Searching \ for '[EE] Capacitor-selection' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=capacitor+selection
Search entire site for: 'Capacitor-selection'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Capacitor-selection'
2007\03\31@041521 by Rikard Bosnjakovic

picon face
For my next project I've just built a PSU using a small
9V-transformer. Hooking it up to a Graetz-bridge for rectification, I
measure 9.5 volts over Vcc/GND using my DMM.

Now I'm going to add filter caps. What I'm unsure of here is if
10v-ratings will work, or if 0.5 volts is a way too small safety
margin. I have big (~4000uF) 10v-caps in my junkbox, and since the
junkbox serve a purpose (using older but working components) I feel
that I want to pick those instead of new 16v-rated caps instead.

So, will 10v-ratings work or will it be a too narrow margin?

The PSU will be connected to a 7805, so I'm not expecting too much
variations (except for perhaps inferior) in the voltage.


--
- Rikard - http://bos.hack.org/cv/

2007\03\31@043753 by wouter van ooijen

face picon face
> So, will 10v-ratings work or will it be a too narrow margin?

No, it is way below what you need.

- 9V AC gives (at most, which is what you want to calculate with) 9 x
sqrt(2) = ~ 12.7 V
- an unloaded small 9V transformer can give much more than 9V (measure
or find the specs), calculate with at least 12V, maybe even 18V
- mains has a wide margin. I don't know your countries specs, but I
would calculate with at least -20 +20 %, maybe (much) more
- elco's don't like too high volatge. If my max-max calculations would
yield 15V I would never choose a 16V, maybe not even a 22V.

> The PSU will be connected to a 7805, so I'm not expecting too
> much variations (except for perhaps inferior) in the voltage.

I don't get that. An 7805 will limit the output variations, but it can't
infuence its input voltage (except by doing weird things when it is not
decoupled, or too high, or too low).

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu



2007\03\31@050225 by Jinx

face picon face
> For my next project I've just built a PSU using a small
> 9V-transformer. Hooking it up to a Graetz-bridge for
> rectification, I measure 9.5 volts over Vcc/GND using my
> DMM

9VAC will give you more like 13VDC. I suspect your meter is
not showing you the true value, especially as you have no filter
caps. If you were to look at it on an oscilloscope I'm sure you'd
see much higher peaks than 9.5V

> I have big (~4000uF) 10v-caps in my junkbox,.....want to pick
> those instead of new 16v-rated caps instead

You can put two in series to get a 20V 2000uF capacitor

http://www.play-hookey.com/dc_theory/series_capacitors.html

> The PSU will be connected to a 7805

If you're experimenting, a diode from 7805 output to input is
a good idea, and recommended. Anode goes to o/p, cathode
to input. See the 7805 datasheet, Figure 1, page 11 (980kB)

http://www.national.com/ds/LM/LM340.pdf

In case you accidentally short the input, any filter cap you have on
the o/p holds the voltage on the o/p pin higher than the 0V now on
the i/p pin. This is a bad thing and can blow the chip. With a diode
there, the o/p cap can discharge into the short on the i/p

2007\03\31@050814 by Rikard Bosnjakovic

picon face
On 3/31/07, wouter van ooijen <spam_OUTwouterTakeThisOuTspamvoti.nl> wrote:

> - an unloaded small 9V transformer can give much more than 9V (measure
> or find the specs), calculate with at least 12V, maybe even 18V

Using my Fluke 179 (RMS-DMM), it said 9.6VAC for the unloaded
transformer (no bridge). Measuring over "+" and "-" on the bridge, it
said 9.5VDC. Adding a load (100k) over the bridge yielded no change
for the voltage (9.5VDC still).

Could it be the transformer is marked with its effective voltage? Else
I should have seen sqrt(2)*9 on my DMM, no?


