Use bleeder resistors in audio opamp design?

[Grant]

I'm making some audio filter circuits, using 1% resistors and capacitors.

Capacitors only available in 22n, 33n or 47n cheaply, so I use them in
combinations to get a reasonable impedance points. One filter would
use three caps in series -- what happens to the DC isolated nodes on
the series connected capacitors here?

Would you put a high value resistor, 10M or something, across caps in
series? I recognise the need to balance DC midpoint with electros in a
power supply, but what of signal nodes around an opamp?

Around one opamp I have three in series caps to ground, two in series
on the feedback (8.25n and 16.5n from 22n and 33n caps).

Anyone have a horror story of what happens if one leaves these nodes
'floating' DC wise. Or do nothing at all, board leakage will fix it?

Thanks,
Grant.

 

[Phil Allison]

"Grant"

I'm making some audio filter circuits, using 1% resistors and capacitors.

Capacitors only available in 22n, 33n or 47n cheaply, so I use them in
combinations to get a reasonable impedance points.

** Where do yo get your 1% tolerance 47nF caps ??

One filter would
use three caps in series -- what happens to the DC isolated nodes on
the series connected capacitors here?

** Nothing much - it will acquire a voltage near to zero DC.

I recognise the need to balance DC midpoint with electros in a
power supply,

** Many designers do not like to see electros with more than rated DC volts
across them - so they add " ballast " resistors to swamp electro leakage
in series strings. But long as the electros are the same type and age and
have a reasonable margin of voltage over that needed - such resistors
hardly matter.

I regularly put 350 volt electros in series and run them across a 500 volt
rail with no ballast.


..... Phil

 

[Rich Grise]

I'm making some audio filter circuits, using 1% resistors and capacitors.

Capacitors only available in 22n, 33n or 47n cheaply, so I use them in
combinations to get a reasonable impedance points. One filter would
use three caps in series -- what happens to the DC isolated nodes on
the series connected capacitors here?

Would you put a high value resistor, 10M or something, across caps in
series? I recognise the need to balance DC midpoint with electros in a
power supply, but what of signal nodes around an opamp?

Around one opamp I have three in series caps to ground, two in series
on the feedback (8.25n and 16.5n from 22n and 33n caps).

Anyone have a horror story of what happens if one leaves these nodes
'floating' DC wise. Or do nothing at all, board leakage will fix it?

If the total DC voltage end-to-end of the series string is less than
the voltage rating of the individual caps, then it shouldn't be necessary
to do any "balancing: - let each cap develop as much voltage as it needs,
and don't worry about artificially "balancing" them.

If it's such a critical circuit that ESR becomes an issue, then that's
beyond my purview, and you'd need a "real" engineer.

Hope This Helps!
Rich

 

[Grant]

"Grant"

** Where do yo get your 1% tolerance 47nF caps ??
http://australia.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=1666465

** Nothing much - it will acquire a voltage near to zero DC.


** Many designers do not like to see electros with more than rated DC volts
across them - so they add " ballast " resistors to swamp electro leakage
in series strings. But long as the electros are the same type and age and
have a reasonable margin of voltage over that needed - such resistors
hardly matter.

I regularly put 350 volt electros in series and run them across a 500 volt
rail with no ballast.

Usually present in PC power supplies where they'll put a couple 200VDC in
series for the rectified mains, 150k is common. Wide variation.

Grant.

 

[Phil Allison]

"Rich Gripe"

If the total DC voltage end-to-end of the series string is less than
the voltage rating of the individual caps,

** How ambiguous.

then it shouldn't be necessary
to do any "balancing: - let each cap develop as much voltage as it needs,
and don't worry about artificially "balancing" them..

** Where a series string of poly film caps must withstand a DC voltage that
is comparable with the SUM of the individual cap voltage ratings - the use
of balancing resistors is almost essential.

The alternative is a kind of Russian roulette to decide which cap will be
overvoltaged enough to make it suffer an internal, insulation failure
discharge next.

Snap, crackle, pop.

If it's such a critical circuit that ESR becomes an issue, then that's
beyond my purview, and you'd need a "real" engineer.

** Instead of a know nothing jerk off like you.



.... Phil

 

[Phil Allison]

"Grant"


** Do you ever take any notice of what folk say to you ??

No reply required....



