# 90 degree phase shift for capacitor heating

Discussion in 'Electronic Design' started by Industrial Electrician, Apr 14, 2006.

1. ### Industrial ElectricianGuest

I am an electrician doing qualification testing of a 10KW power amplifier.
Internally it has a single phase 220VAC input transformer driving a full-
wave bridge rectifier and then large filter capacitors on the DC line.
The end-use will be 50/60Hz ballast testing.

I need to set up a worst-case test for capacitor heating. Please correct me
if I am mistaken, but I think that the worst case is when the output feeds
a resistive load bank with a sine wave that is phase shifted 90 degrees from
the AC power input.

The amplifier input has a 1 MegOhm input impedence and requires a 10V peak-
to-peak input to reach the rated output voltage.

I also have a 25KW load bank on the output that can be set to draw whatever
current I want it to. I can get a capacitor load bank if needs be but it
would be a hassle.

If I remember correctly from my electronics 101 class, I can use a resistor
and a capacitor to generate a 10VPP (7.07VAC RMS) sine wave with close to a
90 degree phase shift from my 220VAC input.

I am not quite sure how to calculate the resistance and capacitance to get a
7VAC 90 degree phase shifed output from my 220VAC input. What values should
I use? Does the capacitor or the resistor go from the 220VAC to the amp
input? Does the capacitor or the resistor go from the the amp input to
common?

I am connecting at the service entrance where 220VAC GND and 220VAC NEUTRAL
are connected together, and I am seeing less than 1/20th of a volt AC between
them at the amplifier AC power input.

Here is what I am going to test next week unless someone tells me differently:

220VAC GND -> AMPLIFIER CHASSIS
220VAC NEUTRAL -> AMPLIFIER AC POWER INPUT 1
220VAC HOT -> AMPLIFIER AC POWER INPUT 2
220VAC HOT -> 10 Microfarad 300V AC-Rated Capacitor 1
10 Microfarad 300V AC-Rated Capacitor 2 -> 3300 Ohm 25W Resistor 1
3300 Ohm 25W Resistor 2 -> 100 Ohm 10W Rheostat (Potentiometer) CW
100 Ohm 10W Rheostat (Potentiometer) CCW -> 220VAC NEUTRAL
100 Ohm 10W Rheostat (Potentiometer) WIPER -> AMPLIFIER SIGNAL INPUT HIGH
100 Ohm 10W Rheostat (Potentiometer) CCW -> AMPLIFIER SIGNAL INPUT LOW

Will this give me a 7VAC sine wave with close to a 90 degree phase shift
between 220V and Amplifier input?

I also have some 12V Zener diodes (current rating unknown). Should I put two
of them connected back-to-back parallele across Rheostat CW and Rheostat CCW?

2. ### colinGuest

I would stronlgy suggest adding a safety isolation step down transformer to
derive the signal. (for safetys sake obviously)

Colin =^.^=

3. ### Phil AllisonGuest

"Industrial Electrician"

** I take it you mean the DC filter caps inside the amplifier?

** Sounds very likely to be true. Charging peaks coincide with amplifier
output zero crossings and load current peaks will be half way between
charging peaks - creating worst case ripple current in the cap.

** Bet the amp would be unhappy too.

** Please DO NOT DO THIS !!!!

** For safety sake you MUST use an *isolation transformer* between the AC
supply and the amplifier input !!!

The simplest way to get a 90 degree phase shift relative to the AC supply is
to use a transformer with a secondary that has a centre tap and connect that
centre tap to ground. Then you will have two, safe low voltage AC supplies
to use. One will be in-phase with the AC supply and the other in exact anti-
phase.

( I suggest you get a small ( ie 20 or 30VA) *toroidal* transformer with a
15 volt secondary that is centre tapped OR has two 7.5 volt secondaries
connected in series for your testing )

THEN, connect one end of 3.3 kohms resistor to one 7.5 volt supply and one
end of a 1uF ( plastic film ) capacitor to the other, join the remaining
cap and resistor leads, this point creates a 7.5 volt supply that is 90
degrees shifted.

Slight adjustments to the resistor or cap value will get the result spot on
at 50 or 60 Hz.

