Class 2 ceramic buzzing aging reversal?

[coppercone2]

So In alot of old stuff near transformers I find ceramic capacitors that whine. If replaced with foil capacitors they are fine. Its not the main capacitors because I replaced all the electrolytics once, and the ceramic was still buzzing. You can also replace them with modern ceramics but I usually just put foil in there. They are usually 50nF or so.

Anyway its kind of annoying to replace them, you would think a piece of ceramic would be fine.

Does this have anything to do with aging of the capacitor? Or is there some kind of other mechanism that makes it buzz like a de lamination of the metal or something?

I just got done replacing a bunch but I kind of wonder if I heated them up to 150C with a hot air station, would the noise go away. I don't have any more noisy equipment right now. I mean the class 2 disk caps with cheap dielectric.

I mean its a piece of ceramic, what the fuck goes wrong with it? Has anyone ever managed to do a on-pcb repair of these with a hot air station or something? They are often in awful places to service (you end up like major kong).

I read you get a full crystalline structure annealing or whatever if you heat it up past 150C.

The disk ceramic capacitor looks like it should be very robust. The best I can figure its either something to do with the crystal being weird, or like something is unstuck and flapping around.

I checked a bunch of the caps I desoldered in the last few weeks. The 0.05 ceramics read between 0.04-0.05. When I took a 0.04 and heated it with a 250C iron for a minute on top of an altoids can, 15 minutes later I measured it at 0.045. So they are aged, but I really don't know if I want to put them back in the circuit now that it works right. 

The new film cap are practically inaudble, but the desoldered z5us can be heard a fair length away from my ear when connected to a source. The heated one does not sound different then the others.

I wonder if these supplies were all noisy to begin with, or if this made less noise in 1965.

ha ha, just imaging that piezo disk vibrating on top of the PCB is amusing. Its kinda like one of those inflatable car lot sales statues. how completely ridiculous.


i present you, the Z5U man.


The more I think about it, the more I am thinking I am going to bother to go through the dirty work and replace them with foil capacitors. Eww.

[jonpaul]

Céramique disc caps can act like piezoelectric transducers, especially the high capacity material

An NPO is less noisy and Z5U more.

We have seen HP precision linear like 6114A (?) buzzing loudly from old film 0.47 Mf noise bypass caps.

Check out the chassis and hardware for loose metal and screws

Jon

[wraper]

As heating resets capacitance aging, they will have more capacitance when cool down. But I don't noise will go away because of that unless it somehow changes ripple frequency in the circuit. Replacing garbage like Y5V with X7R or X5R nearly always fixes or drastically reduces the issue. Also circuit may be just unlucky to have a capacitor part number which have mechanical resonant frequency close to the ripple current frequency.

[coppercone2]

so.. you think that the noise that bugs me is standard for a new instrument made in the 1960s? (all my HP supply amplifiers do this, the PSU are usually fine though)

I figured something had to have went bad because it just sounded terrible. Maybe that was acceptable then.

is there some kind of damage mechanism that causes the capacitor to make noise?e I replaced all the electrolytics with high quality ones before I found the ceramics were doing it. 


Could it be something relating to the diode aging or something that makes the capacitor make more noise (they are near rectifier). ? or some transformer aging that electrically stimulates the capacitor to make more noise? ( I verified its the capacitor by putting a hose in my ear and checking what is the loudest). *

I just can't imagine them selling something that loud and obnoxious, I would imagine the marketing team would immediately make them replace all the ceramics with foil capacitors once they heard it. But its hard to imagine since ceramic is such a invariant material.


jonpaul, yes, the film cap in the latest amplifier repair was by far the worst offender. The ones that look like a pack of chewing gum (the rectangular axial ones, I replaced with round ones that are new from digikey). those things were terrible.

