Solid State Relays and EMI

I'm using some solid state relays, and having some problems with (I
believe) EMI.

http://www.cel.com/pdf/datasheets/ps7141a.pdf

I'm using this part in the B and C configurations.

These items are located near some RF sources operating at about
148MHz. During transmit (.5W power), I'm getting some false "on"
states. I'm pretty sure this isn't due to the equipment on the load
side, as removal of the SSR's eliminates the effect.

1. Anyone know if these type of SSRs are subject to EMI that can
cause false on-states?
2. Would induced AC voltages (from the RF source) in the load lines
(around 148MHZ) decrease the R-off resistance?
3. If the problem is in the SSR's, anyone have any suggestions
regarding how to lessen/prevent this? (EMI shielding, other PN's that
are immune to this, inductors/caps that could be placed in the load
lines, etc).

Thanks for your time!
Dave Harper

 

Be advised that relays made with copper (a *solid*), iron (a *solid*),
sometimes plastic (a *solid*) and sometimes glass (a *solid*) ARE SOLID
STATE (certainly *not* gas state or liquid state)...
Also there are a number of variations that handle RF very well with
low VSWR, and some have response times under a millisecond.

 

"David Harper" ...

What type _exactly_? URL?

Not if it's a mains load switching type with a triac or thyristor output,
they have no "R-off", they are either ON or OFF.

SSR's with an optocoupler inside are easily triggered by RF voltages
occurring between in-and output. You might try adding a ceramic Y-rated
capacitor (470pF ... 1nF should do), between the ground on the input
(optocoupler) side and the supply pin on the output (contact) side, as short
a connection as possible. Also one over the output pins.
**** Warning **** this may solve the problem but will add leakage current
between controller and mains. Check you stay on the safe side!

Another test could be just the 1nF Y-rated cap over the output (less likely
to help).This does not introduce leakage to the controller ground, just some
extra OFF-state current in the load.

Regards,
Arie de Muynck

 

Well, obviously. Solid state generally means made of semi-conducting
material and containing no moving parts. Are you thinking of magnetic
(coil) relays? The ones I'm using involve a LED and a MOSFET, not a
coil.

http://www.cel.com/pdf/datasheets/ps7141a.pdf

These have an average response time of about .35ms. Also, I'm not
running an RF signal through them. They are being affected by a
nearby RF source.

Thanks,
Dave

 

From www.m-w.com:

Main Entry: solid-state
Function: adjective
1 : relating to the properties, structure, or reactivity of solid
material; especially : relating to the arrangement or behavior of ions,
molecules, nucleons, electrons, and holes in the crystals of a substance
(as a semiconductor) or to the effect of crystal imperfections on the
properties of a solid substance <solid-state physics>
2 a : utilizing the electric, magnetic, or optical properties of solid
materials <solid-state circuitry> b : using semiconductor devices rather
than electron tubes <a solid-state stereo system>

So no, the molecular make-up of a device does not necessarily categorize
it as solid-state. It's solid state if every atom in the device is
relatively stationary, meaning no moving parts (e.g semiconductors).

 

I *do* know the difference, and that the terminology "solid state
relay" is used for relays using some kind of semiconductor technology -
starting from incandescent lamp used to illuminate a CdSe or similar
cell.
With sufficent RF field, all "solid state relays" will be affected -
ie: act as if some non-zero input had been applied.
Sometimes, the RF field can even make the "solid state relay" act as
if it were linear and/or erratic.
In the case of LED/FET combo, the FET is the first to be affected.
Proper shielding and RF bypassing will help reduce the effect.
If the RF field is very strong, then the FET and perhaps the LED would
be destroyed.
Depending on the severity of the problem, one could go back to the
older "solid state relay" technology and use the lamp/CdSe scheme OR use
"iron state" relays (to coin a term), which can be rather sensitive.

 

Let us see - - the definition you quoted maid no mention, even
indirectly concerning *moving* parts.
And the last time i looked, each (individual) piece of copper , iron,
plastic, glass, whatever does not move...
And the combination of the copper and iron uses the magnetic
properties of (solid) iron; which fits the quote exactly...
BTW, iron has a rather definitive crystalline structure, and since
annealed OHFC is not used for the coil, then the copper also has a
crystalline structure.