.... Phil

 

[Rich Grise]

"Grant"

** Do you ever take any notice of what folk say to you ??

Do you?

Thanks,
Rich

 

[Rich Grise]

"Rich Grise"

** How ambiguous.

What part don't you understand?

** Where a series string of poly film caps must withstand a DC voltage
that
is comparable with the SUM of the individual cap voltage ratings - the
use of balancing resistors is almost essential.

The alternative is a kind of Russian roulette to decide which cap will be
overvoltaged enough to make it suffer an internal, insulation failure
discharge next.

Snap, crackle, pop.

If you've got three capacitors, each with a breakdown voltage of, say, 100V,
in a circuit where the maximum total voltage across the string is, say, 50V,
which one gets the overvoltage?

Where do the extra volts come from?

Thanks,
Rich

 

[Phil Allison]

"Rich Gripe"

What part don't you understand?

** How ambiguous.....

If you've got three capacitors,

** Then you can tell a funny capacitor joke.

Hey:

No fuckwit hypotheticals .

Ever.

You stinking TROLL.


.... Phil

 

[Rich Grise]

Film caps can go from near zero leakage to punch-through shorted in
microseconds, so it's not safe to let them balance themselves.
Electrolytics and ceramics generally have greatly increasing leakage
at or above rated voltage, so strings self-balance. Startup transients
must be considered, of course.

What kind of insane voltages are you using in your opamp circuits where
you could punch through a series string of 100V caps? For punch-through,
from where are all those volts coming?

Thanks,
Rich

 

[Grant]

I'm not doing that around an opamp circuit

Film caps can go from near zero leakage to punch-through shorted in
microseconds, so it's not safe to let them balance themselves.

But at low voltage/power levels? These are 160VDC, 100VAC caps in a
nominal 1V level audio signal chain. Perhaps I should have mentioned
the caps I have are polypropylene film type 1837 MKP and although
there's not much power involved around an opamp, I simply wonder if
there's some effect that would be heard? Like a hissing or crackling
as the nodes balanced themselves over board leakage, or whatever.

Electrolytics and ceramics generally have greatly increasing leakage
at or above rated voltage, so strings self-balance. Startup transients
must be considered, of course.

I've experimented a little with polymer aluminum caps. They have very
low leakage up to roughly 2x rated voltage, then die suddenly. So they
probably don't equalize and aren't very well suited for series-string
use. Pity.

Not used polymer electros. Replaced a lot of low ESR electros in PC
related gear (mobos and monitors) that seem top be selected at too low
a rating for the job. There I try to fit higher voltage caps, but that's
a different topic.

Another OT: I've replaced failed mains 400V electro with a pair of series
electros and ballast/balance resistors, to fit the space of the blown
original. That monitor still running too after couple or more years.

Grant.

 

[Grant]

Shouldn't be a problem... just let them float. Though, for
simulation, you'll need paths to ground


If your OpAmps "feature" class-B outputs you may want to bleed enough
load current to force the stages into class-A. I've experienced
problems with some brands of LM324 having enough dead-band to produce
wild spurs in active filters.

Thank you, I did discover that in the early 80s when designing and
prototyping a new portable LVDT readout unit. The LM324s were great on
12V single supply, apart from needing that resistor to ground to kill
the easily seen crossover distortion.

For these filters I plan to use AD847, it's all high level audio signal
and the opamps in follower mode (S-K filters). I bought a fair few at a
decent price and individual opamps easier to force into class A than a
quaddie on +/- 15V, for dissipating the extra heat?

Probably use current mirrors for constant current sink to -15V. There I
wonder at signal cross-coupling via the current sinks, maybe I should
plan on individual sinks? This is not a production design where I have
to minimise component count.

Years ago (1981/2) I designed a LVDT portable readout unit, used the
'proper' synchronous rectifier and related stuff to get measurement
performance, and the LM324 on 12V single supply.

Discovered the crossover distortion for myself and the obvious cure of
loading the output with resistor to 0V, since they have that large source,
lower sink current output design, just begging for more current to ground
when run on single supply. Turned out to be the most accurate instrument
the boss used, and first that showed an LVDT's 'S' curve, quite nice

(LVDT: Linear Variable Displacement Transducer, not the modern signaling
technique

Grant.

