capacitor

[ivan]

I have short circuited R and saw an increase of 15Vpp. By changing the cap
type I see from worst (50Vpp) to best (120Vpp). Would ESR variance be high
enough? What are the typical values of ESR? Unfortunately I do not have the
means to do a proper study on this here...

---
Earlier you stated that your circuit looked something like this:
(View in Courier)

Er
/
+----[7.37E-4H]--+--[2.2E-9F]--+--[22R]----+
| |
[GEN1] [GEN2]
| |
GND GND


Where GEN1 and GEN2 are square-wave sources with their outputs 180
degrees out of phase and Er is the voltage measured, to ground, from
the junction of the capacitor and the resistor.

From the values you've given, the circuit is resonant at:

1 1
f = -------------- = ------------------------------------
2pi sqrt(LC) 6.28 * sqrt (7.37E-4H) * (2.2E-9F)


~ 125kHz.


and the reactances of the capacitor and inductor will be:


Xc = Xl = 2pi fL = 6.28 * 1.25E5Hz * 7.37E-4H ~ 578 ohms


Now, since the signs of the reactances are opposite to each other,
at resonance they'll cancel and you'll be left with only the
resistance of the 22 ohm resistor, the resistance of the wire in the
inductor and the equivalent series resistance (ESR) of the capacitor
to limit the current through the circuit. Just for grins, if we
assume an ohm for each, the total series resistance in the circuit
will be 24 ohms.

Now, in order to determine the current in the circuit we'll need to
determine its impedance, thus:


Z = sqrt (R² + (Xl - Xc)²)


= sqrt (24² + (578 - 578)²)


= 24 ohms


Assuming you're driving your circuit with 5V sources, the maximum
current you'll be able to pump through the 24 ohms will be:

E 5V
I = --- = ----- = 0.208 ampere
R 24R

but because the reactance of the capacitor is 578 ohms, the voltage
dropped across it will be a startling:


E = IR = 0.208A * 578R ~ 120V

ok I understand, It looks like this capacitor has a 2R series resistance
then.
and the other which swings at 80V has a series resistance of 14.125R as:

80V / 578R ~ 0.138A

5V / 0.138A = 36.125R

36.125R - 22R = 14.125R

yeah I see what you are saying now...

Now, assuming that your generators can deliver the current and you
short out the 22 ohm resistor, the reactances will still cancel and
we'll be left with 2 ohms of resistance to oppose the current. That
means the current in the circuit will be:

E 5V
I = --- = ---- = 2.5 ampere
R 2R

and since the reactances haven't changed, the voltage dropped across
the capacitor will be:


E = IR = 2.5A * 578R = 1445V.

I get about 160Vpp here actually...

for those who asked, the complete circuit is roughly as shown at
http://www.atmel.com/dyn/resources/prod_documents/doc4684.pdf
in Figure 4-1. The signal and the resonant branch I refer to are COIL1 and
COIL2...

So, it looks like just shorting out the resistor will get you a
voltage increase of:


1445
------ = 12 = 1200%
120

There are also some second order effects like the inductance of the
inductor changing as the current (and frequency) through it changes,
and the parametric effects of dielectrics causing capacitance change
with voltage which will spoil resonance.

Of course there are always the primary effects, which in your case
involve the capacitance tolerances of the capacitors you're using
and their ESR's (losses) which could easily explain the results
you've been getting.


Here's an LTSPICE circuit list of your circuit which you may find
interesting:

ok I'll try this...thanks!

 

[ivan]

AND he better tell us what kind of cap meter he's using. Many $4.99
DVM's from the corner gas station have a cap measuring feature, but
they are actually measuring how quickly it oscillates in a RC
oscillator. Said frequency is only vaguely related to the actual
capacitance, especially for lossy or leaky capacitors.

it is a fluke 185, it has three digits after comma when measuring nF...

 

[ivan]

E 5V
I = --- = ----- = 0.208 ampere
R 24R



E = IR = 0.208A * 578R ~ 120V

I changed the resistor with a 18R so I would expect 144.5V swing, but I get
only a few volts increase... why?

ivan

 

[John Fields]

I changed the resistor with a 18R so I would expect 144.5V swing, but I get
only a few volts increase... why?

---
Dunno...

You might be off resonance, your cap may have a higher ESR than you
think it does, the coil's resistance might be higher than you think
it is, and/or the impedance of your scope probe might be loading
down the circuit.

Also, take a look at the spec's for the coil drivers in the U2270.

They're rated for an absolute maximum of 200mA out, but rated for a
typical output voltage of 4V with a 100 mA load, so you'll be
running them pretty much balls-to-the wall with that 22 ohm resistor
in there, (neglecting the real part of the LC impedance) and
certainly you'll be bottoming them out when you short out the
resistor!

 

[ivan]

---
Dunno...

You might be off resonance, your cap may have a higher ESR than you
think it does, the coil's resistance might be higher than you think
it is, and/or the impedance of your scope probe might be loading
down the circuit.

Also, take a look at the spec's for the coil drivers in the U2270.

They're rated for an absolute maximum of 200mA out, but rated for a
typical output voltage of 4V with a 100 mA load, so you'll be
running them pretty much balls-to-the wall with that 22 ohm resistor
in there, (neglecting the real part of the LC impedance) and
certainly you'll be bottoming them out when you short out the
resistor!

60-70Vpp is safe enough then...thanks!

 

[Ancient_Hacker]

it is a fluke 185, it has three digits after comma when measuring nF...

Nice meter, BUT:

Most meters have better accuracy on voltage ranges than on cap ranges.
So all those digits on the display don't mean the cap readings are
that accurate.

The manual doesnt seem to mention exactly HOW it measures capacitance.
If it uses a RC oscillator, as most of them do, rather than an actual
bridge, all bets are off. It's not measuring capacitance so much as
charge transfer rates. If you put a battery or resistor or inductor
across the terminals, it will *still* read out a capacitance value.
That should be a clue it's not exactly the world's best cap meter.

The manual states the accuracy isnt so good on the cap ranges, like
plus or minus one percent plus or minus 5 counts.

If you really want to investigate the properties of capacitors, get
yourself a real cap meter, one that can show actual capacitance and
dissipation factor. The better ones can show the characteristics to
like four decimal places. A good old HP 4xxx goes for not very many
bucks, has neat nixie-tube readouts, and won't fool you.

 

[jasen]

I have found this capacitor that perfoms much better than the others I have
tried, so I am interested to know what are its characteristics in order to
confirm my guess about the type of capacitor needed. That triangle must be
some kind of logo so I was hoping that someone knew where it is from.

http://bos.hack.org/ic_logos/

Bye.
Jasen

 

[Michael A. Terrell]

But remember to use a non-magnetic screwdriver

What good would that do? You need a non metallic tip to keep from
changing the capacitance, unless the rotor is grounded. These are
commonly called "Alignment tools" by professionals, and "Diddle Sticks"
by people who don't know any better.


--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

 

[Rich Grise]

What good would that do? You need a non metallic tip to keep from
changing the capacitance, unless the rotor is grounded. These are
commonly called "Alignment tools" by professionals, and "Diddle Sticks"
by people who don't know any better.

Also "tweakers". ;-) I had a cow-orker once who kept losing his, so he'd
grab my X-acto knife and break off the tip, and use it. This was only
for trimpots, but still! My X-acto knife! >:-(

;-)
Rich