Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Tantalum Capacitors in audio applications

Status
Not open for further replies.
Interesting. Maybe the new ones are more robust than the old ones. I never tried reversing the voltage... at least, never intentionally.
that's the first spec I have seen written on Vr for for caps at 5%. This allows a safety margin, but my experience has been -10% in low current applications over the last 40 yrs.
 
I also bricked a laptop with a beer spill thru the top keyboard.
I tried fixing twice. 2nd fix worked. My secret was eliminate all residue trapped under IC's where solvent is trapped and air flow is poor.
So copeous amounts of isoprop. solvent and forced air with heat drying time.

Reassuring to know it is not just me :)

We had a reliability engineer at work who was a bit of an animal- big bushy beard and hair to match. He used to drop coffee, tea, pie, cakes ... into his keyboard. The IT guys used to go mad when he kept phoning up to say that his computer was not working. They were changing his keyboard every couple of weeks at one time. But then they solved the problem by getting him a huge sealed industrial keyboard which could be hosed down a couple of times a year. :arghh:

re: beer induced faults
Having worked as Test Eng / Mgr in HDD industry for 11 years, I am familiar with most failure modes. Look for a SMT picofuse around <1A and perhaps an unmarked black 3mm square SMT near power jack. Spindle motor drivers draw more current when wet from square wave frequency and hi-K factor of beer. Hopefully you never turned the spindle in reverse which may damage the heads. in other words, beer makes a good capacitor but may draw unexpected current from large high RPM square waves with contaminants.

Thks for info. The heads were nicely parked when I turned the platter.

As said, I have now got a TORX small driver set, all the way from China, so when I get over the present bug that is going around, I will whip the controller board off and see what's going on. I haven't got a microscope, but I do have a x20 loupe which is quite good.

spec, I thought you got daily viagra emails so I suggest hardening your email spam filter settings instead.
:happy: got it. If only they knew how ironic it is sending ME viagra emails :D:D

Cheers

Chuck
 
Reassuring to know it is not just me :)


Thks for info. The heads were nicely parked when I turned the platter.

Cheers

Chuck

I used to inspect HDD heads once using our SEM in shop( Works just like an old vacuum tube scope) Under 50kV and 50k x magntfication I could measure the gap in the magnetic core and also burn holes in the ceramic sliders with a plasma charge cloud looming over the surface using only 100uA focused in a tiny beam.

I also remember horizontal recording heads ripping off in the landing zone, when the platter direct was reversed. ( like trying to downhill ski backwards with square heels )
I also remember testing Fujitsu 8" drives with dry ice at -40'C for potential military clients and it worked but the spindle bearings squealed for a minute. Usually all HDD"s are rated to 0'C because condensation of 1 water droplet looks like a tank at 100 kph surface velocity to a HDD head flying 1 u" above the surface.

But perhaps the vertical recording heads are more robust on their sliders. They usually buzz them radially before starting the rotor to break stiction threshold, then start spinning.
Dust and moisture trapped under an SMD chip can act as unwanted capacitance. Usually you can trace +12V or +5V now to the chips to ensure power is avail, but also possibly the crystal OSC. is overloaded with beer residue capacitance and there is no clock.
 
I used to inspect HDD heads once using our SEM in shop( Works just like an old vacuum tube scope) Under 50kV and 50k x magntfication I could measure the gap in the magnetic core and also burn holes in the ceramic sliders with a plasma charge cloud looming over the surface using only 100uA focused in a tiny beam.

I also remember horizontal recording heads ripping off in the landing zone, when the platter direct was reversed. ( like trying to downhill ski backwards with square heels )
I also remember testing Fujitsu 8" drives with dry ice at -40'C for potential military clients and it worked but the spindle bearings squealed for a minute. Usually all HDD"s are rated to 0'C because condensation of 1 water droplet looks like a tank at 100 kph surface velocity to a HDD head flying 1 u" above the surface.

But perhaps the vertical recording heads are more robust on their sliders. They usually buzz them radially before starting the rotor to break stiction threshold, then start spinning.
Dust and moisture trapped under an SMD chip can act as unwanted capacitance. Usually you can trace +12V or +5V now to the chips to ensure power is avail, but also possibly the crystal OSC. is overloaded with beer residue capacitance and there is no clock.
Why don't you live near Weston super Mare Tony? Perhaps you could have looked at my drive- many beers, or even a bottle of scotch, would have been the reward for the extraction of the data off the faulty HDD :joyful:

spec
 
Here is my LCR meter and I am calibrating my finger pressure to get 1.0 pF on a pair of measurement wires after null calibration.
THis is useful when you want to tune XTAL clocks with your fingers and find out why the frequency is too high by x ppm.
But I have my fingers calibrated from 10kOhm to 10MOhm and 1pf to 50 nF.

It's very handy to debug designs with your LCR calibrated fingers.
upload_2016-4-7_10-17-44.png


Note in order to measure ESR with a Cap with external bias , it MUST be coupled with a larger series cap with a lower negligible ESR
THis tester does not have this. But it can be done externally. And yes all caps except in general plastic and COG ceramic have a sensitivity of dC/dVdc which is the cause of AC coupled audio distortion that "analog guru" and others have reported accurately.
 
Thanks Tony- looks like a nice job :cool:

spec

(did you post on the wrong thread?)
 
