Siemens Oven Control Board Repair - Sourcing the Relays - Planned Obsolescense

[TrickyNekro]

Hi!

The question is not how to repair it, that I know already. The problem is to source the relays.
This oven has the bad habbit of doing an Oopsie every 3.5 - 4 years, just when it was just out of warranty (big surprise no wonder).

This is the second time the relays go bad, first time we changed the whole board as the failure was much more catastrophic and
someone might have also played with the RCD pass-through (I know unacceptable but it failed so horribly that it might as well be terrible design).

Well, this time I was around and I am a huge fan of the RCD, we are.... close friends... no pun intended.

But I can not have the board failing every 3 - 4 years of something that costs 1000+ €. So I am looking for replays that have a bit more oompf but are in the same form factor.
Offending Relays:
http://www.hongfa.com/pro/pdf/HF3FD_en.pdf
http://www.hongfa.com/pro/pdf/HF152F_en.pdf

These are miniature 9VDC (9V yeah they are being a$$holes also with the parts sourcing) 1 Form A (NO) SPST relays.
So far for the smaller ones I have found this that might fit the bill:
https://gr.mouser.com/datasheet/2/307/en-g5le-1131193.pdf

I am not just trying to repair the board though, the shot relay is only one. I am rebuilding it with stronger parts.
I dremeled the offending relay and it was pretty charcoal-ed. Again unacceptable for something in that price range.
But consumer products... what do you expect...

I would be happy if you got ideas of models and suppliers. Truth be told I would like to stick to Mouser or Digikey as suppliers, but right now I need a place to cook my meals so, it´s indifferent.

[fzabkar]

Have you considered redesigning the switching using optically coupled Triacs?

Just one example incorporating zero crossing detection …

https://www.vishay.com/docs/83627/il410.pdf

[TrickyNekro]

Yeah, I though of many things, some also like this, but nah, that would over complicate things.

I can do it, and most probably it will work 10 times better than the original design, but still... You got to think about the redesign, space constrains cause these Triacs
would certainly require thermal design etc etc etc. Hell I could do it, with back to back IGBTs or MOSFETs if I really wanted... but... Do I?
Also power supply limitations, possible mitigation of power supply glitches etc etc etc... so no... I do not want to do it...

A replacement board from Siemens costs around 70 - 80E... I just don´t want to give them the money out of spite.
If necessity becomes more than spite, then ok. But I can hold on to my spite for a little longer.

I actually have found a replacement part.
https://gr.mouser.com/ProductDetail/Song-Chuan/207-1AH-S-9VDC?qs=sGAEpiMZZMtSzCF3XBhmW0jFivbULW3rWSOxOuDOl88%3d

Direct replacement for both styles of relays, will do a great job! But it is non-stocked. So the problem is... How do I source it? I have contacted Mouser,
I will be also contacting the manufacturer, but biggest bet is somewhere from the debts of China. That´s what I wanted to avoid though...


Cheers, Lefteris

[TrickyNekro]

I made some photos just to drive my point home.

These are the relays which are all on the same connector. These are probably the relays controlling the heating elements.
From left to right.
The two smaller relays have significant carbon build after 4 years of use. The third small one is the one that failed (.....duh). The forth one is a bigger relay with much less carbon build up!

The one with the most kilometers left based on carbon build up, is of course the big relay. And since both relay are present on the same board and their price and coil resistance is not much different,
I wonder! Is that planned obsolescence by Siemens? It´s another thing having the oven fail every 6 - 7 years and another thing having it fail at 4+ years. Judge for yourselves.

[floobydust]

It's not the relay. Any relay there will fail the same way.
The heating elements have inductance which make large arcs across the contacts when opening, which destroys the contact metals.

The fix is to add a RC snubber network or MOV across the contacts. I add 10-47nF X-class capacitors  33-47R 1W resistors depending on the inductance. I do this to many appliances to stop them from eating relays. It's too bad appliance manufacturers could care less and save pennies leaving this out - planned wearing out so you buy a new appliance instead of replace control board $$$.

One safety problem caused by this is added leakage current through the heating elements.
If you can remove (unplug) the heating elements for say cleaning, with a wet rag, there is a shock hazard with larger capacitors or MOV's even though the heat is turned off.

[TrickyNekro]

Well then... why use a MOV and not use a TVS? Bidirectional that is.
Well yeah, I have not thought of it like this, The current is large enough for inductance to start playing a role, that´s correct.

But I am clinging more to a TVS solution now that you mention it. It should be better also with aging compared to a MOV.



Edit: Now that you mention it... They do not have any kind of protection even for the induction motors. Granted it´s a much much smaller load, but hey a TVS costs almost nothing
for such small currents.

[floobydust]

It's hard to know how much inductance (energy storage) there is in the heating element.
Assuming an induction motor is 1/10 the current but 100X the inductance, same energy amount to deal with.

