

I wrote back, indicating that the problem was not resolved in the trimmer setting, but this time gave a little more detail about the problem, noting that it came on after about 5 minutes.  I wondered if it was maybe a bad cap, or a heat-related problem, since both of those tend to be time-sensitive.  The repair guy (Brian Fecteau) wrote back and said that my description of the time aspects tripped a recollection of an issue they had with an earlier run of the pedal.  The pedal has three 3-pin regulators in it: 5, 12, and 15 volt.  During an earlier batch, they found that the On-Semi regulators they had used had unusually thin heat-sink tabs, and were causing problems because of that.  He sent a picture of what it would look like, and sure enough, this unit had the offending regulators for both the 5 and 12-volt sources.  The LM7815 had the normal thickness.  The regulators are not thermally-coupled to anything, including the chassis, but are essentially free-standing.

Some time this weekend, I hope to cure the pedal by de-installing the regulators and replacing them with some nice thick LM340T-5 and 12 units.

The things you never suspect, eh?  What could go possibly wrong with a voltage regulator?

[Tony Forestiere]

Isn't it amazing what background, experience, and a sharp memory can do? Good job to Brian!

[Rob Strand]

Sounds like that company provides a good service.

The things you never suspect, eh?  What could go possibly wrong with a voltage regulator?

The correct use of regulators is very underestimated.  Many things can go wrong.  People often don't put the right caps around them or meet all the requirements set out in the datasheet. 

In production you end-up finding all kinds of weird stuff.  Especially issues with marginal designs where people haven't exactly done the right thing - that *is* where the wrongs come back to haunt you.

I have doubts that the thickness of the tab being the actual cause.   The heat removed from a free standing regulator largely due to the surface area which doesn't depend much on the plate thickness.    While there is heat conduction temperature drops over the plate, the drops are extremely small compared to the surface to air temperature drop  (I'm assuming a TO220 ?).   If it's temperature related I suspect the unit would work on cold days and not hot days (or with the back off).

Having said that there probably *is* something different about those regulators.  The batch of regulators with the thin plates are the ones that show this problem *but* the thin plate isn't the actual cause.   To me thin plates = cheap = maybe a dodgy china part.

[Here's the insides, regulators on the left:
http://files.effectsdatabase.com/gear/pics/diamond_memorylane2_002.jpg
]

[Tony Forestiere]

"Bean Counters." 

[Rob Strand]

"Bean Counters."

In the past when you bought parts you had no reason to suspect one regulator is different to another.  Nowadays you can't trust anything, cheap parts or expensive parts.

[Mark Hammer]

Sounds like that company provides a good service.

The things you never suspect, eh?  What could go possibly wrong with a voltage regulator?

The correct use of regulators is very underestimated.  Many things can go wrong.  People often don't put the right caps around them or meet all the requirements set out in the datasheet. 

In production you end-up finding all weird kinds of stuff.  Especially issues with marginal designs where people haven't exactly done the right thing - that *is* where the wrongs come back to haunt you.

I have doubts that the thickness of the tab being the actual cause.   The heat removed from a free standing regulator largely due to the surface area which doesn't depend much on the plate thickness.    While there is heat conduction temperature drops over the plate, the drops are extremely small compared to the surface to air temperature drop  (I'm assuming a TO220 ?).   If it's temperature related I suspect the unit would work on cold days and not hot days (or with the back off).

Having said that there probably *is* something different about those regulators.  The batch of regulators with the thin plates are the ones that show this problem *but* the thin plate isn't the actual cause.   To me thin plates = cheap = maybe a dodgy china part.

[Here's the insides, regulators on the left:
http://files.effectsdatabase.com/gear/pics/diamond_memorylane2_002.jpg
]

1) The pic shows "good" regulators, in terms of tab thickness.
2) It's possible that tab thickness is not a cause, but merely a marker of a problematic component.  Possible.  I suspect one would have to gather data over some reasonable period, say 10 minutes, of regulator temperature, and stability of output voltage over the interval that transitions from no whine to whine.