*I am absolutely certain it was the ceramic caps because as I desoldered them (there was four), it got quieter and quieter on the ear-hose, dead quiet with no ceramic caps, and made no perceptible noise when I replaced it with 400V film caps). Now it just has a faint pleasant buzz of the transformer, no high pitched whine. In this specific case, there was 4 ceramic on the PCB near the rectifier diodes. Those capacitors sound exactly like tinnitus. This thing managed to piss me off from across the room. No way it got through HP like that unless they put a cement mixer on the board room. this one was last year, and the ceramic caps in the thing I am working on now were not nearly as bad, but still audible. I can hear them from 6 inches away when plugged into a 8KHz function gen set to 20Vpp

[wraper]

It may be a failure too, if they did not sing before. If ripple current increases or frequency changes, noise may appear. With modern equipment where high ripple current is present like graphics cards, it may change from batch to batch or manufacturer. Even though the same reference PCB was used. Just unlucky combination of parts and parameters and capacitors start singing.

[jonpaul]

Rebonjour, all the HP precision 1980s PSU 6114A etc have the cap  buzz.

https://xdevs.com/doc/HP_Agilent_Keysight/HP%206104A,%206114A,%206105A,%206115A%20%20Operation%20&%20Service.pdf

I fixed three of mine with terrible buzz of film caps


Jon

[coppercone2]

Yea the supply amplifier i have is built the same including the terrible way to mount the displays on the front panel (garage door construction).

BTW I totally HATE that power supply you have because the switch matrix is utterly terrible. When you combine that switch with the other parts that are in there it feels like a joke. And the fact that none of the displays have bezels with screws and stuff and just wedge in there is really obnoxious too.

I recall that supply having round hermetic LM741 amplifiers in there for the feedback loops, alot of precision resistors, and those utterly terrible switches. If they put the switches that have the lever on them to flip between the digits it would be a whole different story (like gigatronics).

Do you know if those square capacitors they wedge under the transformer eyelets are naturally square? I see they are not available in that form factor anymore. I wonder if someone just put them in a hot plate and squished them into a mold or something in the factory

[coppercone2]

It may be a failure too, if they did not sing before. If ripple current increases or frequency changes, noise may appear. With modern equipment where high ripple current is present like graphics cards, it may change from batch to batch or manufacturer. Even though the same reference PCB was used. Just unlucky combination of parts and parameters and capacitors start singing.

Well with a linear supply, what could possibly happen that leads to this condition? All that comes to mind is something with the diodes, the transformer windings shrinking (inductance change), and perhaps down stream decoupling capacitors having lower capacitance. Its hard to imagine those things leading to excess capacitor noise (unlike a switch mode power supply). Maybe if I get another one of these singing linear devices I will try replacing the rectifier diodes first, to see if the noise is related to that. I can't help but wonder if I am just masking a problem by putting foil capacitors down to mute it, especially if the ceramic caps are fine, but are improperly loaded by some circuit condition. How do you damage a damn ceramic disk?? it seems invulnerable.

I wonder if its just insufficent filtering at the input because the power grid got nasty compared to 1960 in terms of harmonics because of all the switchers (solar, car, small smpsu, furances, hvac etc). So maybe if you brought one folward in time, it would be just as loud because it was not designed to handle such a chaotic power grid situation.

Maybe a big beefy low noise EMI filter would be a good piece of test equipment, to see if plugging it in between the PSU and the wall would help.


******
Just to clarify how I got down to the bottom of this last year (I confused myself in the thread by now)
1. Replace all power transistors and put new thermal grease (one was blown), disassemble and clean all the potentiometers, repolish all screws, grease all contacts with deoxit grease after careful polishing to get rid of tarnish.

2. replace all electrolytic capacitors with good quality ones from digikey (including using axial and radial as needed)
-still hear noise, blamed transformer (somewhat quieter)
3. Desolder transformer and vacuum varnish it (after cursing a lot)
-part of the noise went away, but the high pitched tinitus one was still there
4. fix PCB trace and repair amplifier because solder joint cracked on transformer replacement'
5. Snoop around with tube to determine the ceramic caps were buzzing, check performance of amplifier and it seems fine (no excessive electrical noise, good regulation, plays audio music through a speaker fine)
5. replace ceramic caps with film and enjoy silence

So I feel like I basically replaced everything that could have been making the capacitors act badly. I measured alot of the parts when I was going through it when the trace broke, all the diodes looked OK on Vf.
*****

Wraper,

Could you give me a plausible combination of factors that could lead to the capacitor singing based on drift of various part in a typical linear amplifier?? (assuming the above list was checked where the power transistors, electrolytic caps, potentiometers and interconnects were serviced). Maybe you need a small series resistance between the filter caps and the rectifier or something?