Some of the past is coming back: I think I may have been in error about the use of tantalum capacitors at work. I now realise they were mainly ceramic, and I also vaguely remember the danger of putting tants in an unlimited current situation.:wideyed:

Our boards were mostly digital and we had a standard filter/decoupling regime: a pi filter into the board comprising two ceramics and and a lossy inductor and then ceramic caps placed across the supply lines according to chip current consumption/number of chips.

spec
 
A Vishay FAQ document does say that a short duration reverse voltage is tolerable: (page 5 lower/page 6 upper) https://www.vishay.com/docs/40110/faq.pdf.
'Solid tantalum capacitors are capable of withstanding short duration peak voltages in the reverse direction limited to 5 % of the DC rating at +25 °C'

spec

Cornell-Dubilier says (page 5, right column): https://www.cde.com/resources/catalogs/AEappGUIDE.pdf

"Like aluminum electrolytic capacitors, solid tantalum capacitors
are polar devices (1 V maximum reverse voltage), having distinct
positive and negative terminals and are offered in a variety of
styles"

Regarding aluminum electrolytics, they say

"Aluminum electrolytic capacitors are polarized and must be
connected in the correct polarity. They can withstand reverse
voltages up to 1.5 V
. Higher reverse voltage can cause failure
by pressure build up and rupture of the capacitor’s safety vent
structure."

This does seem different than "Although never specified, all electrolytic capacitors can safely handle 10% rated V in negative bias."


Wikipedia: https://en.wikipedia.org/wiki/Electrolytic_capacitor

says:

"Standard electrolytic capacitors, and aluminum as well as tantalum and niobium electrolytic capacitors are polarized and generally require the anode electrode voltage to be positive relative to the cathode voltage.

Nevertheless, electrolytic capacitors can withstand for short instants a reverse voltage for a limited number of cycles. In detail, aluminum electrolytic capacitors with non-solid electrolyte can withstand a reverse voltage of about 1 V to 1.5 V. This reverse voltage should never be used to determine the maximum reverse voltage under which a capacitor can be used permanently.[49][50][51]

Solid tantalum capacitors can also withstand reverse voltages for short periods. The most common guidelines for tantalum reverse voltage are:

  • 10 % of rated voltage to a maximum of 1 V at 25 °C,
  • 3 % of rated voltage to a maximum of 0.5 V at 85 °C,
  • 1 % of rated voltage to a maximum of 0.1 V at 125 °C.
These guidelines apply for short excursion and should never be used to determine the maximum reverse voltage under which a capacitor can be used permanently."
 
I know this thread is about tantalum capacitors, but I found this Maxim report about ceramic capacitors both interesting and surprising: https://pdfserv.maximintegrated.com/en/an/TUT5527.pdf

spec
Very interesting. I knew the smaller ones were worse but I didn't know how bad. The morale to the story is:

"Note, first, that as the package size increases, the capacitance variation with applied DC voltage
decreases"

There is no free lunch. At one point I forced our marketing guys to shut up and let me publish warnings on our data sheets about not using Z5U and Y5F type ceramics because they are so bad that when used near rated voltage, they have less than 10% of rated capacitance. Of course a Z5U capacitor of a given capacitance was smaller and cheaper so everybody bought them. Then our parts oscillated because we required a specific amount of capacitance to be stable and the caps were nowhere near it. The linear reg data sheets I wrote have details about this problem. We recommended only X5R and X7R but the article shows that going "small and cheap" on those does the same thing.

There is no free lunch in electronics.
 
Very interesting. I knew the smaller ones were worse but I didn't know how bad. The morale to the story is:

"Note, first, that as the package size increases, the capacitance variation with applied DC voltage
decreases"

There is no free lunch. At one point I forced our marketing guys to shut up and let me publish warnings on our data sheets about not using Z5U and YFV type ceramics because they are so bad that when used near rated voltage, they have less than 10% of rated capacitance. Of course a Z5U capacitor of a given capacitance was smaller and cheaper so everybody bought them. Then our parts oscillated because we required a specific amount of capacitance to be stable and the caps were nowhere near it. The linear reg data sheets I wrote have details about this problem. We recommended only X5R and X7R but the article shows that going "small and cheap" on those does the same thing.

There is no free lunch in electronics.

I read your stuff while on the benches. I had a hell of a job persuading some of the engineers about ceramic capacitors variation with temperature and voltage. The bushy tailed grads thought I was some old dodger from the past. I did not know about the case size versus capacitance though. Mind you, the very small surface mount cases were only just coming in when I left design and moved over to pen pushing.

Whadya mean no free lunches- I had tons and free chips to go with them too. :D

spec
 
Last edited:
I read your stuff while on the benches. I had a hell of a job persuading some of the engineers about ceramic capacitors variation with temperature and voltage. The bushy tailed grads thought I was some old dodger from the past. I did not know about the case size versus capacitance though. Mind you, the very small surface mount cases were only just coming in when I left design and moved over to pen pushing.

Whadya mean no free lunches- I had tons and free chips to go with them too. :D

spec
I got a few free lunches provided by salesmen who were trying to get me to recommend their junk to our purchasing department...
 
I got a few free lunches provided by salesmen who were trying to get me to recommend their junk to our purchasing department...
It's a wicked wicked world. Sometimes you would find that the quality of the lunches were inversely proportional to the quality of the product.

The best lunches ever were when the company were buying some Ferranti mini computers for a project. They were delivered once a month and I had to go to their factory in Bracknell UK to witness the acceptance testing and sign off the paperwork for each delivery- silver service every time.

hp were consistently the best for lunches though and their, so called, staff canteen was like an Egon Rhone five star restaurant. Their products were pretty good too, but we all preferred Tektronix scopes.

spec
 
RC highpass filter on the output of the audio generator running at 5Vrms @ 160Hz. R was 600Ω, C was 1µF and the distortion analyzer was connected across the resistor.

Remember though that this was tested below the cutoff frequency of the RC circuit (265 Hz), while any normal audio coupling circuit would operate significantly above the cutoff, where voltage across the cap is largely constant and distortion would be minimal.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top