will this cause a voltage spike?

[Eric R Snow]

I need to interface a relay to an old CNC machine. The machine sends a
24 volt pulse as a signal to operate turrets, indexers, etc. I need to
keep a relay on for about 1 second after being actuated by the pulse.
To do this I connected a capacitor across the relay coil. When the
relay recieves the pulse it now stays on for the time I need. But I
have only bench tested the setup. Is there any danger of sending a
voltage spike back to the machine with the above setup?
Thanks,
Eric

 

[default]

I need to interface a relay to an old CNC machine. The machine sends a
24 volt pulse as a signal to operate turrets, indexers, etc. I need to
keep a relay on for about 1 second after being actuated by the pulse.
To do this I connected a capacitor across the relay coil. When the
relay recieves the pulse it now stays on for the time I need. But I
have only bench tested the setup. Is there any danger of sending a
voltage spike back to the machine with the above setup?
Thanks,
Eric

You "bench tested it?" the spike aside for a moment, how does one
bench test it?

A "pulse" may be any pulse width - needs to stay on long enough to
charge the cap to hold the armature down when the pulse is gone - did
you consider that? One assumes this is a DC pulse - since it wouldn't
work with AC . . .

The output impedance of the pulse generator may be low or high - if
high you have a time constant between the impedance and capacitor that
may limit charging.

Two of several failure modes is that a spike goes back to the pulse
generator and eats it - unlikely because the cap would absorb the
energy - there is no spike to speak of, the cap just peters out, it
doesn't switch off abruptly. The cap would appear to be a short
circuit to the pulse when initially charging and that may eat the
pulse generator.

So, without knowing more about the device generating the pulse there's
no good way to tell.

If you are controlling an AC device with the relay, you may want to
consider a solid state relay. Many take an input of 3-32 volts for
the coil at 10 milliamps and will switch the output down to 24 VAC or
up to 240 VAC.

The 10 milliamp spec means it wouldn't load the generator as much as
any coil you're likely to find, 3 volt spec would take a smaller cap
to stay on, and wouldn't generate any reverse EMF when the circuit
opens.

One good idea might be to use a solid state relay.

The one second will be highly variable with this jury rig, one assumes
that would be tolerable?

 

[Eric R Snow]

You "bench tested it?" the spike aside for a moment, how does one
bench test it?

A "pulse" may be any pulse width - needs to stay on long enough to
charge the cap to hold the armature down when the pulse is gone - did
you consider that? One assumes this is a DC pulse - since it wouldn't
work with AC . . .

The output impedance of the pulse generator may be low or high - if
high you have a time constant between the impedance and capacitor that
may limit charging.

Two of several failure modes is that a spike goes back to the pulse
generator and eats it - unlikely because the cap would absorb the
energy - there is no spike to speak of, the cap just peters out, it
doesn't switch off abruptly. The cap would appear to be a short
circuit to the pulse when initially charging and that may eat the
pulse generator.

So, without knowing more about the device generating the pulse there's
no good way to tell.

If you are controlling an AC device with the relay, you may want to
consider a solid state relay. Many take an input of 3-32 volts for
the coil at 10 milliamps and will switch the output down to 24 VAC or
up to 240 VAC.

The 10 milliamp spec means it wouldn't load the generator as much as
any coil you're likely to find, 3 volt spec would take a smaller cap
to stay on, and wouldn't generate any reverse EMF when the circuit
opens.

One good idea might be to use a solid state relay.

The one second will be highly variable with this jury rig, one assumes
that would be tolerable?

The bench testing was done simply by waving a wire past a contact such
that it made contact very briefly. The pulse from the machine is 24VDC
and 100ms long. The pulse comes from a relay in the machine control
that connects to the 24VDC rail. The relay I need to switch on could
control an AC load because it's up to me whether to use a DC or AC
solenoid. The on time of the relay can be longer than 1 second but not
shorter. This is because the solenoid that the relay controls actuates
an air valve that then operates a clamping device. This device needs
just a little less than 1 second to open completely. The power for the
solenoid is separate from the CNC control. I like your idea of using a
solid state relay. Since the pulse from the machine is pretty constant
doesn't that mean that the solid state relay on time will be as
constant?
Thanks,
Eric

 

[Jamie]

I need to interface a relay to an old CNC machine. The machine sends a
24 volt pulse as a signal to operate turrets, indexers, etc. I need to
keep a relay on for about 1 second after being actuated by the pulse.
To do this I connected a capacitor across the relay coil. When the
relay recieves the pulse it now stays on for the time I need. But I
have only bench tested the setup. Is there any danger of sending a
voltage spike back to the machine with the above setup?
Thanks,
Eric

You may want to put in a flyback diode (free wheeling voltage some call
it) to absorb the high voltage from the coil in case something happens
to the capacitor.
Also, you may want to consider using a One-Shot timer so that a small
load is introduced from the pulse of the CNC machine. The method you're
using now maybe taxing the output circuit of the CNC machine. I can only
assume that you're using a rather large value to be able to hold the
relay? that is, a relay suited for the load you're applying.