The problem with using the AC supply as an input signal is the waveform is
not a perfect sine wave - however it will likely be close enough or your
purpose.

Good luck testing the temp rise of the filter caps - make sure to monitor
the air ambient near the caps as well as the actual cap body temp.

........ Phil

4. ### Industrial ElectricianGuest

Not a bit. AC power conversion amplifiers can handle any power factor,
including 100% capacitive, 100% inductive, and nonlinear (cheap) switching
power supplies.
I know how to stay safe. This is in a controlled lab condition.
We do things that we would b=never do in a real-world installation.

In addition, the amplifier input has an input transformer with a
15KV isolation rating.
Anti-phase is 180 degrees, not 90 degrees.
Ah. That is where you are getting the 90 degree shift?
So you are saying I got the component values wrong? That is the part I am
unsure about, but I calculated 10 microfarads, not 1. I believe that the
-3db point (con=rner frequency) has to be at least 1/10 of my 50/60Hz.
72 degree controlled climate, huge fan blowing air through the amp,
capacitor near the inlet. But still, big caps don't dissapate well.

5. ### Phil AllisonGuest

"Industrial Electrician"

** Like hell.

What you describe is a fatal shock hazard to any operator.

** Asinine bullshit.

** You completely missed the safety issue.

** I see you have no patience at all.

Bad thing in a electrician if he wants to stay alive.

** Yep, with anti-phase inputs a simple RC network that would otherwise
produce a *45 degree* phase shift at the equal value frequency jumps to 90
degrees.

** You got the electrical circuit theory wrong.

Yours is the mad sparkys' damn dangerous way to do a very simple job.

** With forced air cooling like that, the DC filter caps surface temp will
barely rise.

........ Phil

6. ### John PopelishGuest

Putting the capacitor to the hot line will give you a nearly 90 degree
phase lead, but will also differentiate any noise spikes on the line
and pass them to the resistor with little attenuation. Connecting the
capacitor at the neutral side will integrate the line voltage, causing
an almost 90 degree phase lag, but will low pass filter any noise,
giving you a much cleaner sine wave signal.

A 10 uF capacitor has about 265 ohms impedance at 60 Hz. As long as
the series resistor has a much higher resistance than that, the phase
shift (across the resistor) will approach 90 degrees.

In order to have about 10 volts appear across the capacitor impedance,
you need about .038 amperes to pass through it. A resistor of about
6000 ohms rated for at least 10 watts would provide about that current.

Then you could connect a 100k pot across the capacitor, and use the
wiper to adjust the output voltage to exactly whatever value you
wanted that any fraction of what is across the capacitor. That high
resistance load will alter the phase shift almost not at all, but will
provide a low enough impedance to drive the 1 meg input impedance with
little sag (measure the final voltage while connected to the input, or
substitute a 1 meg load across the pot to set the voltage).

It would be a very good idea to have an isolation transformer between
the line and this circuit, so the output low side has no voltage with
respect to the amplifier common.

7. ### Industrial ElectricianGuest

I only had time for a quick test before we closed down Friday, but
I did have time to try it both ways, and the noise is much, much
lower doing it the way you suggested. Thanks! My oscilloscope trigger
circuit thanks you as well. ;=)
That makes sense. Just to see if I understand, the above relates to
what I read about the -3DB point because the bigger resistor moves
that frequency down, and if I then lowered the signal frequency I
would not only moving closer to the -3DB point but also lowering the
impedance of the capacitor. Am I getting all of this right?
I see how that works. Just to see that I understand, that also means that
the resistor/capacitor phase shift has a voltage loss, and that someone
who needs the phase shift at the original voltage would need to use another
technique such as an operational amplifier based circuit. Am I getting
this right?
That's a great idea. All of my high wattage pots are much smaller in
resistance but using a parallel pot like you suggest beans that I can
use a regular pot. That's very helpful.
This is a temporary setup on a wooden bench, and I have separate
3 foot lengths of 8 AWG wire connecting chassis, line neutral, and
amplifier common to an earth ground that directly connects to a 12
foot ground rod in moist soil, which in turn connects through a
short length of grounding braid to the rebar in the concrete
foundation and connects through another short length of grounding
braid to the building water pipes. I have tested this earth ground
with 250 amps. I seriously doubt that there will be a noticeable
voltage at anything that I have connected to earth ground.

