[John Larkin]

I recommend a study of good audio practice too.

"Good audio practice?" That's funny.

John

 

[Walter Harley]

That's because only *microphone level* mixer inputs have phantom power.

You don't normally plug *line level* outputs into *mic level* inputs !

Hi, Graham.

Many stage devices such as bass amps, guitar effects, and the like, have DI
outputs using active electronics rather than transformers. In live
performance these outputs are commonly plugged into stage snakes. The other
end of the stage snake is commonly plugged into a mixer, often not in the
control of the band. Often the mixer has a single switch for phantom power
on all its inputs, rather than a switch per input (e.g., Mackie, Behringer).
Even if it has a switch per input, often the sound guy doesn't know when to
use it; or the band doesn't know which way it's set when they plug in.
Although phantom power is generally not present on the line inputs, the
sound man may not have the proper adapters or the awareness or knowhow to
switch the snake line from the mic input to the line input. These may not
be desirable facts, but they are indisputable facts.

I recommend a study of good audio practice too.

Of course in a perfect world anyone using audio gear would know what they
were doing, and for that matter, all gear would be designed to be compatible
with the situations it was likely to encounter. As you know, we don't live
in that world.

My interest in this is not as a musician nor as a sound man; in both roles,
I have in fact studied good audio practice as well as good engineering
practice, and I know how my gear works, what's inside it, and how to use it
safely and appropriately. My interest is as a designer and maintainer of
audio gear.

For instance: should I advise my clients to replace the output caps in their
equipment with ones rated for 50V or more? Prudent engineering would seem
to say yes, since I know that those caps are in fact getting exposed to 48V.
But experience would seem to say no, since I don't see a lot of gear coming
in with damaged output caps. The discrepancy between those perspectives is
what's driving my interest.

-walter

 

[legg]

Followup to a question I posted on sci.electronics.components - trying for a
larger audience here. If you're interested in more details the s.e.c.
subject is "electrolytic mystery".

What is the consequence of exceeding the working voltage on an electrolytic,
if there is a current-limiting resistor in the circuit? I am testing a
circuit right now with a 48V supply driving a 15uF 25V capacitor through a
6.8k resistor. It's been running for a bit more than a day, with no
apparent consequences to the capacitor so far. I'm wondering whether the
cap's life is being reduced (and if so, by how much), or what.

http://www.evox-rifa.com/n_america/technotes_electrolytics.htm

http://www.evox-rifa.com/technote_pdf/lytic_voltage_limit.pdf

'Reforming' procedures, with series limiting resistors, may be
required for electrolytic devices, after shelf storage. There is no
guarantee that they can be reformed to the rated voltage - failing
parts will be those whose leakage current does not drop below the
specified level at the rated voltage.

Small capacity, low voltage electrolytics, will not exhibit dramatic
failure modes in energy-limited circuits. If energy is available, they
may overheat or explode.

There is no profit in intentionally abusing perfectly good components.

RL

 

[James Meyer]

There is no profit in intentionally abusing perfectly good components.

RL

Maybe no profit, but it can sometimes be fun. A small tantalum cap can
make clouds of noxious white smoke and a pretty display of flaming sparks if it
is properly abused.

Jim

 

[Martin Schönegg]

Hi Walther,

| I'm pretty perplexed. Obviously the leakage current is very low; so
at this
| point (now that the cap is charged) it wouldn't make any difference
whether
| the 6.8k resistor was there or not. Perhaps it would make a
difference if I
| discharged the cap and then hit it with 48V with no current
limiting. The
| ESR of the cap is about half an ohm, so it would certainly dissipate
a lot
| of power for a little while. Maybe WV basically amounts to a surge
power
| rating - in this case, about 1kW for about half a usec.

You formed this 25 V Cap to 48V one, thats it. The Capacity will be a
little lower than the orginal type. Thats all. Every manufacturer does
this step to create the dielectricum. The more voltage you set up the
thicker the dielectricum will be. If you discharge the cap this is
stable, if you change the polarity the dielectricum will shrink until
it is gone. By forming and disforming it depends on the power behind
this process. Very old electrolytics in devices that where not used
over several years may explode on their first contact with the power
line. There is no effective current limitter before these Caps so they
make a big BANG. If you have a resistor in series the current may be
limitted enough -> No BANG.

