Repairing a 'Pod' Disco Light & Replacing Filament Bulbs with LED Equivalents

Status
Not open for further replies.

Thread Starter

edwardholmes91

Joined Feb 25, 2013
210
Hello,

I recently repaired a Pod Disco Light, because, instead of sequencing each lamp in turn, they were all lighting at the same time, appearing slightly brighter as the sequence rotated.

On inspecting the PCB, I noted that many of the resistors and capacitors had rusty component legs (possibly due to storage in a damp environment, or excess flux when soldering). Initially I was unsure what the fault was, so I ended up replacing the 100nF capacitors, the electrolytic capacitors (in case they'd dried up), resistors, ICs, triacs and transistors, just in case. Upon further inspection, I noticed that one of the 4 diodes (placed in a bridge rectifier configuration) was actually reversed, thus only providing half wave rectification of the mains to the logic circuitry. I also made sure that there were no dry joints, by heating up all of the solder connections and allowing the solder to re-flow. After giving the underside of the board a clean, I was ready to test the board again and was pleased to see that this had indeed rectified the issue.

I then decided that it would be a good idea to replace the filament lamps with Equivalent LED Lamps. However, after only a short time operating, the same fault re-occurred. I then tested the transistors and triacs and discovered that one transistor and 4 of the triacs had failed. So, at this point, I swapped the LED bulbs back for the filament lamps and replaced the damaged components. As the transistors were a type that I haven't come across before (S9013), I decided to try swapping one out for a BC549B (as I have loads of these) and the circuit still worked. It wasn't until I measured the gain of both a new S9013 and new BC549B, that I realised the collector and emitter were reversed! Therefore, I needed to put them in back to front, for them to work as required (I suspect that the incredibly low gain, when used in reverse was still sufficient to switch the triac, but not enough to amplify the microphone input, for sound operation).

I've attached a picture of the front of the board and one which I started to make, showing the tracks, as I intend to draw out a schematic at some point, to fully understand how the circuit works... So far, my understanding is that the input from the microphone is being amplified and driving the schmitt NAND oscillator at full speed, which sequences a decade counter through the six outputs and I've attached a CSV, detailing the components that I've been able to identify.

So, now I have a dilemma. Do I leave the circuit working as is, with the filament bulbs, or should I look for a solution to get it to work reliably with the LED replacements? I noticed on the PCB, that each output is missing a series resistor/capacitor across the anodes of the triacs. I suspect that the LED replacement's driver circuitry is inducing a voltage spike, that exceeds the 600V rating of the triacs? It would seem the series resistor/capacitors were removed from the original design (presumably to reduce cost), however I don't what values they would have been? But I'd welcome any thoughts on this?

I look forward to hearing your thoughts/suggestions.

Kind Regards

Edward

MOD:Made the Front image viewable.E
 

Attachments

Last edited by a moderator:

Thread Starter

edwardholmes91

Joined Feb 25, 2013
210
hi Edward,
Look at Page 16.3 of this PDF for suppression components.
E
Hello Eric,

Thank you for the PDF. I have to admit, I'm struggling to understand it. Do you think that the LEDs are giving an inductive load, which is rising above the 600V rating of the triacs? And do you think the missing series resistor/capacitor across each output was originally to provide suppression, thereby preventing this? As the original fault, prior to trying the LED bulbs was the same.

I could replace the triacs, for ones with a higher voltage rating, but if I can fit some additional resistors and capacitors to resolve the problem, then I think that'd be best.

Kind Regards

Ed
 

Thread Starter

edwardholmes91

Joined Feb 25, 2013
210
A friend shared this thread with me, which could 'shed some light' on this problem ;)

It would appear that the problem is a combination of the omission of the snubber circuit on the PCB and the inductive nature of the LED lamps. I've since contacted the manufacturer to see if they're able to supply suitable values for the omitted components.

P.S. Attempts to calculate the required values, using an online calculator and values from the triac datasheet have not yielded any meaningful results so far.

Kind Regards

Ed
 

ericgibbs

Joined Jan 29, 2010
10,200
hi Ed,
Those LED lamps appear to have inbuilt SMPS, they would appear as an inductive load.
Use a 39R and a 10nF cap X2 rated for mains.
E
AA1 05-Feb-19 12.31.gif
 

Attachments

Thread Starter

edwardholmes91

Joined Feb 25, 2013
210
Use a 39R and a 10nF cap X2 rated for mains.
Hi Eric,

Thanks, I'm assuming X2 means twice mains voltage i.e. 600V would be suitable? Or does it need to be twice peak voltage i.e. closer to 1000V? Also, what wattage resistor, I'm assuming it doesn't need to be rated as per the lamp wattage... I've got some 0.25W handy or can buy higher wattage if required?