Isolation transformers are very useful if you are poking around inside
a TV, but when working with high current and high voltage systems, you
are far safer by enclosing everything in an earth-grounded case, having
neutral connected to earth ground instead of floating, and never
energizing the circuit without having everything closed up inside earth-
grounded finger-proof grounded conduits and boxes. If you or "Phil
Allison" disagree, please describe, in detail, the exact path that you
think current will take through the human body from a system using the
industry-standard grounding system I am describing. I am open to being
shown that I am wrong, but not without a good explanation showing why I
and everyone else in my industry is wrong.

8. ### Industrial ElectricianGuest

I only had time for a quick test before we closed down Friday, but
I did have time to try it both ways, and the noise is much, much
lower doing it the way you suggested. Thanks! My oscilloscope trigger
circuit thanks you as well. ;=)
That makes sense. Just to see if I understand, the above relates to
what I read about the -3DB point because the bigger resistor moves
that frequency down, and if I then lowered the signal frequency I
would not only moving closer to the -3DB point but also lowering the
impedance of the capacitor. Am I getting all of this right?
I see how that works. Just to see that I understand, that also means that
the resistor/capacitor phase shift has a voltage loss, and that someone
who needs the phase shift at the original voltage would need to use another
technique such as an operational amplifier based circuit. Am I getting
this right?
That's a great idea. All of my high wattage pots are much smaller in
resistance but using a parallel pot like you suggest beans that I can
use a regular pot. That's very helpful.
This is a temporary setup on a wooden bench, and I have separate
3 foot lengths of 8 AWG wire connecting chassis, line neutral, and
amplifier common to an earth ground that directly connects to a 12
foot ground rod in moist soil, which in turn connects through a
short length of grounding braid to the rebar in the concrete
foundation and connects through another short length of grounding
braid to the building water pipes. I have tested this earth ground
with 250 amps. I seriously doubt that there will be a noticeable
voltage at anything that I have connected to earth ground.

Isolation transformers are very useful if you are poking around inside
a TV, but when working with high current and high voltage systems, you
are far safer by enclosing everything in an earth-grounded case, having
neutral connected to earth ground instead of floating, and never
energizing the circuit without having everything closed up inside earth-
grounded finger-proof grounded conduits and boxes. If you or "Phil
Allison" disagree, please describe, in detail, the exact path that you
think current will take through the human body from a system using the
industry-standard grounding system I am describing. I am open to being
shown that I am wrong, but not without a good explanation showing why I
and everyone else in my industry is wrong.

9. ### Industrial ElectricianGuest

I only had time for a quick test before we closed down Friday, but
I did have time to try it both ways, and the noise is much, much
lower doing it the way you suggested. Thanks! My oscilloscope trigger
circuit thanks you as well. ;=)
That makes sense. Just to see if I understand, the above relates to
what I read about the -3DB point because the bigger resistor moves
that frequency down, and if I then lowered the signal frequency I
would not only moving closer to the -3DB point but also lowering the
impedance of the capacitor. Am I getting all of this right?
I see how that works. Just to see that I understand, that also means that
the resistor/capacitor phase shift has a voltage loss, and that someone
who needs the phase shift at the original voltage would need to use another
technique such as an operational amplifier based circuit. Am I getting
this right?
That's a great idea. All of my high wattage pots are much smaller in
resistance but using a parallel pot like you suggest beans that I can
use a regular pot. That's very helpful.
This is a temporary setup on a wooden bench, and I have separate
3 foot lengths of 8 AWG wire connecting chassis, line neutral, and
amplifier common to an earth ground that directly connects to a 12
foot ground rod in moist soil, which in turn connects through a
short length of grounding braid to the rebar in the concrete
foundation and connects through another short length of grounding
braid to the building water pipes. I have tested this earth ground
with 250 amps. I seriously doubt that there will be a noticeable
voltage at anything that I have connected to earth ground.