MArtin

 

[legg]

Maybe no profit, but it can sometimes be fun. A small tantalum cap can
make clouds of noxious white smoke and a pretty display of flaming sparks if it
is properly abused.

Some day, you're going to be asking the wrong guy for samples......

RL

 

[Walter Harley]

Some day, you're going to be asking the wrong guy for samples......

Homeland security?

 

[Rich Grise]

Hi Walther,

| I'm pretty perplexed. Obviously the leakage current is very low; so
at this
| point (now that the cap is charged) it wouldn't make any difference
whether
| the 6.8k resistor was there or not. Perhaps it would make a
difference if I
| discharged the cap and then hit it with 48V with no current
limiting. The
| ESR of the cap is about half an ohm, so it would certainly dissipate
a lot
| of power for a little while. Maybe WV basically amounts to a surge
power
| rating - in this case, about 1kW for about half a usec.

You formed this 25 V Cap to 48V one, thats it. The Capacity will be a
little lower than the orginal type. Thats all. Every manufacturer does
this step to create the dielectricum. The more voltage you set up the
thicker the dielectricum will be. If you discharge the cap this is
stable, if you change the polarity the dielectricum will shrink until
it is gone. By forming and disforming it depends on the power behind
this process. Very old electrolytics in devices that where not used
over several years may explode on their first contact with the power
line. There is no effective current limitter before these Caps so they
make a big BANG. If you have a resistor in series the current may be
limitted enough -> No BANG.

MArtin

I once either saw or hallucinated a construction project for a home-made
electrolytic make of a 5-gallon tub and a stack of aluminum pie tins.
The tins were separated, of course. I don't remember what they said the
electrolyte was, but it did sound like a way fun project.

Cheers!
Rich

 

[Walter Harley]

Wow. In that case, power dissipation is zilch. I wonder if there's any
longterm thing going on, dendrites or gas evolution or something. Keep
it up and let us know what happens.

After two days at 48V, leakage was unchanged (still about 0.6uA). ESR was
slightly *improved*, down from 0.49 ohms to 0.45 ohms - probably not
measurement error, as there was a distinct trend in measurements.
Capacitance was very slightly *increased*, up from 18.7uF to 19.0uF. (The
cap was nominally 15uF 25F, rated for 85C. Manufacturer was Nichicon.)

If nothing else, this one data point suggests that this particular Nichicon
cap was conservatively rated.

Since I have also seen 16V caps used in audio DI outputs, I just tried the
same setup with a 16V 33uF cap; but with 48V across that and the same 6.8k
resistor, leakage current was around 1.2mA and slowly climbing after two
minutes, so I shut it down. That experiment will have to wait until I can
clear a bit more area around my "blast shield" and attend it more closely...

 

[R.Legg]

After two days at 48V, leakage was unchanged (still about 0.6uA). ESR was
slightly *improved*, down from 0.49 ohms to 0.45 ohms - probably not
measurement error, as there was a distinct trend in measurements.
Capacitance was very slightly *increased*, up from 18.7uF to 19.0uF. (The
cap was nominally 15uF 25F, rated for 85C. Manufacturer was Nichicon.)

If nothing else, this one data point suggests that this particular Nichicon
cap was conservatively rated.

Since I have also seen 16V caps used in audio DI outputs, I just tried the
same setup with a 16V 33uF cap; but with 48V across that and the same 6.8k
resistor, leakage current was around 1.2mA and slowly climbing after two
minutes, so I shut it down. That experiment will have to wait until I can
clear a bit more area around my "blast shield" and attend it more closely...

Below a certain can size, it may not pay to specifically manufacture a
lower voltage type. Just change the label to fit the bin type.

This was the case with Philips and Siemens back in the 80's. I assume
other manufacturers run with the same economies in mind.

RL

 

[Roger Gt]

