SMT Diode id & home automation system repair

[Resqueline]

Hey, that's not bad at all! Congrats!
Well, at least you'll have the parts if it decides to blow a few years down the line.
Do put something in series with the wiring to the sounder, just to be safe.
It could be interesting to see what the upgrade might consist of. Maybe there was a weak spot.

 

[mypcbis$crewed]

You had recommended a test light and that is a good suggestion, but I think that I will brave crawling through the rafters, yet again, to rewire with a larger gauge wire. Apparently 24 gauge is too thin for this application - 20ga or 18 would be a better choice according to manufacturer, although it did work! Albeit briefly. :O As for the upgrade, I havent received it yet, but judging from a picture, all the parts look the same. It would be interesting to see if the diode that blew along the burnt trace is the same or an upgrade. Will let you know as soon as I get it

 

[mypcbis$crewed]

I received the board - the trace was jumpered via a thin wire and the LM was replaced. Everything else on the board is the same except for an upgraded IC. I installed the unit and left out the sounder to perform a power up check. I think there is still an issue. There is a small disc type capacitor that gets hot to the touch very quickly! It is located very close to the LM that was repaired. I will get a photo up of it in the next few days... Needless to say, I powered down and started to troubleshoot a little. I removed everything and powered the system up bare (no inputs or outputs just the board - disc cap was ok, didnt heat up). I tried with just the sensors attached, same, disc cap. ok. When I added the two keypads to the load, the cap started to get very hot... The wiring to the keypads is also cat5 and according to spec should be larger gauge. I think that my first step is to rewire those and then see if I still have the same problem. Thats where I am at in the saga... any thoughts?

Thanks for keeping up with me

 

[Resqueline]

No prob's!
Hmm, I see no disc cap, only F1 & F3. And I can't see thin wiring causing an overload.

 

[mypcbis$crewed]

It is F3 that is getting hot. I am off of work tomorrow, so I will power up the board with one keypad and see if it gets hot. If it doesn't I will check the draw on the keypad to see if maybe the cumulative draw between two keypads and all the sensors is causing the cap to get hot... The system worked for two + years before so I am not sure that this is the issue, however, it might have been hot before and I never noticed it either....
If the 1 keypad causes the cap to get hot, I can remove the keypad from the wall and wire it at the board location with a real short run and see if that makes a difference. I will let you know tomorrow
- John

 

[Resqueline]

The F's are PTC fuses that are supposed to get hot as they approach their spec'd current.
But I came to think of one possible reason for thin wiring overload, and that is if the load has a built-in switching regulator.
Say the keyboard is supposed to receive 12V but runs on 5V 1A internally. It would try to draw 5.5-6W from the supply regardless of its voltage.
So if it gets 12V it would draw 0.5A but if it receives only 6V it would draw 1A. The main board would have to supply 1A at the full 12V's - making 12W.

 

[mypcbis$crewed]

Interesting point, I will check what the voltage going out to the keypads is registering at!

 

[mypcbis$crewed]

The plot thickens....
So, the F3 cap gets warm only with one specific keypad and not the other. I removed the affected keypad from wall, checked continuity on wiring (looped 2 pairs of wire, each had about 3.4 ohms resistance per leg) - ok. Backside of keypad shows a burnt diode! So that is why I am getting the F3 hot. It must have burnt out at the same time the mainboard blew out. The question is now, how they worked together to kind of see why that occured and to prevent it from happening again. This was not the keypad that I was using when the board blew so it wasn't like it was actively being used to draw more current. It is about 30ft longer of a wiring run than the other keypad. The diode that blew on the keypad is a smc made by General Semiconductor with markings of BM then directly underneath that 45 and the General Semi logo. I have asked DigiKey to id the component for me because I couldn't find any more info on it as far as what would replace it.
Here is a photo of the keypad

A close up of the damaged area, I couldn't get it clearer but its D5 that is burn out.

One side of D5 has continuity with the red lead. The black lead has continuity with D4,3,5 on the side closest to the edge of the board.

 

[Resqueline]

Always nice to find the reason behind problems.
I find that it's practically the same transient protection diode as on the main board.
SMBJ15A = SMBJ15CA = SMB = LM = BM
Being connected across the power (?) confirms this function.
You say you looped 2 pairs of wire for measuring, each having about 3.4 ohms resistance per leg. Does that mean the panel in normal use sees 3.4 ohms worth of wiring for its power supply? This would give a short-circuit current of 13.8V/3.4=4A.
But if the diode is shorted how could the panel work? (Or did it?)
I don't see any switchmode regulator on the panel btw., only a linear one if that's what it is.

 

[mypcbis$crewed]

Well, at least getting closer to finding the problem

I am with you on converting the part number from the smbj15 until you get to the last step. I see it referenced in the datasheets to LM, but how did you make the last leap to BM. Also I couldn't see a marking to tell me where the cathode is, does that mean that the part is actually bi-directional?

Yes, D5 has continuity with the red lead on the interior side of the board while the black lead has continuity with the side closest to the edge of the board.

There where 2 pairs of wires coming in, so I looped them to create continuity and tested back at the main panel. There was continuity and the total reading per pair was 3.4 ohms or 1.7 ohms per leg. (i.e. red lead would have 1.7ohms resistance and so would the black lead)

The diode is toast, but the odd thing was that the keypad still lit up and display looked as if it was receiving data from the main board! There was an alarm from the keypad when the unit was powered on whereas the other keypad did not set off an alarm, I assume to warn the user of a fault.

Yes, below C4 is a LM78MOS cdt. What are your thoughts on why you are asking about whether there was a switchmode regulator on the keypad or not? Were you thinking if there was one then the full power would not have been drawn onto the board?

