A little stuck during repair.

Thread Starter

Richie121

Joined Jan 12, 2014
27
Hi folks.
I'm replacing a sticky relay on a power saver socket but the problem seems to be on the control side. It's 230v AC and the logic side uses a switching regulator IC to provide 5v DC for control; This in turn provides +5v to one side of the relay to energise, but what is confusing me a bit is that the -ve side of the relay is connected directly to the Neutral AC rail. The main problem I am getting is there is +5v DC to the relay when ON, but 2.5v AC to it when Off. This seems enough to hold the relay on providing a fault.
Smoothing caps have now been replaced as originals had vented, but I feel the fault lies elsewhere.
Rectifier diodes seem ok when checked off board at low volts, but is it possible they fail at high potential allowing leakage?

Many thanks, and an excellent forum btw.
 

alfacliff

Joined Dec 13, 2013
2,458
A dc relay should not hold in with ac. if you give the module a bit of a "hit in the head", does it release? a little knock, not enough to bust anything. an old trick to get relays to release better is to put a little thin stickey tape on the end of the relay core so the armature dosnt stick to it from residual magnetism.
 

Thread Starter

Richie121

Joined Jan 12, 2014
27
It's possible it may be half rectified AC, so a rippling DC. I only used a DVM on AC setting. Either way I don't like it being there. Yes just handling it was enough to free it. Relay is a sealed board mount one so not something I can play with.
Good ideas though.

I've replaced the original 5v 15A relay (rated for 250v) with a 5v 30A one (same 250v), thinking bigger, heavier, stronger spring but had the same problem with it holding after 1 day use, so still think it is the control. For a general overview, these are the instructions - http://www.savasocket.com/support/user_manuals/SavasocketUserManualInstructions.pdf

It's out of waranty - so fair game.

Maybe some method to supress welding of the contacts?
 

MaxHeadRoom

Joined Jul 18, 2013
28,576
The fact the relay is connected to the neutral side is that evidently they have referenced the DC side to the AC side, (no isolation?).
If in fact there is a ripple on the relay when it is in the 'off' posn. it will be enough to retain it after switch off.
A DC relay or solenoid etc, has hysteresis, it takes much less energy to retain it than to pull it in.
A 'reverse engineering' of the circuit may disclose the reason?
Max.
 

ian field

Joined Oct 27, 2012
6,536
A dc relay should not hold in with ac. if you give the module a bit of a "hit in the head", does it release? a little knock, not enough to bust anything. an old trick to get relays to release better is to put a little thin stickey tape on the end of the relay core so the armature dosnt stick to it from residual magnetism.
A DC relay fed with AC should get hot enough to let the magic smoke out.

An AC relay has a laminated core just like a power transformer because a solid core would be the same as a shorted turn to the alternating magnetic field.
 

alfacliff

Joined Dec 13, 2013
2,458
and ac relays have a shorted turn on the core just under the armature to prevent them from humming. the shorted turn holds the magnetic field while the ac passes through zero.
 

MaxHeadRoom

Joined Jul 18, 2013
28,576
A DC relay fed with AC should get hot enough to let the magic smoke out.
Usually the reverse, the DC relay has much higher DC resistance than the AC type so not much happens.
With the reverse, a DC powered AC relay, then it usually gets hot or burns out, the AC relies on inductive reactance to limit current.

and ac relays have a shorted turn on the core just under the armature to prevent them from humming. the shorted turn holds the magnetic field while the ac passes through zero.
This is a shading ring, with DC the copper turn is not seen.
Max.
 

alfacliff

Joined Dec 13, 2013
2,458
old time pinballs (the relay type) used copper rings on dc relays for time delays, more rings, longer delay. the old relay systems used had a lot of delay, and needed a method of delaying a relay from dropping out for however long it needed.
 

MaxHeadRoom

Joined Jul 18, 2013
28,576
That would be on drop out only, up to a point any delay would cease, there is only so much energy stored to be released at switch off, a reverse diode has the similar effect on a DC relay.
I suspect the OP is seeing a DC ripple and it is indicating on his sensitive meter on the AC scale.
Max.
 

Thread Starter

Richie121

Joined Jan 12, 2014
27
Interesting info guys thank you.

After a bit of trace following, I see the generator of the 5v is a LNK302GN 7 pin IC - http://www.farnell.com/datasheets/10044.pdf and the circuit is straight off the front page of the datasheet. This clearly shows it using the neutral line in as the base reference for the output with no isolation.

The relay is marked as 5v, but as far as I can see the switcher outputs 12v as standard?
The relay has it's protection diode across it, but apart from that is connected to the +ve output of the LNK302GN after the inductor, and then uses a BC847 to connect it to the negative (neutral) rail.
Measuring voltage from base to neutal it is getting 0v when off, and 0.83v when turned on.

Do you think the transistor is leaky or the traces / inductors are picking up stray AC due to proximity to it?

I've now replaced all three electrolytics with new, and now I have a larger error voltage of 3v to the relay when it's supposed to be off.

Next step is replacing the transistor anyway as I'll destroy it taking it off to check I think. (SOT23)
 
Last edited:

radiohead

Joined May 28, 2009
514
Relay contacts "stick" because of bounce. When the relay is activated, the armature contact will bounce against the common contact before coming to rest. If there's enough current through the contacts, the arc between the contacts will heat up the contacts and will weld themselves together. Replace the relay. The contacts are done. It will always have the propensity to stick now.
 

Thread Starter

Richie121

Joined Jan 12, 2014
27
I have replaced the relay, but the fault seems to be the supply. 5v When on, but 3v when supposed to be off.
Waiting on my transistors to try one.
 
Top