--
- Rikard - http://bos.hack.org/cv/

2007\03\31@053520 by Rich

picon face
Is it possible that you have made a wiring error?  
----- Original Message -----
From: "Rikard Bosnjakovic" <.....rikard.bosnjakovicKILLspamspam@spam@gmail.com>
To: "Microcontroller discussion list - Public." <piclistspamKILLspammit.edu>
Sent: Saturday, March 31, 2007 5:08 AM
Subject: Re: [EE] Capacitor-selection


{Quote hidden}

> --

2007\03\31@055011 by Jinx

face picon face
> Adding a load (100k) over the bridge yielded no change for the
> voltage (9.5VDC still)

100k isn't much of a load really. Even the very smallest chassis-
mount transformer would be perhaps 0.1VA. 100k is loading it
to barely 1/100th of that. IOW it should be able to hold the voltage
up with a 1k or less load

The 9VAC is likely to be the output voltage at the stated VA output,
(although the actual manufacturer's ratings are often over-optimistic)
meaning the unloaded pk-pk voltage could be much more. For example
I use a nominally-rated 12VAC 0.5A plug-pack that puts out 22VAC
unloaded. Not until you start drawing some real current does the
voltage drop closer to 12V. This is hazard to watch out for if the
circuit you intend to power does not load the transformer enough.

All that said however, you did say you'd measured 9.5VAC. But your
DMM reckons 9.5VDC after the bridge ? Something's not right and
I'd be cautious

If you *are* drawing enough current to pull the voltage down, then
the next issue is that the regulator has to dissipate that wattage as
heat. It may be that you'll need to add a series resistor to drop the
voltage (ie waste some of the difference as heat) before the regulator
if the regulator is eg too hot to touch


2007\03\31@060227 by Vasile Surducan

face picon face
On 3/31/07, Rikard Bosnjakovic <EraseMErikard.bosnjakovicspam_OUTspamTakeThisOuTgmail.com> wrote:
> For my next project I've just built a PSU using a small
> 9V-transformer. Hooking it up to a Graetz-bridge for rectification, I
> measure 9.5 volts over Vcc/GND using my DMM.
>
> Now I'm going to add filter caps. What I'm unsure of here is if
> 10v-ratings will work, or if 0.5 volts is a way too small safety
> margin. I have big (~4000uF) 10v-caps in my junkbox, and since the
> junkbox serve a purpose (using older but working components) I feel
> that I want to pick those instead of new 16v-rated caps instead.

So you're dimensioning the filtering caps using your left ear or your nose ?
:)

Do you have a clear picture what will be happening with your
transformer if you're using a small power one and a high value
filtering cap ?

2007\03\31@062945 by wouter van ooijen

face picon face
> > - an unloaded small 9V transformer can give much more than
> 9V (measure
> > or find the specs), calculate with at least 12V, maybe even 18V
>
> Using my Fluke 179 (RMS-DMM), it said 9.6VAC for the unloaded
> transformer (no bridge).

OK. Now try a gain with a representative load.

> Measuring over "+" and "-" on the
> bridge, it said 9.5VDC. Adding a load (100k) over the bridge
> yielded no change for the voltage (9.5VDC still).

measuring DC without a filter cap makes nos ense.

> Could it be the transformer is marked with its effective
> voltage? Else I should have seen sqrt(2)*9 on my DMM, no?

with a cap, yes.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu



2007\03\31@070818 by Russell McMahon

face
flavicon
face
>> - an unloaded small 9V transformer can give much more than 9V
>> (measure
>> or find the specs), calculate with at least 12V, maybe even 18V

> Using my Fluke 179 (RMS-DMM), it said 9.6VAC for the unloaded
> transformer (no bridge). Measuring over "+" and "-" on the bridge,
> it
> said 9.5VDC. Adding a load (100k) over the bridge yielded no change
> for the voltage (9.5VDC still).
>
> Could it be the transformer is marked with its effective voltage?
> Else
> I should have seen sqrt(2)*9 on my DMM, no?

With no filter capacitor you tend to see a full wave rectified sine
wave - ie the negative half cycles are inverted so you get a series of
positive half sine cycles. The mean DC level of this is less than the
steady DC you'd get with a filter and what your meter reads on a DC
range depends on how it responds to the unexpected AC fluctuations.
There will be some stray capacitance and this plus meter load will
lead to an unknowable end reading. Add even a 10 uF cap and the result
will change.

Aluminium electrolytics should ideally be run at near but below their
rated voltage. A 10 volt capacitor rating is too low in the
application you describe.


