Snubber network question

Hello all,


3 weeks ago, I had the thermostate of my hot water boiler replaced because
it generated sparks right before switching on the circulating pump.
(sparks to be heard via the radio as crackles).

After 1 week of good working, the new thermostate now also sparks.

Could this be solved by placing a snubber network (R en C in series) over
the switching contact of the thermostate ? And if so, how should I
dimension this RC network, knowing that the mains voltage here in Belgium
is 220V/50Hz. Personaly I was thinking of a 1k resistor in series with a
capacitor so that the RC value falls in the range of 10 kHz, i.e. a 1 k
resistor in series with a 15 nF cap.

But is this effective ? And what should the voltage of the cap be ? I
would guess, since the pump motor acts as a coil, that the voltage can
rise up to as much as then times the line voltage, so about 2000V would be
in order. Is this correct ? Should the resistor have specific features ?

Come to think of it, now that I do the math for the resistor, maybe it
should be 100k, question of limiting the current.

Anyhow, I'm a little (or a lot) lost here, and I'd be grateful for any
suggestions.

TIA

Danny

 

[CBarn24050]

100nF + 100R

 

[CFoley1064]

Subject: Snubber network question
From: daniel.hendrickx@alcatel.be
Date: 3/1/2004 5:52 AM Central Standard Time
Message-id: <1078141975.648245@slbhw0

Hello all,


3 weeks ago, I had the thermostate of my hot water boiler replaced because
it generated sparks right before switching on the circulating pump.
(sparks to be heard via the radio as crackles).

After 1 week of good working, the new thermostate now also sparks.

Could this be solved by placing a snubber network (R en C in series) over
the switching contact of the thermostate ? And if so, how should I
dimension this RC network, knowing that the mains voltage here in Belgium
is 220V/50Hz. Personaly I was thinking of a 1k resistor in series with a
capacitor so that the RC value falls in the range of 10 kHz, i.e. a 1 k
resistor in series with a 15 nF cap.

But is this effective ? And what should the voltage of the cap be ? I
would guess, since the pump motor acts as a coil, that the voltage can
rise up to as much as then times the line voltage, so about 2000V would be
in order. Is this correct ? Should the resistor have specific features ?

Come to think of it, now that I do the math for the resistor, maybe it
should be 100k, question of limiting the current.

Anyhow, I'm a little (or a lot) lost here, and I'd be grateful for any
suggestions.

TIA

Danny

Good morning, Danny. Looking at your post, I'd first like to ask a couple of
questions. When the first thermostat switch was replaced, was radio
interference the only problem, or was there something else? Water heaters
usually don't come with any guarantees that they're not going to generate RFI
when they turn on. I'm assuming the first switch failed, and that the RFI is a
symptom you picked up investigating along the way. I'd also like to know if
possibly the thermostat is driving the motor directly, or it's driving a relay
that's driving the motor. Also, if the RFI is occurring on turn-on or turn-off
of the motor.

One thing you might want to look at is a partial short in the relay coil or
motor coil, and the possibility of a partial coil short to ground. That can
jack the current way up, and cause contact arcing where you normally wouldn't
have any. Check the wiring, and ohm things out. It might help to find out the
typical resistance values of any coils or motors in the circuit, and measure
them to check. Measure to ground too. The standard principle is that, if it's
designed to work well and something doesn't work right, you should look for
something that's broken first. You might have to temporarily disconnect things
to make measurements.

Many control circuits will use a transformer to generate a lower control
voltage which is switched by the sensor. This is especially common in the EU,
because of different wiring standards. That is used to drive a relay coil,
which drives the line voltage motor. Before you make any mods, you need to
know the electrical wiring, and exactly what's being switched where. You
mentioned that the problem occurs on turn-on rather than turn-off. That might
mean something's going on you're not aware of. Get the wiring diagram, look at
it.

When a contact opens across an inductive load, the voltage across the inductive
load being turned off will rise (Lenz' Law). Since the mechanical contacts are
moving away from each other slowly, a spark will be initiated across the
contacts, and will continue until the contacts are far enough away to
extinguish the arc. Without a snubber, the inductive kick of the load is being
absorbed by the energy dissipated by the spark. Not a good way to do things,
because it chars and pits switch contacts (and generates RFI). At worst, the
arc can eventually spot weld the contacts together on an intermittent make of
the contacts. A snubber will absorb some of the inductive kick power when the
switch opens, keeping the voltage across the contacts low enough for long
enough that a spark doesn't form across the contacts while they're separating.
In other words, the rise in load voltage remains lower than the arc-over
potential. The rest of the power of the inductive kick will eventually be
absorbed by the coil itself.

Now you can look at specifying a snubber. A snubber is an R in series with a
C, and is placed across the load (or across the switch contacts, if it's not
possible to put it across the load). The snubber will conduct normally when
the load is on (or off, if you've put it across the contacts). The current is
limited by the cap rather than the resistor (which is good, because it means
you can use realistically sized resistors).

