Bathroom Fan over-run circuit repair

[jokskot]

The pcb controlling the over-run of my bathroom fan has failed. I've removed it from the fan which runs fine with the light switch, but I would like to repair and replace the over-run circuit..
On examining the pcb I noted it was discoloured in the area close to resistor R3. I think this had a value of 2.2k but it was difficult to see if the bands were red or brown. I found the resistor to be open circuit and when I replaced it with a 2.2k resistor this glowed as soon as I applied 240v ac. What component of the board will have failed which is causing overall failure and specifically caused R3 to fry? Which component should I replace first?
I include a scan of the board - I tried to convert this into a circuit diagram but couldn't guarantee it would be accurate! - and a commentary on the components and the connections....

 

[throbscottle]

It would be much easier for you to draw a schematic that it is to try and work out what is going on from your notes. If it doesn't make sense at least we have your board scan and notes to compare against. Otherwise anyone trying to help has to work out a schematic from the information you've given, which is incomplete since we can't see which way around the triac is, which is pin 1 of the ic, or the orientation of the diodes, and the tracks are not terribly clear either, I can't tell where all of them go.

 

[jokskot]

Thanks for your response.
As I said in the OP, I can't guarantee converting the pcb into an accurate drawing. I did start with this approach but got very confused. I'll have another go and publish this and a photo of the component side of the board. Equally if you could point me at, or provide, a suitable circuit - I've done a lot of unproductive Googling - please do!
I would then build it.
The fan is <20w the voltage 220-240 and the overrun time needs to be up to 30 mins.

 

[KeepItSimpleStupid]

One of our own (In this thread even) has put together a comprehensive reverse engineering method here: https://www.instructables.com/id/How-to-reverse-engineer-a-schematic-from-a-circuit/

In princple it's easy in concept especially for a single layer, but in practice it's not. For a single sided board, you need to take a photo of the component and solder side and then reverse the solder side image and add some transparency to the component side.

Then build a three layer image of solder side and the translucent component side and a translucent drawing layer. They all have to be scaled and aligned.

Now draw the connections on the translucent top layer.

==

I have purchased a delay on break timer for a bathroom fan many years ago, but the timer is located in the attic and after 30 years, it needs to be replaced. e.g. https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0CD8QFjAA&url=https://www.ssac.com/files/downloads/30/TDMB.pdf&ei=DdwHVPyPLc-4ggSq-ICgAg&usg=AFQjCNGOt4_Ubzj362GbINdXkiRmARU_Bg&sig2=jSbgiL0JoPxxK4OAIi_8dg&bvm=bv.74649129,d.eXY

This isn't what I used the first time, but similar to what I plan to replace it with.

==

If I were to make my own, I'd probably use one of these **broken link removed**

and a PICAXE. www.picaxe.com

 

[throbscottle]

One of our own (In this thread even)

I got a warm cosy family glow now

 

[Nigel Goodwin]

One of our own (In this thread even) has put together a comprehensive reverse engineering method here: https://www.instructables.com/id/How-to-reverse-engineer-a-schematic-from-a-circuit/

In princple it's easy in concept especially for a single layer, but in practice it's not. For a single sided board, you need to take a photo of the component and solder side and then reverse the solder side image and add some transparency to the component side.

Then build a three layer image of solder side and the translucent component side and a translucent drawing layer. They all have to be scaled and aligned.

Now draw the connections on the translucent top layer.

Jesus! - that's vastly over-complicated

I've drawn out the schematics of numerous items over the years, all you need is a pencil or pen, and a sheet of paper - no pictures, or anything else - a possible aid is a torch, so you can shine it through the board, but this is rarely needed. It usually takes about three iterations to arrive at a nice tidy circuit

In this particular case it's an extremely simple circuit with a low component count, dead easy to draw out.

I had a circuit I drew out in the early 70's still on the old workshop wall until we moved buildings, it was for a 'Style 270' stereo record player - a really cheap nasty little system, that used something like AC127/8 output transistors - no service information was available (and we sold 100 of them) so I had to draw the circuit out.

It's a skill any electronics hobbyist really needs to learn.

 

[tunedwolf]

I agree, should take 10 mins over a cuppa to draw out a schematic from that wee board. Of course a couple of decent front and rear photos would help the rest of us draw it out too

 

[jokskot]

Thanks for your input everyone.

Had I used the Instructables Reverse Engineering approach, I don't think I'd have had a circuit drawn out before next Christmas; absolutely fine when applied to the right problem but it's application to my problem made me think of sledgehammers and nuts.
fyi the fan is a Greenwood Airvac 100T mark 2 - I've contacted the manufacturers who were unable to help. It may amuse you to know that one of these fans was sold new on eBay recently and included a competition to win tickets for the 1986 World Cup!