       Russell


2007\03\31@071135 by Rikard Bosnjakovic

picon face
On 3/31/07, Vasile Surducan <piclist9spamspam_OUTgmail.com> wrote:

> Do you have a clear picture what will be happening with your
> transformer if you're using a small power one and a high value
> filtering cap ?

Not really, no. I'm aware that the cap will vanish into a puff of blue
smoke, but I don't know how the tranformer will behave or derate.


--
- Rikard - http://bos.hack.org/cv/

2007\03\31@072516 by Tony Smith

picon face
> Now I'm going to add filter caps. What I'm unsure of here is
> if 10v-ratings will work, or if 0.5 volts is a way too small
> safety margin. I have big (~4000uF) 10v-caps in my junkbox,
> and since the junkbox serve a purpose (using older but
> working components) I feel that I want to pick those instead
> of new 16v-rated caps instead.


Use 2 caps in series, that'll give you a 20v 4000uF cap, give you a reason
for having a junk box, and save you having to buy 16v ones.  

As Wouter pointed out, an unloaded (or lightly loaded) transformer can put
out a voltage way above what the nameplate says.  12v-16v wouldn't be too
much of a surprise.

Tony

2007\03\31@073530 by Jinx

face picon face
> > Do you have a clear picture what will be happening with your
> > transformer if you're using a small power one and a high value
> > filtering cap ?
>
> Not really, no. I'm aware that the cap will vanish into a puff of
> blue smoke, but I don't know how the tranformer will behave or
> derate

I think Vasile may be hinting at in-rush current, but I'm not sure
any damage will be done here. For higher currents you might use
a thermistor, or relay+resistor, or current limiter like the LM317,
or just a plain resistor. These moderate current coming out of the
transformer for the short period when it sees a big empty capacitor

2007\03\31@081002 by Dario Greggio

face picon face
Tony Smith wrote:

> Use 2 caps in series, that'll give you a 20v 4000uF cap, give you a reason
> for having a junk box, and save you having to buy 16v ones.  

more likely "2000/20V" :)
--
Ciao, Dario

2007\03\31@092238 by Rich

picon face
And twice the ESR
----- Original Message -----
From: "Dario Greggio" <@spam@adpm.toKILLspamspaminwind.it>
To: "Microcontroller discussion list - Public." <KILLspampiclistKILLspamspammit.edu>
Sent: Saturday, March 31, 2007 8:09 AM
Subject: Re: [EE] Capacitor-selection


> Tony Smith wrote:
>
>> Use 2 caps in series, that'll give you a 20v 4000uF cap, give you a
>> reason
>> for having a junk box, and save you having to buy 16v ones.
>
> more likely "2000/20V" :)
> --
> Ciao, Dario
> --

2007\03\31@104629 by Tony Smith

picon face
Not that ESR is a huge consideration for a 7805...

Ok then, add 2 more capacitors in parallel.  That'll give you a 20v 4000uf
cap (ie adding 2x2000uF, as Dario pointed out), with the ESR back (or close
enough for Government work) where you started from.

Tony

> {Original Message removed}

2007\03\31@105129 by Vasile Surducan

face picon face
On 3/31/07, Rikard Bosnjakovic <RemoveMErikard.bosnjakovicTakeThisOuTspamgmail.com> wrote:
> On 3/31/07, Vasile Surducan <spamBeGonepiclist9spamBeGonespamgmail.com> wrote:
>
> > Do you have a clear picture what will be happening with your
> > transformer if you're using a small power one and a high value
> > filtering cap ?
>
> Not really, no. I'm aware that the cap will vanish into a puff of blue
> smoke, but I don't know how the tranformer will behave or derate.

You didn't even thought it's possible that blue smoke come form the
transformer isn't it ?

1. the voltage capacitor must be dimensioned to be over the maximum
peak voltage when there is no load and the outlet voltage is at
maximum.
Assumin you're living in the Europe, that is around 220V +10% which means 242V.
That will be the maximum standardised  potential at the transformer input.
You should check or compute wich is the secondary voltage  V2 of your
transformer in this situation. The voltage on filtering cap will be
Vcap=0.707 x V2. And you must choose a capacitor with a bigger voltage
than Vcap.