Find the rated switching current of the switch. Choose a resistor which would
provide the rated current across the switch if placed in series with the line
voltage. For a 24VAC control voltage and a 1A switch, that would be 27 ohms.
For 240VAC at 1A, that would be 270 ohms. A 1 watt minimum resistor should be
used here for 240VAC, 2 watt for 24VAC. A 3 watt wirewound is always OK --
just make sure you don't use carbon film, they are unable to handle the power
of current spikes. Now look at the cap. You need to get a self-healing AC cap
rated for more than the AC control voltage. For a 24VAC circuit, you would
like a cap rated for at least 50VAC; for 240VAC you should choose 350VAC or
400VAC rated. To calculate the capacitance, you need to characterize the
opening and closing time of the switch, with distance over time, as well as the
inductance and resistance of the motor. Too much here, especially for a
hobbyist or do-it-yourselfer. It might be easier to start with an 0.1 uF cap
for 240VAC, or 0.47 uF if you've got 24VAC, and just try it. If you're still
getting sparking across the contacts, bump up the capacitance just until there
isn't a visible spark across the contact. Assuming you've got a low voltage
control transformer switching a relay, your snubber circuit might look
something like this (view in fixed font or use M$ Notepad):

Thermostat Control Circuit

Thermostat
SW1
_/
-. ,--------o/ o-----o---.
220VAC )|(24VAC or | |
)|(12VAC .-. |
-' '----. | | |
| R | | C|
| '-' C|RY1
| | C|
| | |
| --- |
| C --- |
| | |
'-------------o---'

Summary -- play safe with line voltage, look for something that's broken before
you make mods, select the resistor based on switch rating, choose cap based on
ability to quench the spark and don't pick a bigger cap than you need.

Good luck
Chris

 

[Jamie]

get into the thermostat and put a ceramic disc cap across the contact
points.
a .01 uf should be ok but a .1 uf would be better. (one that can
handle at least 400 Volts.




CFoley1064 wrote:

Subject: Snubber network question
From: daniel.hendrickx@alcatel.be
Date: 3/1/2004 5:52 AM Central Standard Time
Message-id: <1078141975.648245@slbhw0

Hello all,


3 weeks ago, I had the thermostate of my hot water boiler replaced because
it generated sparks right before switching on the circulating pump.
(sparks to be heard via the radio as crackles).

After 1 week of good working, the new thermostate now also sparks.

 

Hi Chris,

thanks for your response.

As for your questions, the thermostate switch switches the mains directly,
and as for the sparks, they are not generated in the boiler itself as I
hear the switch sparking. I don't know if I can find the specs of the
motor coil. But the system works fine up to now. I'm only afraid for the
pits on the contacts (which you also mention) which may weld them together
or destroy them entirely.

Your remark about placing the snubber over the load or the switch has me
wondering: is it really possible to place the snubber over the load ? I
thought a snubber was something typical to be placed over switches (triacs
etc) ?

Anyway thanks for your reply already.



Danny

 

[Jamie]

i am sorry, that was intended for the originator.


Jamie wrote:

get into the thermostat and put a ceramic disc cap across the contact
points.
a .01 uf should be ok but a .1 uf would be better. (one that can handle
at least 400 Volts.




CFoley1064 wrote:

Subject: Snubber network question
From: daniel.hendrickx@alcatel.be
Date: 3/1/2004 5:52 AM Central Standard Time
Message-id: <1078141975.648245@slbhw0

Hello all,


3 weeks ago, I had the thermostate of my hot water boiler replaced
because it generated sparks right before switching on the circulating
pump. (sparks to be heard via the radio as crackles).

After 1 week of good working, the new thermostate now also sparks.

 

[CFoley1064]

Subject: Re: Snubber network question
From: daniel.hendrickx@alcatel.be
Date: 3/2/2004 2:41 AM Central Standard Time
Message-id: <1078216896.95827@slbhw0

Hi Chris,

thanks for your response.

As for your questions, the thermostate switch switches the mains directly,
and as for the sparks, they are not generated in the boiler itself as I
hear the switch sparking. I don't know if I can find the specs of the
motor coil. But the system works fine up to now. I'm only afraid for the
pits on the contacts (which you also mention) which may weld them together
or destroy them entirely.

Your remark about placing the snubber over the load or the switch has me
wondering: is it really possible to place the snubber over the load ? I
thought a snubber was something typical to be placed over switches (triacs
etc) ?

Anyway thanks for your reply already.



Danny

Snubbers can be placed either across the load or across the switching
mechanism. For solid state switching, it usually makes more sense to put the
R-C snubber across the load, because that's where the circuit board is. The
snubbing action of the R-C is essentially the same. For a snubber across the
load, it's conducting as an RC element when the power to the load is on. For
an RC snubber across the switch contact, current will flow through the snubber
when the switch is off.

If you're making a mod to an existing piece of equipment, you probably want the
snubber to be across the load, to avoid epithets, swearing and general bad
karma from the poor guy who's repairing the circuit next time. If you have
voltage (even through an RC snubber) across the load when the switch is off,
and he doesn't know about it, he may have a bit of a surprise when he touches a
wire that's supposed to be de-energized. Of course, the source will be
current-limited, and it probably won't be fatal, but it probably won't be the
highlight of his day.

Remeber that, if you make unauthorized mods to a circuit, mark it off. I've
always believed there's a rather toasty sub-ring in a very warm place (probably
fifth or sixth circle) reserved for the afterlife of those who make mods to
circuits or code without leaving very plain notice for unsuspecting future
tinkerers. Probably they are forced to repair each other's work over and over
for eternity, with each rat's nest and indigestible block of spaghetti code
getting worse and worse and worse.... ;-)

Good luck
Chris