Isolation transformers are very useful if you are poking around inside
a TV, but when working with high current and high voltage systems, you
are far safer by enclosing everything in an earth-grounded case, having
neutral connected to earth ground instead of floating, and never
energizing the circuit without having everything closed up inside earth-
grounded finger-proof grounded conduits and boxes. If you or "Phil
Allison" disagree, please describe, in detail, the exact path that you
think current will take through the human body from a system using the
industry-standard grounding system I am describing. I am open to being
shown that I am wrong, but not without a good explanation showing why I
and everyone else in my industry is wrong.

10. ### John PopelishGuest

All but the capacitor part. The impedance of a capacitor foes up as
frequency goes down. Xc=1/(2*pi*f*C)

At the -3db frequency (where Xc=R) the phase shift is 45 degrees, at
frequencies much higher than this, the phase shift approaches 90
degrees. Operating where a capacitor has an impedance of 265 ohms, in
series with a 6000 ohm resistor is far enough above the -3db frequency
that the approach to 90 degrees is pretty close.
Yes. this technique involves an essential voltage loss.
(snip)

Very well.

11. ### Phil AllisonGuest

"Industrial Electrician"

** Your smug arrogance is only exceeded by your sparkies pig ignorance.

Isolation transformers are included inside EVERY electronic and electrical
device where human contact with external wiring is possible - as is the
case with your 90 degree phase shifted "signal generator".

** Irrelevant crapology !!

Using an isolation step-down transformer means you AVOID working with a
dangerous high voltage system in the first place.

** Q. What happens if the link to neutral ever becomes disconnected ?

A. The previous low voltage (ie 90 degree phase shifted supply) jumps to
220 volts AC with plenty of current available to cause electrocution.

NOW:

Stop being such a conceited, bloody ass !!

A small isolation transformer and one R and one C network method is far
CHEAPER and well as enormously SAFER than your asinine scheme.

......... Phil

12. ### Industrial ElectricianGuest

Nothing of any importance. The system will stop working, but the fact
that it is fully enclosed in earth-grounded conduit and boxes makes it
impossible for the now-hot neutral to harm a human. Your rejection of
the basic safety scheme used in all modern electrical wiring (earth
grounded boxes and conduit enclosing a neutral that is earth-grounded
only at the service entrance and a hot that is protected by a fuse or
breaker) is based on ignorance. Your suggestion that an isolation
transformer be added and the neutral be allowed to float is dangerous
and illegal in most municipalities. Stay away from AC power, please;
you are sure to hurt someone with your lack of knowledge and refusal
to learn from those who know what they are talking about.
There already exists a hot wire with 220VAC on it inside the grounded
conduit and boxes. Your belief that an open neutral that creates
another wire with 220VAC on it inside the grounded conduit and boxes
is a safety issue is magical thinking; both hot and neutral are fully
protected from contact with a human, as is the capacitor and resistor
that comprises the phase shift network. Your inability to describe,
in detail, the exact path that you think current will take through
the human body says it all. Until you can describe that path, all your
blustering and magical thinking will only lead you farther astray.
Use of personal attacks to hide an obvious lack of technical rigor in
your claims noted. Bozo Filter now engaged.

Plonk.

13. ### Phil AllisonGuest

"Industrial Electrician"

** WHAT A PIG ARROGANT ASSHOLE !!

** The previous 7 volt AC signal becomes 220 volts and lethal.

** Shame about those connections to the amplifier's input - they are now
lethal for anyone to touch.

Dumbfuck sparkie assholes like YOU have no training in or comprehension of
basic electrical safety.

** Go shove your posturing opinions up your stupid arse.

** Only I did not make that suggestion.

The use of an isolation step down *removes* the need for a connection to
neutral to exist to guarantee safety - since loss of a neutral connection
merely disables the transformer.

Don't have the SLIGHTEST idea why electronic appliances have isolation
trasnformers inside ?

** You are an ARROGANT MORON who ought to be taken out and shot through
the head before YOU kill someone.

Likely you will electrocute yourself real soon.

I sincerely hope you do not kill a co-worker.

........ Phil