: Martin Schönegg wrote:
:
: > Hi Walther,
: >
: > | I'm pretty perplexed. Obviously the leakage current is very
low; so
: > at this
: > | point (now that the cap is charged) it wouldn't make any
difference
: > whether
: > | the 6.8k resistor was there or not. Perhaps it would make a
: > difference if I
: > | discharged the cap and then hit it with 48V with no current
: > limiting. The
: > | ESR of the cap is about half an ohm, so it would certainly
dissipate
: > a lot
: > | of power for a little while. Maybe WV basically amounts to
a surge
: > power
: > | rating - in this case, about 1kW for about half a usec.
: >
: > You formed this 25 V Cap to 48V one, thats it. The Capacity
will be a
: > little lower than the original type. Thats all. Every
manufacturer does
: > this step to create the dielectricum. The more voltage you set
up the
: > thicker the dielectricum will be. If you discharge the cap
this is
: > stable, if you change the polarity the dielectricum will
shrink until
: > it is gone. By forming and disforming it depends on the power
behind
: > this process. Very old electrolytics in devices that where not
used
: > over several years may explode on their first contact with the
power
: > line. There is no effective current limiter before these Caps
so they
: > make a big BANG. If you have a resistor in series the current
may be
: > limited enough -> No BANG.
: >
: > MArtin
:
: I once either saw or hallucinated a construction project for a
home-made
: electrolytic make of a 5-gallon tub and a stack of aluminum pie
tins.
: The tins were separated, of course. I don't remember what they
said the
: electrolyte was, but it did sound like a way fun project.
:
: Cheers!
: Rich
:

Really old HAM publication. Used marbles to separate the pie
pans. I also do not remember the electrolyte.

 

[Walter Harley]

Below a certain can size, it may not pay to specifically manufacture a
lower voltage type. Just change the label to fit the bin type.

This was the case with Philips and Siemens back in the 80's. I assume
other manufacturers run with the same economies in mind.

Maybe so. But I'm pretty sure that for a 15uF cap (what I was trying), the
25V and 50V can sizes are different. I'm not glancing at a catalog or mfr
data sheet as I say that, so perhaps I'm wrong.

Over time they appear to have improved the technology, so that caps of the
same rating are much smaller these days. Perhaps when that happens there is
a transitional time when the caps of a certain physical size are
under-rated.

 

[legg]

Maybe so. But I'm pretty sure that for a 15uF cap (what I was trying), the
25V and 50V can sizes are different. I'm not glancing at a catalog or mfr
data sheet as I say that, so perhaps I'm wrong.

Over time they appear to have improved the technology, so that caps of the
same rating are much smaller these days. Perhaps when that happens there is
a transitional time when the caps of a certain physical size are
under-rated.

I note in the Nippon Chemicon (~United Chemicon) catalog that case
sizes for 15uF parts rated 16, 25 and 35V still typically sport the
same case size.

Perhaps when you refer to the catalog for the type that you've been
'ass'essing, you'll find the same general tendency.

RL

 

[Walter Harley]

I note in the Nippon Chemicon (~United Chemicon) catalog that case
sizes for 15uF parts rated 16, 25 and 35V still typically sport the
same case size.

Perhaps when you refer to the catalog for the type that you've been
'ass'essing, you'll find the same general tendency.

The part I was using was drawn from my junk bin - remember, I expected it
might explode, didn't want to use a good one

Glancing through the Mouser catalog, the pattern I see is that for
particular cap types, there seems to be a "smallest size"; above that,
voltage or value makes a difference, but there's a threshold (say, 5x11mm)
below which all the caps of that type are the same physical size. Some 50V
caps seem to fit under the threshold, for some types of cap. Not
surprisingly, quantity price appears to be related more closely to case size
than to value.

The cap that I was testing (Nichicon) was probably small enough to be below
that threshold.

I remain interested by the fact that 50V-rated components aren't routinely
used in an application in which the cap can be expected to often be exposed
to 48V; but the rest of the mystery is clearing up. Thanks for your help.

 

[Tam/WB2TT]

The part I was using was drawn from my junk bin - remember, I expected it
might explode, didn't want to use a good one

Glancing through the Mouser catalog, the pattern I see is that for
particular cap types, there seems to be a "smallest size"; above that,
voltage or value makes a difference, but there's a threshold (say, 5x11mm)
below which all the caps of that type are the same physical size. Some 50V
caps seem to fit under the threshold, for some types of cap. Not
surprisingly, quantity price appears to be related more closely to case size
than to value.

The cap that I was testing (Nichicon) was probably small enough to be below
that threshold.

I remain interested by the fact that 50V-rated components aren't routinely
used in an application in which the cap can be expected to often be exposed
to 48V; but the rest of the mystery is clearing up. Thanks for your help.

Walter,
Actually, when you factor in manufacturing variances, temperature, current,
and product liability concerns, 100% over rating does not seem unreasonable.
I had measured the access times of DRAMs a few years ago, and all operated
about 2X better than spec. I have also measured transistor reverse Vbe and
Vce, and 90% beat the spec by 2X - except for one unit that passed by 1V.