The circuit is shown in attachment jls1004.pdf; it's not conventionally drawn but the linkages and nodes are correct and I think it should be intelligible. I can't positively identify the value of some of the resistors because of discolouration - reds might be purple and vice versa and diodes D1 and D2 are unidentified but I think the polarity is correct. I've also included a rather fuzzy photo of the component side of the board; the soldered side was provided in the OP.

What I want to do is either:
1. Replace any component which may have failed causing R3 to fry. What is/are the most likely cause(s)? What else may have been damaged and should be replaced as a precaution? or
2. Build a simple circuit to replace this which allows the fan to run on for a period of 1 to 30 minutes after the light is switched off. The time need not be continuously variable, discrete values in the stated range would be fine. The momentary switch operated by a pull cord over the shower is a "nice to have" rather than a given.

I can buy from a fan specialist a timer which would probably fit the bill, but it would cost around $20 including postage. Where's the fun in that and I can buy a replacement fan for under $30.

I look forward to your responses.

 

[Nigel Goodwin]

The circuit is shown in attachment jls1004.pdf; it's not conventionally drawn but the linkages and nodes are correct and I think it should be intelligible.

That's why I said three iterations above

Although a circuit a simple as this would only need two.

 

[throbscottle]

Ok, it's helpful to be familiar with at least the E12 series of values, that will help guide you towards the real values of some of those resistors. D1 and D2 are probably 1N4148. I'm not sure about the orientation of D3 and D5, looks like you've drawn them the wrong way.

It could be that R3 has simply failed - it looks like it should be dissipating a couple of watts. I would try putting in a 22K 3W one, then see if the circuit still doesn't work.

And now I have to sleep. Work in morning - Boooooo....

 

[KeepItSimpleStupid]

Comments for now:

My initial comment is that the IC itself would be prone to failures. I ran into a similar situation where this nice fancy thermocouple scanner would refuse to scan when used near a 1000 W SCR controlled IR lamp. The scanner's schematic and the board had a full regulator based design, but none of it was populated. I just clamped the power supply to the CMOS chip. The fix worked on all of the scanners I fixed.

Two things I would add is a snubber for the triac ( an RC network). The large resistor that's popped should be of the metal oxide type and lifted off the board. Clamps should be added across the IC and add a varistor across the line. So, those parts will be optional.

Metal oxide resistors basically basically open when stressed and may actually totally break into two parts. They are also sold as fuseable or flameproof resistors.

There is a good chance that the IC is damaged and possibly the Triac and the popped resistor.

 

[jokskot]

Thanks to all for the input.
Throbscottle...
I got all the diodes the wrong way round on the diagram. I attach an amended version.
I'm not certain that D1 and D2 are identical, but D2 certainly has 148 printed on it so you're correct about its identity
and KISS...
Your point about R3 acting as a fuse may well be correct as the resistor which I originally replaced it with (which glowed on application of power) was only 1/4w. I've ordered a 22k metal oxide 3w as suggested and will install and report back.
If it turns out that neither replacement of this nor the triac make the circuit functional I think I'll call a halt as replacing the IC would exceed my soldering skills.
Should this be the case, can anyone provide a simple circuit which I could build from scratch which would achieve the same end? It could include an IC as I'd use a socket for any new build

 

[throbscottle]

D1 could actually be a zener, it makes sense like that. So, you should check D1 for open or short, and C1 for short. I'm wondering if a 555 timer could be used for an alternate circuit, only 8 legs instead of 14!
HTH

 

[KeepItSimpleStupid]

The IC is easy to remove destructively. i.e. you destroy the IC in the process. You cut the pins and then remove each pin separately.
A desoldering tool such as braid or a hand pump help.

See: https://www.amazon.com/s/ref=nb_sb_noss_1?url=search-alias=aps&field-keywords=desoldering

 

[Diver300]

The power for the IC and the triac gate come from a transformerless supply, using R3. R3 has 240 V across it for half of each mains cycle, when the diode is conducting. That gives about 1.3 W dissipated in the resistor, all the time, whether or not the fan is on.

Along with the fact that there isn't much cooling around the resistor, is why it gets so hot. It really needs a higher power resistor, or a circuit that actually shuts down.

Fans like that take as much standby power as most TVs!

 

[jokskot]

Excuse the long delay since anything added here. A weeks holiday and two weeks in hospital to blame.
I've replaced R3 with a 22k 5W and replaced the triac, but get only 5v at the output pins.
I think the time has come to build a circuit from scratch around an inexpensive (555?) IC. Can anyone provide a suitable circuit diagram, please?

 

[throbscottle]

Not to rain on your parade, but it would be pretty quick to just test all the other components with an ohm-meter, you'd soon spot the dud. OTOH, rolling your own is fun Sorry I don't have any suggestions in my pocket right now...

 

[KeepItSimpleStupid]

And you have never replaced the IC. Cut the pins at the top of the IC, use a solder sucker and /or solder braid to remove. CMOS chips are very fragile. Cutting the pins isn't easy either, but it is with the right tools.