2. the filtering cap is used for "eating" the ripple current after
graetz rectification.
As high capacitance, as low ESR, as high current through filtering cap
So high it could be equal with the transformer's short circuit current
if that one is a very small 2...3W transformer. The result: blue puff
from the transformer and not from the capacitor.
You should dimension the capacitor value in accordance with the load
current and with the maximum ripple you could get at the three
terminal stabiliser.

A 7805 from last generation need minimum 2 V dropout, one from old
generation need 4V dropout (7805 from various manufacturers are not
the same !).
As an example if you have a load of 100mA load and 12V AC at the
transformer output, you could use a very small filtering cap as 100uF
or so after the bridge, because the riple low value on the 7805 input
will be bigger than 7V.
But if you need 1A (assuming the transformer is dimensioned for such currents)
the capacitor must be at least 1000uF because the lowest value of the
input riple should be at least 7V.
There is a simple theory about how to dimension the capacitor based on
the riple value.


'[EE] Capacitor-selection'
2007\04\05@090155 by Howard Winter
face
flavicon
picon face
Rikard,

I'm late to the party as usual, but...

On Sat, 31 Mar 2007 11:08:13 +0200, Rikard Bosnjakovic wrote:

> On 3/31/07, wouter van ooijen <TakeThisOuTwouterEraseMEspamspam_OUTvoti.nl> wrote:
>
> > - an unloaded small 9V transformer can give much more than 9V (measure
> > or find the specs), calculate with at least 12V, maybe even 18V
>
> Using my Fluke 179 (RMS-DMM), it said 9.6VAC for the unloaded
> transformer (no bridge).

Right, so if this is a "True RMS" meter, it would be reading 9.6V RMS - so the *peak* voltage would be 9.6 * sqrt(2), so
way above your capacitors' 10V maximum.  Nobody has emphasised this:  the peak voltage is what matters for a
capacitor's survival.

> Measuring over "+" and "-" on the bridge, it
> said 9.5VDC. Adding a load (100k) over the bridge yielded no change
> for the voltage (9.5VDC still).

But it isn't DC.  Without any capacitance it's full-wave-rectified AC, so a meter set to a DC range won't give you an
accurate result, although it may be close to the RMS voltage depending on the design of the meter, so even then
you're a factor of sqrt(2) too low for calculating the capacitor's rated voltage.

> Could it be the transformer is marked with its effective voltage? Else
> I should have seen sqrt(2)*9 on my DMM, no?

No, unless it's a peak-reading meter (very rare) it will read RMS or something like it.  The transformer markings show
its output as an RMS voltage, and the actual output will vary with the input voltage and the load, but you need to rate
for the peak, not the RMS voltage.  So even if the transformer is spot-on in its output, the peak voltage will be 9 *
1.414, about 12.73V so you need at least a 16V rated capacitor even before considering tolerances on input voltage and
so on.

Cheers,


Howard Winter
St.Albans, England


2007\04\05@091122 by Howard Winter

face
flavicon
picon face
Vasile,

On Sat, 31 Mar 2007 06:51:27 -0800, Vasile Surducan wrote:

>...
> Assumin you're living in the Europe, that is around 220V +10% which means 242V.
> That will be the maximum standardised  potential at the transformer input.

Actually, it used to be this on the Continent (in the UK it was 240V +/- 6%) but a few years ago the nominal voltage in
the UK and on the Continent was harmonised to 230V +/- 10%.  No actual voltages were changed, but the specification
was!  :-)  So the voltage can be anwhere from 207 to 253V and still be within specification.

I normally see around 245V in my house, and I have seen it right up to the specified limit of 253V (usually in the early
hours of the morning).  So if Rikard's transformer is specified for 230V input, it could give 9.9V output with everything is
within spec.  (Reinforcing the conclusion that a 10V capacitor will not have a nice day! :-)

Cheers,


Howard Winter
St.Albans, England


2007\04\05@094631 by wouter van ooijen

face picon face
> Right, so if this is a "True RMS" meter, it would be reading
> 9.6V RMS - so the *peak* voltage would be 9.6 * sqrt(2), so
> way above your capacitors' 10V maximum.  Nobody has
> emphasised this:  the peak voltage is what matters for a
> capacitor's survival.

Did you notice the sqrt(2) in my first reply?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu



More... (looser matching)
- Last day of these posts
- In 2007 , 2008 only
- Today
- New search...