 

[Grant]

Why is Phil such an asshole anyway? The few posts I've looked at
recently rapidly turn into an argument with Phil being, by far, the
least civil and throwing around the most four letter words.

Statistically speaking, it seems highly unlikely he's the rational one
here, while everyone else has behavior and anger issues.

Phil is himself, and doesn't stand nonsense, a bit overboard in his
responses, but I worked on building sites as a teenager, the language
doesn't fuss me. Some people have short fuses (fuzes?), I don't have
to buy into that.

In this particular case I dunno if he was impatient for my reply to
his query (where to buy 1% 47n, which I answered) or what?

I don't see what else he's fussed about unless it's because I don't
see the point in replying to posts I agree with, nor do I see a point
responding to Rich G. when he was talking about a different point to
what I'm interested in. I find myself quite busy lately (surprisingly
for someone left the workforce in '93), so I don't spend a lot of
time here.

It's nice to pop in the odd questions and see responses, but I'm not
here to play flame games. Unless I am

Grant.

 

[Grant]

[about forcing op amps into class A operation with output current bias]

Probably use current mirrors for constant current sink to -15V. There I
wonder at signal cross-coupling via the current sinks

If you need a lot of sinks, it's easy to generate another, circa -24V,
supply, and strap base of NPN to (-15V), then an emitter resistor
to the (-24V) programs the sink current of the collector.

Interesting thought. Worth a look at.

Since these currents are regulated, there oughtn't be any
ripple due to load at the (-24V) supply. Also, the sinks are better
compliance at the (-) rail than most op amps; your source-versus-sink
decision might be guided by the improved behavior-near-rail
that this offers.

Except the next available higher voltage is from power amplifiers, very
ripply, so there's some interesting stuff for me to think about in there.

For the cautious, a base resistor can be added to eliminate all-silicon sneak
paths.

That too.

Grant.

 

[Grant]

So, what's your opinion of using electros back to back to form non-
polarized caps?

What's wrong with using bipolar caps, the leakage? I've never liked
back to back as they sort of expect diode action to keep the electros
nice, so what does that non-linearity do to a signal? Or the cap itself
after a while.

Then there's all those circuits out there with little, none or slight
reverse bias on tantalums in audio signal paths, how do they sound
after a while?

Grant.

 

[Phil Allison]

"Grant"


** Thanks - they were formerly very hard to get items in Australia.

Farnell had some ( expensive) 1% polystyrenes up to 10nF and that was it.


..... Phil

 

[Phil Allison]

"Grant"

In this particular case I dunno if he was impatient for my reply to
his query (where to buy 1% 47n, which I answered) or what?

** My oversight, now corrected, in not replying.

I don't see what else he's fussed about unless it's because I don't
see the point in replying to posts I agree with,

** You have a reading comprehension problem.

The concept of "context' seems beyond you.



...... Phil

 

[Phil Allison]

"Bill Bowden"

So, what's your opinion of using electros back to back to form non-
polarized caps?


** Ask the NG, not me - shithead.

 

[Grant]

That's a huge number of parts, most of which affect gain and offset. I
like S-K filters because they can have a DC gain of exactly 1,
independent of any resistor values.

Too right

There's more practical factors in my mind, like building a well known
topology that will work first time.

Grant.

 

[Grant]

....

SK filters are for people that don't know how to design filters.

That's why I'm using a couple windoze apps to do it for me

Anyway, as I was uncomfortable with series caps I ordered some 10n today
for where I was going to series up to get lower value. I might even use
the remainder of the 50 caps one day on ADC/DAC stuff.

I did say up front I'm processing audio, though I do like the DC unity
gain because I'll possibly use the low pass on data, dunno yet. Direct
coupled is good too. Better than back to back electros.

If I was a tad brighter, younger and better at maths I'd look further
into other topologies, but you need to try harder to scare me into
doing more than taking a brief glance at your leapfrog whatsits
I never heard of them before. I used to work in industrial electronics,
was called on to make a filter card once, multistage low pass thing
with lots of trimpots, yuck. Stayed out there so it must've worked.

That was DC accurate stuff for signal conditioners too. So there's my
bias showing. If it don't work I'll chalk it up as lesson learned, be
here back with a sad story to tell.

Grant.