Out of interest, are the values derived or rule of thumb?

Kind Regards

Ed
 

Thread Starter

edwardholmes91

Joined Feb 25, 2013
210
Look at this AAC link for X2 definition.
Hi Eric,

Thank you for the link to the article, it's an interesting read and something I've not come across before, so very handy to know. I've not had chance to read the PDF you attached yet, but I will do and be sure to file it for future reference.

Also, I was just wondering, what wattage resistor do I need to use?

Kind Regards

Ed
 

Thread Starter

edwardholmes91

Joined Feb 25, 2013
210
I would use a 0.25W or 0.5Watt resistor.
Hi Eric,

Thanks for confirming this. I purchased the following items from RS Components:
The components arrived today and I fitted them in series across the anodes of each triac, in the un-used holes on the PCB and tested with the filament lamps. All seemed to work as expected. However, when I tried replacing the red lamp (channel 1) with an LED lamp, problems started to re-occur almost instantly.
  1. The red LED lamp (channel 1) flickered on and off, but otherwise appeared to work for a short while before the familiar fault re-occurred, presumably due to current leakage?
  2. Then I replaced the LED lamp with the original filament lamp and changed the triac on channel 1
  3. This appeared to fix the problem for short while, then channel 4 failed
  4. I replaced the triac in channel 4 and it seemed to work ok for a short while, then channel 6 failed
  5. I then replaced the triac in channel 6... So far, so good!
  6. Spoke too soon, channels 2 and 3 also failed
  7. Finally, I replaced the triacs in channels 2 and 3 - it worked for a short time, before channels 1 and 3 failed again.
I've got one more spare triac, but I'm reluctant to use it, until I can understand what is causing them to blow? On reading the PDF attached to Post #7, I notice that it mentions a minimum value of 47R for the resistor and values of up to 620R. Could it be that the resistor value needs increasing? Or should I try a different triac altogether?

Note that, some of the triacs that I've been using to replace blown triacs had been used in the light before and although they seemed to switch ok, may have sustained some damage previously, which may have led to their eventual breakdown?

It seems almost as though adding the snubber circuits has exacerbated the problem, rather than improved the situation? Although it's difficult to tell, because I can't be sure about the state of all of the triacs.

If you've got any further ideas/suggestions, I'd be keen to hear them.

Kind Regards

Ed
 

ericgibbs

Joined Jan 29, 2010
10,200
hi Ed,
I would say there is an underlying problem/fault on the control PCB or the original Triac's are under rated.
Did you manage to sketch out a circuit diagram as you noted on post #1.?

E
 

Thread Starter

edwardholmes91

Joined Feb 25, 2013
210
...or the original Triac's are under rated.
They're rated for 0.8A and 600V, so no problems current wise as far as I know, but if voltage spikes are occurring, then I guess they could quite easily exceed 600V?

...Did you manage to sketch out a circuit diagram as you noted on post #1.?
I haven't got round to it yet, but ought to, as I still don't understand the full details yet. As can be seen on the front side photo, it looks as though there are a few more missing components... a big capacitor (presumably the same as the other yellow ones), an inductor/transformer and 'Ym1' across the supply, perhaps another safety capacitor?
 

ericgibbs

Joined Jan 29, 2010
10,200
hi,
I have been going over your PCB images, I would say a circuit diagram is not required, there is enough detail on those images.
I guess you know that it is a non isolated mains power supply.?

Which diode did you say was the 'wrong' way around.?
also check D5 the Zener it looks broken in the image.??
E
 

Thread Starter

edwardholmes91

Joined Feb 25, 2013
210
Which diode did you say was the 'wrong' way around.?
also check D5 the Zener it looks broken in the image.??
Hi Eric,

I worked out it was non isolated, as there wasn't a transformer.

With regards to the diodes, it was one of the 1N4007 diodes which are arranged in a bridge rectifier configuration. The one where the diode body is the opposite side to the other 3 is the one I switched round. Now all of the 1N4007 diodes match the silkscreen legend. The zenor diode seems to be working ok... when I measure the voltage on the output side of the bridge recrifier comes out around 10VDC.

Kind Regards

Ed
 

ericgibbs

Joined Jan 29, 2010
10,200
I worked out it was non isolated, as there wasn't a transformer.
hi Ed,
We have a problem.
Now that it is confirmed that this a non isolated mains supply, it contravenes the T&C's of AAC, so we must Close the Thread.

Eric
 
Status
Not open for further replies.
Top