[jonpaul]

Rebonjour cher CopperCone...

We have used HP 6114/5A precision 40V and 100V over a decade.
Were my only encounters with buzzing caps,   Flat,  oval sides, mylar film, US made.

The  HP precision linear PSU use  film caps (not ceramic) as noise bypass on the  primary (or secondary?) transformer  windings. Caps are stressed in ways the DC bus caps and coupling are not.
Perhaps the  AC voltage continuous V applied caused  dielectric adhesive to age or separate, allowing the foils to generate the buzz? Or, plastic encapsulant/end seals plastic might have dried out or cracked.
 
Just my memoires...

Jon

PS: Afraid, that I am not in agreement with you  about the HP 611x PSU...

I am well aware of the flaky and hard to access thumbwheels on the very well designed HP classics from 1980s.
Acquired off ebay and silent key estate,  needed cleaning, especially the thumbwheels.
One had a damaged 10 turn current limit Beckman Helipot.

They are VERY good as secondary references, and general use, with the large meters and precision V and I.
IMHO, very fine instruments (HEAVY!) for the analog/PSU/linear aficionados.

REF  HP Tech journal 1972 p. 16-20
https://www.hpl.hp.com/hpjournal/pdfs/IssuePDFs/1972-11.pdf



 

[wraper]

Well with a linear supply, what could possibly happen that leads to this condition?

Something oscillating. If it's a purely analog circuit, with no fancy power reduction features that include some form of switching, then it can only be an unstable feedback loop. Quite often PSUs have oscillation problem in CC mode.

Replace ceramic caps with film and enjoy silence

Overall film caps are worse for decoupling purposes. As I said using better ceramics usually solves the issue (in a circuit which is not faulty). However as it's a linear PSU, I don't think ceramic capacitors are a problem. Very likely it's not even they what is singing.

[coppercone2]

When I move a hose around inserted to my ear, it was easy to identify the capacitor as being the loudest thing. I scanned every structure from like two angles, and the ceramics were clearly making the noise.

The hose works well too becuase it does not transfer the transformer hum very well (you need to push it into the transformer body to really hear the hum) but its like having mrs incredibles ears for finding high pitched shit.  It sounds like 9+ KHz

Try it with a disk cap connected to a signal gen, a simple flex hose (3/8 thin wall) has very impressive fidelity. But a foot long section of tubing will deafen you to the transformer hum, so you can only pick up the high pitched stuff clearly. 

[coppercone2]

wow those front panels are just something else as far as serviceability goes on that style of powersupply 6xxx. thats some damn good bush. Thats like if you had a buncha of trees and bushes growing up against the house and you need to change a garden hose faucet

I like taking that stuff apart but I think I am going to wait till something breaks on this one, a little bit of deoxit and some oil for the switches is enough for me on this one thank you very much

[coppercone2]

Rebonjour cher CopperCone...

Oh the specifications I think on it are wonderful and so is the control circuit, its just the input switches are flaky as hell, and it has some other shoddy mechanical deficiencies. Its like a ferrari with a floor tile for a steering wheel. I was wondering what might happen if you replace those 741's with better op-amps too. The idea that you need to clickity click your way to a different voltage is crazy too.

Don't you think that if they put a bezel with 4 screws for the each panel meter, a solid front panel and nice metallic switches it would be alot better? The panel meters just get cleverly... uhmm.. tensioned by two hooked plates. It makes me think of a garage door, but without the pin. Its a side ways jigsaw puzzle with extrusions!

That would be a great design, for plate mail on a knight's armor, but not for a front panel. D-

The other HP power supplies have front panels that get clipped into place by a plastic clip, which also sucks, thats another repair task I Have to deal with, I think I want to try making a replacement part out of brass stock or something. That click breaks and the display gets pulled into the unit at an angle.