You can get HV TVS to 440V but they do not take much current. I never see them used with relay contacts.
Even with the oven off, mains transients will hit the TVS. If the TVS shorts, it will burn up in series with the heating element and make lots of smoke.
I think a RC snubber is the best solution.

[SilverSolder]

What about a solid state relay?  That should be an easy fix, since many of them can be controlled by the existing coil drive?

For example, could something like this be hacked in pretty easily?

https://www.ebay.com/itm/SSR-40DA-40A-Solid-State-Relay-DC-AC-3-32VDC-24-380VAC-for-Temp-Controller-Tool/162587436373

[fzabkar]

Beware …

https://www.instructables.com/id/The-inner-workings-of-Counterfeit-FOTEK-SSRs/

[floobydust]

What is your oven power? Example 2.8kW is 11.7A at 240VAC.
An SSR will dissipate about 1W per amp, of heat. So you would need a large heatsink to dissipate 12W of heat. This is the hassle with SSR's.

[SilverSolder]

Assuming a good quality SSR and an adequate heatsink/fan,  would that not be a more reliable solution in the long term?  -  or are SSRs as prone to fail eventually as a relay?

[Kleinstein]

In the likely relevant current range (like 10-15 A) a SSR will have a significant heat problem. The classical relay is at maybe 0.5 W - 2 W, while the SSR is at 10-15 W. So if space is restricted the SSR would need a fan than can fail.

The obvious solution is a relay with higher power rating mechanical relay. However this might need more space - though not as much as the SSR.

Normally heaters don't have that much inductance that they absolutely need a snubber or MOV.  One could still have the snubber in parallel to the load if leakage currents are a concern.

Another option would be to use the small relay to control an external larger one that than controls the heater.

[SilverSolder]

Another option would be to use the small relay to control an external larger one that than controls the heater.

That sounds like a very easy and good solution - perhaps there is space inside the oven to add a couple of extra relays.  Maybe even SSRs could be used there, if it has a metal area that could sink the heat!

[TrickyNekro]

Nah, I´m not doing an SSR replacement, it´s a build in oven, heat is a problem.

And yeah, TVS - MOV could lead to explosive situations, so that is ruled out :-/

Snubber sounds good, but I do not wish to have the always on load. Although, if I remember right, from last time the thing exploded, they use two relays for the load.
At least they managed to do this right. Not 100% sure though. I really have to try and find a wiring diagram.

The codes on the board do not help much.

And all this could simply be avoided with a proper ZCD, that most times comes free with the microcontroller... but who am I to know this 

Your ideas help a lot so far, keep it coming please :-)

[TrickyNekro]

Another option would be to use the small relay to control an external larger one that than controls the heater.

That sounds like a very easy and good solution - perhaps there is space inside the oven to add a couple of extra relays.  Maybe even SSDs could be used there, if it has a metal area that could sink the heat!

Yeah, thought of that, I just wanted to avoid it, just get some heavy duty "sugar cube" relays. The problem is that mostly Chinese manufacturer "attempt" such a thing and that alone should be telling.
Not to bad mouth these manufacturers or anything, but when you see big player staying much more conservative about their numbers, it makes you wonder.

Still, "nice" idea as much as it can be about driving a relay with a relay, the second relay still needs a power source, and now look at the hassle you created.
No need for that though, there are higher power relays that match voltage and coil resistance, just not the foot-print.

I can do mods, that´s no problem. But I do not want to do "big" mods if I don´t have to.

[SilverSolder]

Still, "nice" idea as much as it can be about driving a relay with a relay, the second relay still needs a power source

That could be mains voltage...  you can easily get relays with 240V coils...  and you could choose really big, quality relays at a good price since you no longer care about the exact shape...

Something along these lines, perhaps...
https://www.ebay.com/itm/JH1A-W-AC220V-Q-AROMAT-AR551898-Power-Relay-Coil-220V-30A/112911300519

[Yansi]

Look up OMRON G5LE.

[TrickyNekro]

Look up OMRON G5LE.

Look at my first post. 

I was just looking for something maybe a bit beefier...


That´s probably what I will end up with though...

[TrickyNekro]

Still, "nice" idea as much as it can be about driving a relay with a relay, the second relay still needs a power source

That could be mains voltage...  you can easily get relays with 240V coils...  and you could choose really big, quality relays at a good price since you no longer care about the exact shape...

Something along these lines, perhaps...
https://www.ebay.com/itm/JH1A-W-AC220V-Q-AROMAT-AR551898-Power-Relay-Coil-220V-30A/112911300519

I know there are 240V coil relays. As I said, I do want to avoid extensive modifications.