Mind you, i am not arguing.
Just point out a few iregularities...

 

True, although it does mention semiconductors, and I can count the
number of semiconductors composed of moving parts on zero hands.

When used in solid-state electronics, correct.

....to move mechanical components in an electro-mechanical device.

Could have fooled me

Well sure, if you want to get technical, vacuum tubes and semiconductors
are both "solid" objects in the sense that they're not made out of gas,
water, or plasma. But if you're going to nitpick subtle nuances of the
English language there are far easier targets (why do we drive on a
parkway and park on a driveway?).

I think the confusion concerns the use of "solid", which is used in the
sense of the device's physical configuration, not so much its molecular
state. We could use the term "finite-state", but that's generally used
to describe electric circuits.

I'm just pointing out convention. In all fairness, the term is still
jargon, but referring to any electro-mechanical device as "solid-state"
will confuse most people.

 

With the turn-on time being ten times the turn-off time, I don't
expect that false turn-on will be a result of LED emitter modulation
at the interfering frequency.

Fet structures usually have a hefty input capacitance; the coupling
capacitance to the gate is piddling by comparison.

Biggest coupling capacitance that could produce false gate turn-on is
from the two drains. Is there a fairly large amplitude of RF voltage
present there? These thinga are really only isolators at low frequency
or DC, unless a voltage close to at least 1/10 their rating exists
when 'off'.

RL

 

I read in sci.electronics.design that Chris S. <>

Not any more. Consider the TI micro-mirror-matrix chips, for example,
and solid-state microphones are appearing on the market, too.

 

Solid-state accelerometers with on-chip bending arms have been
available for about a decade- and have sold in large quantity.

Best regards,
Spehro Pefhany

 

I wouldn't necessarily classify MEMS devices as solid-state. While
certain types make use of "bending" to move small actuators, they still
rely on this movement to perform their function. The characteristic I
typically associate with solid-state is a lack of dependence on physical
motion to function.

 

IMHO, the key characteristic is that the function depends on
semiconductors. The branch of science involved is solid-state physics.

Best regards,
Spehro Pefhany

 

That's more of a technicality. Those microphones use piezo-electric
effects, which is a function of [micro]strain.

Yes, technically the elements are moving a few microns (or less).
There's several "solid state" temperature sensors that experience far
more "movement" from thermal expansion than those microphones do.

They are not moving fractions of an inch (or more), like they are in
electro-mechanical devices.

Dave

 

I think you're missing the original point. In the field of
computer/electrical engineers, "solid state" refers to a branch of
electronics based primarily on semi-conducting (ex. silicon)
materials.

The whole "movement" part is really a minor footnote.

Your original statement about them being "solid" and not "gas" or
"liquids" was obvious, and came from left field. I don't think anyone
reading this thread would go into Radio Shack and ask for "liquid
relays".

Dave

 

Burridge might ;-)

...Jim Thompson

 

OK, but I've seen SSRs used at broadcast transmitter sites running
many kilowatts of RF with no problems. SSRs are like Rank telecines.
When you see it 'misbehave' don't assume it isn't being 'told' to do
it.
GG

 

The ones I'm using involve a LED and a MOSFET, not a

Thanks for answering my question - although on another part of the thread...

You may not know it, but the light from the LED does not directly turn-on
the MOSFET. It shines on a string of small photovoltaic cells that charge
the gates. When the light is off, the photocell leakage current (sometimes
with an extra resistor across them) discharges the gate again.

Any idea what happens if an RF voltage is capacitively coupled from LED pins
to the _string_ of photodiodes? It's called rectification. The resulting
voltage turns on the gate.

Did you alrady do the test with the 1nF between driver size and output side?

Regards,
Arie de Muynck

 

1. It's very likely. In my classes, I've been taught that EMI can bugger
up the proper working of just about any circuit, so it should always be
taken into consideration.
2. Er, I think it's a simple matter of induced current in the line
giving false hi signals.
3. First thing I'd do is make absolutely sure your relay coil is
grounded rather than floating when in the "off" state. Being assured of
that, if the problem persists, you might want to try a simple RC
band-elimination filter. If you do not have it handy, I can give you the
formula to determine the R and C values.

I apologize in advance if your electronics knowledge is more advanced
than mine and I've just wasted your time.

- NR

 

I was having a little fun...