HP seems allergic to having screws on the front panel for some reason!

On this dual 6000 series supply, if you have a loose screw on one tie rod, the entire front panels turns like fluid.

Not everything is terrible though, for instance they managed to make the 3310B function generator to have a solid plate for the front panel.

[floobydust]

I have various equipment that makes buzzing sounds that come and go, and on certain times of day and dates. Then one day I saw it on a scope while looking at mains, a few squiggles that moved around every 15 seconds or so.
Long story short, I traced it (buzzing) down to a power quality problem. It seems to be an industrial heater using phase-control that I traced back to a nearby large medical lab, so I'm assuming it's a dishwasher/sterilizer. Every time it runs, I get buzzing with the heater ramping up and down. From LCD monitors, light bulbs, you name it.

It might be worth while checking your power quality for higher harmonics.

[coppercone2]

Oh yeah this is right up next to a solar inverter. But some of this stuff is treatable because for instance I just refurbished and fixed my old lambda LQD power supply, with new caps and some other repairs.

When I plug in a 0.5A load to it that is fairly digital (a RGB light), the left side makes the annoying whine but not the right side. They are powered from the same transformer... and I have a feeling its one of those film or ceramic caps I failed to replace.

The unit is exactly duplicated on both sides.

But also that PSU was not used for a long time, and I noticed the buzzing started after about an hour. When I switched the load to the other channel, it was nice and quiet. But now the other channel makes the same noise even after it cooled down for a few hours.. so that means that something must have happened from the heat I would imagine that caused a permanent change to the circuit.

[coppercone2]

oo managed to plug a card in backwards the ammeter went bye bye 

problem : 10 hour repair day

[coppercone2]

well now that the amp is repaired with the revarnished transformer, greased transistors etc, with all the buzzing parts replaced, I finally have some information.

The two buzzing foil caps that I replaced with modern foil caps are quiet. However on high load I did hear an annoying buzzing noise, not so much a whine, but its audible on max load.

It turns out the the new single layer ceramic 1uF disk cap (not MLCC) I put in (bought 2x NOS I believe) is making the noise. I think it has slightly less annoying noise then the old one, but it seems that either the ceramic is also aged and also has a problem, or if it could be considered brand new, its just a problem with using the ceramic cap in this circuit.

I will replace it with a 1uF foil cap and see what happens, if I can figure out where to fit it.


So its either:
1) by design,
2) related to a aged transformer or some other part (I checked every resistance, capacitance and diode in the circuit, and if it was out of spec while measured in circuit, I desoldered it to measure.. it seems to be all within spec save a few resistors near the ammeter). I left the small original electrolytics, foil and tantalum but measured their ESR and capacitance out of circuit.
3) the NOS ceramic part I bought  has the same flaw as the original (flaw related to absolute age, not use)
4) dirty modern power

Now since the transformer puts out the correct voltage, and the rails are in order, I think that #2 is unlikely, unless its related to some transformer parasite (capacitance over age?).

I think its #4 or #1. If I ever get a compliance lab 60hz power system, I can rule out #4.

[coppercone2]

Anyone know what a capacitor that is 105K 250G means (Meritek?), I figure its 250V but anyway

I replaced the SLCC with the 1uF foil cap, the noise went down.

There is a little bit of buzzing still on full load that you can hear if you put your head close to the chassis.

I tried to chase this noise down, but it looks like its the vinyl tube that has the transformer wires running through it that is buzzing. Not sure what that is (its the board on the back of the units with the power transistors, its built with slots so  its difficult as hell to inspect that area without ripping a wire.

I think its just the transformer sounds a little nasty at full load DC.

Yeah I spend some time listening to it while 'burn in' testing my repair on a 10 ohm power resistor that got steamy.. it looks like its all coming from the transformer now. But its still hard to hear with a totally open chassis and two missing screws on the side wall. I think its mission accomplished.

I will calibrate it vs a thermal RMS meter or a 34401A with a low thd source