I will probably go with the G5LE relays. Problem is that the "E" high power variant only comes in 5V, 12V, 24V variants.
I could get away with the 12V variant probably. It´s still something that I am thinking about. Problem is the relative high temp environment.
Magnetics do not like high temp environments, and running a 12V relay at 9V already means a weaker field. It´s all debate-able I guess...

[TrickyNekro]

So I started the reverse engineering on a "lighter" level.

The power supply is 9.6V / 12.5W so ~1.3A. That´s unfortunately awfully specific. And Lo and behold! they do actually use a proper ZCD circuit and they have in the power supply!

I do not know if the existence of an ZCD circuit makes matters better or worse. Probably worse. I will be reverse engineering the ZCD circuit and posting it here for the details!

Because if they are using it only for the triac control and not the relays, then it is planned to produce damage on the relays on the long run. Else it´s bad implementation, could also be.

I´ll be investigating more, this seems interesting 

[floobydust]

For long-life mains switching, it is common to use both an SSR and a relay in parallel.
Furnaces and heaters have an endurance spec to meet, typically 50,000 or 100,000 cycles, in IEC 60730 and other safety standards. You can't get this many cycles with an ordinary relay, most are resistive-load rated to ~20,000-30,000 cycles.

Firmware turns on the TRIAC (zero-cross) and then the relay. Since the relay needs time around 20msec to pull in, it sees no load at first. The TRIAC does the work switching (on) the load current. The relay ends up pulling in some time after and takes over passing the load current. TRIAC is switched off sometime later, like 50msec.

To shut off the load, firmware first turns on the TRIAC and then the relay off. Since the relay takes around 20msec to switch off, the TRIAC does the work switching (off) the load current.

It's like an SSR with a relay to short it out, to stop the TRIAC from dissipating any heat, and the relay switches no current- it just holds current.
The TRIAC will only ever see load current for a few mains cycles. It is common for a 25-35A TRIAC to be without a heatsink. This is the first clue, a huge TRIAC with no heatsink.

[TrickyNekro]

Why on earth would you switch on the Triac first? The reason you would use both Triacs and relays it would be to control current and soft switch the relay.


And no... there is only one triac connected to only one relay, probably going to a induction fan for speed control...

[johnkenyon]

Why on earth would you switch on the Triac first? The reason you would use both Triacs and relays it would be to control current and soft switch the relay.


And no... there is only one triac connected to only one relay, probably going to a induction fan for speed control...

Re-read what was written: "Firmware turns on the TRIAC (zero-cross) and then the relay."

Then think about the advantages/disadvantages of a triac and a relay when switching AC

Triac: Can be turned on quickly at the desired part of a mains cycle (preferably at a zero crossing to minimise RFI)
Relay: Turns on when it feels like it, possibly drawing an arc as the contacts come together

Triac: Imperfect switch - dissipates heat when turned on and load current is flowing
Relay: Once turned on dissipates next to no heat (until the contacts get tarnished...)

Triac: Can be turned off quickly at the desired part of the mains cycle (preferably at a zero crossing to minimise RFI)
Relay: Opens when it feels like it, drawing an arc until the contact distance is too long, or a zero crossing reduces the voltage.

Put the two in parallel and operate in the right sequence, and you get to use the advantages of one to mitigate/remove the disadvantages of the other.

[TrickyNekro]

Yeah... I read your comment... still it makes no sense to me...

When you have a TRIAC and a Relay in series, why would you turn on the Triac first? So that you can kill the relay contacts or what?

If the relay is operated when TRIAC is off, then you simply do not have to worry about arcing, subsequently damage. The relay is probably there anyhow,
in case the TRIAC is damaged.

Unless you did not get the part that the TRIAC I am talking about is IN SERIES with a relay on the board that I am wishing to repair. And that the only reason
that I mention the existence of the TRIAC is that I found out that in the power supply section of the oven / stove there is a ZCD circuit, and then I wondered
whether or not, they are using the ZCD circuit only for the TRIAC or also the relays.

Unless you knew already, mains frequency is pretty darn stable, so even with the slow relays, they might as well be doing some Zero-Crossing turn on with the
relays, after acquiring timing and "eye-balling" the turn on time of the relay.

And then I said, If they do not use the ZCD circuit to minimize relay damage over the time and they just switch on and off the relays, they might as well be aiming
at planned repair / obsolescence. If you know that out of 1000 samples, your relays turn on after x ms, then you might as well use this statistic to actually turn on the
relays x ms before the zero cross. It´s not an out of the world concept.

And there are no TRIACS parallel to relays on this board. There is only ONE TRIAC IN SERIES with a relay most probably for fan speed control. Nothing else.
I know the differences between TRIACs and relays... thank you! 

[floobydust]

For long-life mains switching, it is common to use both an SSR and a relay in parallel.

They are not in series, they are in parallel. Does this make more sense?