Not having the schematic, I have no idea what the 5V and 12V regulators feed/power.  Perhaps that might offer some clues.

[Rob Strand]

Interestingly others have their complaints about the "cheap" TO-220 packs,
https://www.eevblog.com/forum/projects/when-the-(expletive)-did-to-220-tabs-get-so-thin/
http://www.eevblog.com/forum/beginners/different-looking-irf-mosfet/

I suspect one would have to gather data over some reasonable period, say 10 minutes, of regulator temperature, and stability of output voltage over the interval that transitions from no whine to whine.

I think you will find the good and bad regulators will be at pretty much the same temperature.  The only difference is the bad ones will misbehave when they heat up.   Why that occurs is due to subtle differences between the innards of the good and bad devices - all beyond our control.   It might be possible to add a cap so it works for all devices but that could be just working around the dodgy device.  (If there were 1000's of units out there then a simple workaround is justified but it is still hard to prove it will be 100% successful.)

Not having the schematic, I have no idea what the 5V and 12V regulators feed/power.  Perhaps that might offer some clues.

Yeah, I had a look but got nowhere. I was interested to see if the Diamond broke any "rules" as I've seen those regs misbehave in the past.    Suppose it doesn't break any rules then all that means is the dodgy parts go off the rails even under normal conditions - I wouldn't be surprised.

[Mark Hammer]

Given the volume of products they make, I imagine they got their regulators from a major distributor, and nobody in the supply-chain spotted the difference.  It's also fair to note that, while TO-220 stye regulators have a hole in the tab to provide the option of securing the regulator against a larger surface to wick away heat, these regulators are freestanding.  I pondered maybe squirting in a wee bit of thermal compound between chassis and tab to let the enclosure serve a heat-wicking function.  But since I would have no means of assuring the tab remains pressed up against the chassis down the line, the problem could easily return.  So I elected to change the regulators.  Or at least I will later today.

[Rob Strand]

I pondered maybe squirting in a wee bit of thermal compound between chassis and tab to let the enclosure serve a heat-wicking function.  But since I would have no means of assuring the tab remains pressed up against the chassis down the line, the problem could easily return.

It's a tricky problem because it's not clear the regulators are shutting down due to over heating, in which case the voltage tends to drop below its normal value, or doing something else like oscillating.  To shut down the regulators need to get pretty darn hot - like you can't touch them too long.  Given replacement is a known solution you might be best doing that.

Maybe only one regulator is causing the problem?

For heatsinking, you really need to screw them down with a very thin layer of thermal compound.   The tabs aren't isolated and the tabs for negative regulators (&9xx) aren't even ground.  You could get unwanted ground connections or shorts.  It is possible to use insulating washers.   If the board isn't rigid it can cause solder fatigue.

[Mark Hammer]

The regulators aren't shutting down, per se, because everything works just fine - relay-switching, delay range, tap tempo, EQ, modulation, etc.  But their functioning would appear to be unstable after 5 minutes of use.

The plan IS to replace them.  I'm just having a bear of time getting all the knobs off, so that I can unscrew the pots and remove the board to wick the solder away underneath (pots are soldered to the board).  Four knobs came off, and all pot and jack retainer-nuts came off with a modicum of torque, but set screws in two of the knobs are giving me grief and appear to have had their hex slot stripped a little by an Allen key that slipped.  I went to a specialty shop this afternoon in search of a suitable screwdriver bit or Allen key, but everything they had was either too small (lotsa wiggle room) or too big.  I'll try another place tomorrow.

[Rob Strand]

The regulators aren't shutting down,

Sometime the voltage drops a bit like 4.9V and causes the circuit to misbehave.

The plan IS to replace them.  I'm just having a bear of time getting all the knobs off, so that I can unscrew the pots and remove the board to wick the solder away underneath (pots are soldered to the board).  Four knobs came off, and all pot and jack retainer-nuts came off with a modicum of torque, but set screws in two of the knobs are giving me grief and appear to have had their hex slot stripped a little by an Allen key that slipped.  I went to a specialty shop this afternoon in search of a suitable screwdriver bit or Allen key, but everything they had was either too small (lotsa wiggle room) or too big.  I'll try another place tomorrow.