Tam

 

[mike]

The part I was using was drawn from my junk bin - remember, I expected it
might explode, didn't want to use a good one

Glancing through the Mouser catalog, the pattern I see is that for
particular cap types, there seems to be a "smallest size"; above that,
voltage or value makes a difference, but there's a threshold (say, 5x11mm)
below which all the caps of that type are the same physical size. Some 50V
caps seem to fit under the threshold, for some types of cap. Not
surprisingly, quantity price appears to be related more closely to case size
than to value.

The cap that I was testing (Nichicon) was probably small enough to be below
that threshold.

I remain interested by the fact that 50V-rated components aren't routinely
used in an application in which the cap can be expected to often be exposed
to 48V; but the rest of the mystery is clearing up. Thanks for your help.

Nobody has asked the obvious question,
"Why would anyone want to do such a (insert your favorite derogatory
here) thing?"

Safety agencies would never approve it in a product.
If you're doing a few, the testing/verification costs will exceed
any possible cost savings.
Why would you build a hobby project that required you to wear safety
glasses all the time?

Back in the old days, we had a product that had huge caps bolted to
an aluminum plate. There was a hole in the plate where the vent was.
Well, eventually, someone put a cap in backwards. That reversed the
voltage and blocked the vent. The capacitor hit the ceiling and then the
$#|+ hit the fan...
mike

--
Return address is VALID.
Wanted, 12.1" LCD for Gateway Solo 5300. Samsung LT121SU-121
Bunch of stuff For Sale and Wanted at the link below.
Compaq Aero floppy,ram,battery.
MINT HP-41CV, 2-METER AMPS, 200CH SCANNER
http://www.geocities.com/SiliconValley/Monitor/4710/

 

[legg]

Safety agencies would never approve it in a product.
If you're doing a few, the testing/verification costs will exceed
any possible cost savings.

Although safety agencies would not reject such a low-power
missapplication, if no demonstrated hazzard occurred in evaluation,
the costs of testing required to demonstrate this places it outside of
practical budgets for standard product development.

Why would you build a hobby project that required you to wear safety
glasses all the time?

Back in the old days, we had a product that had huge caps bolted to
an aluminum plate. There was a hole in the plate where the vent was.
Well, eventually, someone put a cap in backwards. That reversed the
voltage and blocked the vent. The capacitor hit the ceiling and then the
$#|+ hit the fan...

Even used as prescribed, parts can and do fail, not neccessarily with
spectacular results. I recall a type of enclosed 1500W redundant
converter module with two capacitors in parallel in its primary PFC
boost stage. One module was returned after more than six months in the
field because a technician noticed an occassional puff of smoke vented
by the self-contained air movers.

One of the bulk caps had blown a 4mm hole in it's side, and on later
examination, you could see that fully 30% of its internal volume had
been gradually eaten away. The unit tested and ran normally in all
respects with this part in place. The failure process could have
commenced at any time after the lid went on.

Even had the failure caused some kind of output glitche - in the fully
redundant application there would have been no indication at the
system level (that was no accident).

I recall another model that went into production (before I started to
work at that location) with art polarity errors in a small (and
apparently totally useless) radial position. This went 'undetected'
until I had to use the same art for another purpose. No big kabooms
there - but no medals were awarded for that fact, believe me.

Defective parts are also not completely unheard of, and the results
can be spectacular and odiferous. Production testing generally demands
a special consideration for this fact for at least the first
application of power. Repeated misbehaviour in the record, for any
position or part type, requires action to quarantine and track down
the problem.

RL

 

[Spehro Pefhany]

Safety agencies show no interest in the low-voltage stuff or the guts
of the electronics, and generally never see the schematics. They
certify a product based on the AC line insulation and fusing and
grounding and stuff.

John

We were told by a guy at (safety agency) not to bother sending the
schematics because their engineers would just waste a whole lot of
time studying them...

Best regards,
Spehro Pefhany

 

[John Larkin]

Safety agencies would never approve it in a product.
If you're doing a few, the testing/verification costs will exceed
any possible cost savings.
Why would you build a hobby project that required you to wear safety
glasses all the time?

Safety agencies show no interest in the low-voltage stuff or the guts
of the electronics, and generally never see the schematics. They
certify a product based on the AC line insulation and fusing and
grounding and stuff.

John