 

[default]

The bench testing was done simply by waving a wire past a contact such
that it made contact very briefly. The pulse from the machine is 24VDC
and 100ms long. The pulse comes from a relay in the machine control
that connects to the 24VDC rail. The relay I need to switch on could
control an AC load because it's up to me whether to use a DC or AC
solenoid. The on time of the relay can be longer than 1 second but not
shorter. This is because the solenoid that the relay controls actuates
an air valve that then operates a clamping device. This device needs
just a little less than 1 second to open completely. The power for the
solenoid is separate from the CNC control. I like your idea of using a
solid state relay. Since the pulse from the machine is pretty constant
doesn't that mean that the solid state relay on time will be as
constant?
Thanks,
Eric

Something else that occurred to me - putting a cap across an ordinary
coil type relay will have another unintended consequence - the relay
will drop out much more slowly - the contacts will come apart with
glacial speed, that in turn, is likely to promote arcing and shorten
the life of the contacts.

SSR wouldn't have that problem.

They do make SSR's in a variety of flavors these days normally closed
and AC/DC types etc.. But the simplest, least costly, is normally
open to switch AC with a DC control of 3-32

If the pulse is generated by a relay in the machine, there's an
excellent chance it will work the way you want without problems.

 

[Jamie]

Something else that occurred to me - putting a cap across an ordinary
coil type relay will have another unintended consequence - the relay
will drop out much more slowly - the contacts will come apart with
glacial speed, that in turn, is likely to promote arcing and shorten
the life of the contacts.

SSR wouldn't have that problem.

They do make SSR's in a variety of flavors these days normally closed
and AC/DC types etc.. But the simplest, least costly, is normally
open to switch AC with a DC control of 3-32

If the pulse is generated by a relay in the machine, there's an
excellent chance it will work the way you want without problems.

Actually, the way Relays work by nature (most of them) produce their
own Hysteresis. I don't think you'll need to work about that how ever,
putting a snubber on the contacts for the inductive load that is
connected to it will help save the life of the contacts.

For example, a 12 volt relay will pull in at around 9 volts on the
average and will hold down to lets say 6 volts. at the point of where
it releases from the coil head will start the rapid decay of attraction
as the contact level arm then starts to move aways, which also, weakens
the attraction along with the coil voltage dropping.
Since you have DC with a CAP going to the coil, there will be no
vibration effects.
Now, if you were applying this theory using AC, this maybe a
different story.

 

[Eric R Snow]

Actually, the way Relays work by nature (most of them) produce their
own Hysteresis. I don't think you'll need to work about that how ever,
putting a snubber on the contacts for the inductive load that is
connected to it will help save the life of the contacts.

<SNIP>
Greetings Jamie,
I have decided to go with the SSR. It seems like it's the best way if
I can still use a capacitor to keep the relay on for the required 1
second. However, I have another application where a relay is
controlling a contactor that controls an AC motor. The relay does
suffer from burned contacts and sometimes sticks. So I have replaced
the relay a couple times. The hot tub place told me the crummy design
is why the company that made the controller went out of business. So,
to help prevent the contacts burning how do I make a snubber?
Thanks,
Eric

 

[default]

<SNIP>
Greetings Jamie,
I have decided to go with the SSR. It seems like it's the best way if
I can still use a capacitor to keep the relay on for the required 1
second. However, I have another application where a relay is
controlling a contactor that controls an AC motor. The relay does
suffer from burned contacts and sometimes sticks. So I have replaced
the relay a couple times. The hot tub place told me the crummy design
is why the company that made the controller went out of business. So,
to help prevent the contacts burning how do I make a snubber?
Thanks,
Eric

A snubber is a combination of resistor in series with a cap across the
contacts. They can be made or bought in that form, in a single
package.

I have no idea how to size the cap resistor combo . . . typical values
are 100 ohm resistor in series with a point one microfarad cap. The
cap has to be a able to withstand the AC voltage and should be a high
quality low inductance type. The resistor should be flame proof.

The switch on my computer speakers was causing my power supply to
reset due to a transient it created - a snubber fixed it.
(.1uf/100ohm)

Jameco catalog shows six ready made snubber networks as well as some
individual caps suited to transient suppression applications. One
snubber is a one microfarad in series with a 100 ohm 1/2 watt
resistor, the others are all point one microfarad in series with 22,
47, 100, 150 ohms

The cap should be polyester or polypropylene dielectric rated for the
AC voltage you will be using or 2-3 times higher if rated for DC.

 

[Eric R Snow]

A snubber is a combination of resistor in series with a cap across the
contacts. They can be made or bought in that form, in a single
package.

I have no idea how to size the cap resistor combo . . . typical values
are 100 ohm resistor in series with a point one microfarad cap. The
cap has to be a able to withstand the AC voltage and should be a high
quality low inductance type. The resistor should be flame proof.

The switch on my computer speakers was causing my power supply to
reset due to a transient it created - a snubber fixed it.
(.1uf/100ohm)

Jameco catalog shows six ready made snubber networks as well as some
individual caps suited to transient suppression applications. One
snubber is a one microfarad in series with a 100 ohm 1/2 watt
resistor, the others are all point one microfarad in series with 22,
47, 100, 150 ohms

The cap should be polyester or polypropylene dielectric rated for the
AC voltage you will be using or 2-3 times higher if rated for DC.

Thank You Default. I buy from Jameco from time to time so I'll check
out their stuff.
Eric

 

[ehsjr]

<snippety>

You would do better to use an electronic time-delay relay set for a
one-second delay-on-release. The problem with a cap across the coil is
that you introduce some delay on activation as well.

How much delay on activation do you figure will be introduced?

Ed

 

[neon]

you charge the reay coil to 24 volts the cap across is 24 volts when you release the coil will spike bu in the recers direction a diode is more like what you want to clamp the spike to zero the ca may actualy make things worst. for proection fom spikes at the generator source try a snubber.