That's a bit of a pain.

I have a feeling you could probably fix it by adding some small caps to the output and/or input to the regulators.  I'd try 100nF on the output.   Then maybe 100nF to 330n on the input.  You could do it one at a time to identify the problem regulator.  If all the regulators are 78xx the inputs might be tied together.

You can get radial ceramic caps which are small and cheap.  They are very easy to retrofit.  You can solder them to the top or the bottom of the PCB.

[Forgot to mention if there is an issue because of the input cap and all regulators are in parallel you might be able to add just one cap and all the problems are solved.]

[blackieNYC]

You'll probably be fine with a good regulator alone, but you could give it an individual screw-on or clip-on heat sink. If you can bend the legs to make room. 
Three to220 regulators seems like a tremendous amount of deliverable current for an analog delay,no?  They got hold of some really awful regulators. If the guy can barely recall this production issue, I guess not many bad pedals made it out the door. Your lucky friend. Either that or Brian will soon be stuffing replacement regulators into shipping envelopes all day long. 

[Mark Hammer]

It's something they fixed long ago.  I suspect that the move to use 1A voltage regulators was simply because it was easy, inexpensive, and did the required job.  The box is big enough to accommodate them.

I have to laugh, though.  In the fall of 2016, I dropped by the Diamond facility, when I drove our younger son out to Halifax for school.  Had a delightful chat with the production manager Aimish, and on the way out asked if it was possible to buy a few powder-coated boxes from them.  Aimish pointed me to a bin of predrilled boxes that were used for discontinued products and invited me to take a few, on the house.  I didn't realize until this local fellow brought over his Memory Lane 2 pedals that that's what the predrilled boxes I took were originally for. 

In other news, Steve Bragg from Empress is coming over to the house later this week for a personal tour of their Zoia do-everything pedal.  The best part is that, unlike NAMM, it won't be against the deafening din of people wanking away on everyone else's products.  I think what I'm really looking forward to is learning more about those teeny hi-res displays that seem to be showing up in more and more products.

[Rob Strand]

Aimish pointed me to a bin of predrilled boxes that were used for discontinued products and invited me to take a few, on the house.  I didn't realize until this local fellow brought over his Memory Lane 2 pedals that that's what the predrilled boxes I took were originally for. 

That's a good score.  Good to see that stuff not going to waste.

Once I bought some electronics equipment from a company that was closing down its development division.   After buying some stuff, the guy said  "you can take whatever you want from those boxes over there".  Got better stuff out of the boxes than the junk I bought!  In some ways it was good compensation because the stuff I bought wasn't great.  Had a similar experience at an auction - after that I learned my lesson about auctions.

In other news, Steve Bragg from Empress is coming over to the house later this week for a personal tour of their Zoia do-everything pedal.  The best part is that, unlike NAMM, it won't be against the deafening din of people wanking away on everyone else's products.

LOL.  You've got some good connections up there.

[PEPPER!]

Sounds like that company provides a good service.

They really do . . . I do some pedal repair for the shop where I work and Brian wrote me back (during NAMM of all times) with calibration advice on a used Memory Lane 2 we got as part of a trade.  You'd think all the knobs on the front would be enough - some people can't resist the urge to open it up and mess with the trimmers!

When I go back I'll see if this unit has the weak regulators and swap 'em out.

[Rob Strand]

They really do . . . I do some pedal repair for the shop where I work and Brian wrote me back (during NAMM of all times) with calibration advice on a used Memory Lane 2 we got as part of a trade.

I like to support good companies like that.  I've had enough of cheap junk and stories of people being intentionally deceived.

You'd think all the knobs on the front would be enough - some people can't resist the urge to open it up and mess with the trimmers!

It's a bit of a double edged sword.   I screwed-up the settings of my first Flanger.  After that I worked out what the trimpots did and I set it by ear.  Not being happy with that it forced me to read-up on the specs, draw out a schematic, set it up again and later down the track modified it for the better.   So I learned a lot from the experience.  However if someone screws around with the settings, make it sound crap, then sells it, that dodgy unit can propagate bad press for a company - that sucks!

[R.G.]

There's a way to test whether it's the "thin heat tab" or not: screw a postage-stamp sized bit of aluminum on them. A TO-220 with no sink will dissipate about 2W in free air without (usually) cooking the insides of the package. It's easy to double the dissiption area with a bit of aluminum screwed down with heat sink goo.

There is a package that was called "Uni-Watt" as I remember it, and it had a much thinner tab. Maybe they got a good deal on regulators in those. I can't quite make out from the picture. I'd know instantly if you could magnify up the section of the PCB.  Uni-Watt regulators may well have a much lower current limit built in. The long start up time may well be thermal, but in a sneaky way, by modulating the current limit point of the regulator. Of course, the 1A devices in TO-220 cures this. It sounds like when the regulators get hot, they're falling off the threshold of regulation and letting the voltage sag a few tenths, enough to mess up any fine balance.

Just a few guesses.

[Mark Hammer]

Followup.

    To replace the regulators, it required taking the board out to remove solder from both sides of the double-sided board. Taking the board out required removing the retaining nuts on the pots that are soldered to the board and hold it in place. Removing the retaining nuts required removing the control knobs.

    And THAT's where the problem happened. Four of the six knobs came off easily, using an Allen key to loosen the set screw. However the other two knobs were on so tightly, that when I inserted the Allen key and gave a twist, the set screw didn't budge but the hex socket in the set screw got stripped by the Allen key, such that nothing was going to grip it and turn it.

    I drilled a small hole into the set screws, making sure to leave material and not graze the thread in the knob, thinking that I might get the set screw out, pop off to the store and get replacement set screws. I thought that maybe I could find some sort of tool to get in there. Convened an expert panel of four old guys in aprons at Lee Valley Tools, and their best suggestion, after looking at the pedal, was to grind down an oversized Allen key, tap it into place with a hammer, and hope that it would grab the set screw. I did that, but no dice.

    Mulling it over, I realized that if I couldn't remove the board if the pots were still attached to the chassis, maybe I could disconnect the pots from the board and leave them attached to the chassis, but liberated from the board. Worked my way through a couple inches of solder wick, because there is a LOT of solder when standard-size pot lugs are soldered to the board. But I was able to wiggle the lugs with my needlenose, giving the thumbs up sign that the board would not be held down by them.

    Removed the board, replaced the regulators with new ones that had the proper thick tabs, resoldered the pots, and reassembled the pedal. It has been on for a half hour now, and no sign-o-the-whine. Success! I'm going to bed.
     

[Rob Strand]

Success! I'm going to bed.

Yeah! 

[R.G.]

Convened an expert panel of four old guys in aprons at Lee Valley Tools, and their best suggestion, after looking at the pedal, was to grind down an oversized Allen key, tap it into place with a hammer, and hope that it would grab the set screw. I did that, but no dice.
[...]
Worked my way through a couple inches of solder wick, because there is a LOT of solder when standard-size pot lugs are soldered to the board. But I was able to wiggle the lugs with my needlenose, giving the thumbs up sign that the board would not be held down by them.
[...]
    Removed the board, replaced the regulators with new ones that had the proper thick tabs, resoldered the pots, and reassembled the pedal. It has been on for a half hour now, and no sign-o-the-whine. Success! I'm going to bed.   

I have a friend who restores classic cars. As such, much of his life is involved with removing stuck, corroded bolts and such. The car-restorers use some penetrating oil called "S'OK" (... soak ...) which may or may not have some rust-eating stuff inside it. It's relatively magic for removing rusty nuts.

I keep a half-sized solder sucker on hand for the times when some builder has used a pound of solder per joint. It does a good job of getting the big masses of solder off a pot lug and such. I don't use it on PCBs for normal small parts, but for pot lugs, it beats using up enormous amounts of solder wick and many heat/cool cycles on the PCB foil.

And isn't it much easier to fall asleep with something just completed and working?