Thanks again

 

[Resqueline]

I just stumbled across a (cryptic) cross-reference when searching for smd code smb bm diode.
Yes, you are right about the bidirectional bit, I had forgotten that was what the C meant.
Ok I understood you right about the wire resistance then.
It must be a 78M05, a 5V 0.5A linear regulator. It was just to establish how the current draw would react to the wire resistance; that is no change at all (down to 7V) with a 5V linear.
There could be something else wrong with that panel. Could it have been hit by static electricity? Check the in- & output voltage from the 78M05 anyway, and what becomes warm. Knowing the wiring resistance could also enable you to calculate the current draw (if you dare hook it up that long).

 

[mypcbis$crewed]

I am not sure what you mean by 'C' from : "just stumbled across a (cryptic) cross-reference when searching for smd code smb bm diode.
Yes, you are right about the bidirectional bit, I had forgotten that was what the C meant."

I'm sorry, I miss typed. It is a LM78MO5 cdt. not MOS

I guess I could wire the VOM in series at the main panel and measure the current and turn it on briefly and then power it down. As for checking the in and out of the 78mo5, I think I will have to replace the diode first before attempting that one. I could try it at the main panel that way I could power it off quickly but then I would be using a short test lead or alligator clips and that would negate the reason for performing that test no?

Well, just found out the roof was leaking, so of course that is going to take precedence! Hopefully it will be nothing more than a little tuck and squirt ....it's never ending lately.

 

[Resqueline]

The page I found (couldn't find it again right now) had dual rows not referring to bi- & unidirectionality and wasn't clear as to which code belonged to which part number, but when you mentioned it I remembered that the suffix CA means bi- and just an A means uni-. (SMBJ15A vs SMBJ15CA).
Guess that page I found wasn't the only one being cryptic..

I wouldn't hook it straight to the main panel w/o having the protection of the wiring resistance. I wouldn't trust replacing the diode would get rid of the "short". You could get much higher currents and "release the grey smoke". If you're able to connect & disconnect the power (quickly) at the display I'd measure the volts & feel the temps right there.
Since the display kind of works it can't be a full short and then it's not certain the diode is drawing the power (now) but something else.

I know what you mean, problems often comes in clusters. I had to fix leaks in a small roof over the basement exit at my homeplace last summer, and then there's a problem with the main drainage pipe blocking up that needs to be adressed this coming summer.

 

[mypcbis$crewed]

I understand what your saying on the main wiring. I guess I can use the rest of the reel of cat5 I have (about 2-300 feet) to test the keypad at the main panel. That way I am close enough to disconnect the power if something should be getting hot.

As for the diode on the keypad it shows that it was shorted somewhat:

D4, D3, D5 the three diodes closest to the pin/wiring harness connector show the following on VOM set to 20k resistance check:

D4
6.09k forward 6.09k reverse

D3
7.15k forward 7.15k reverse

D5 - burnt diode
5.20k forward 2.24k reverse

All the diodes have the same marking of BM 45 (and apparently they are bi directional).

Alas, more rain.... baby decided to projectile vomit while feeding, out the mouth and nose causing the 2.5 y.o. to freak out right before bedtime....shoot me

 

[Resqueline]

If you have a reel of wire you can use that's a nice precaution & procedure. I guess D3 & D4 are on comm lines so I'd expect them to measure a bit different than the power line anyhow, but it will reveal itself if it gets hot when powered up.

Btw., I took the liberty of expanding the title of this thread to better describe the contents. Hope it's ok with you.

Ah, the blessing of the little ones.. How are they able to make us want & love them anyway? Scary.. I'm not quite sure what to think about me not being "blessed" though..

 

[mypcbis$crewed]

Title is a-ok with me. I hope it helps others in some way.
We feel blessed to be home - all of us Still no answers but the little one is doing better so we are content with that for now. After I catch up with work, bills, taxes, roof leaks and cheating death I will resume this thread. Thanks for your support.
Feel blessed, there is always something you have that someone else doesn't

 

[mypcbis$crewed]

And we're back...

Ok, picking up from where we left off....I wired up the refurbed system with all zones, sensors and only the working keypad - no sounder. These are the readings:

Typical zone outputs on the board have a 13.43 DC voltage when nothing is attached.

All zones that have a 1k resistor on them and a motion sensor attached register at 2.14-2.15 VDC. This is really odd because the manufacturer states that their devices work from 10-14 VDC and consume at standby 12.5mA at 12vdc and 100mA when alarmed. Does this make any sense? There are 6 of these devices.

The horn sounder section of the board has two outputs and they both read 10.32 VDC with no load attached. The sounder is rated to operate between 6-13.8 v and draw 175mA.

This board was fixed by the manufacturer and supposedly tested before being shipped back.

I tried to get a reading on current being used, but couldn't?? I don't know if my meter isn't functioning or operator error The scale goes all the way down to 20 microamps, so I am sure we are covered as far as range goes, but hmmm? I don't know what to make of that.

So far the board has tested out with all devices attached except the sounder and bad keypad and everything works well. Alarm triggers and dials out, remote sensors work, etc. I didn't have a chance to test the bad keyboard yet, but I am really nervous to test the sounder again It might be possible that the devices are drawing collectively more than the 1A limit of the board.
Thanks!

 

[Resqueline]

A colleague came by my office today. I sniffed & asked what's burning? Oh, it's from my clothes, we just smoked the battery tester was the reply. Sigh, what have you done now.. Ok, let's have a look at it.. So I went to their workshop, opened it up and located a couple of burnt (but not off) ground tracks. My conclusion was that they had taken off the positive connection and allowed it to touch the case, thereby shorting out the battery via the negative wiring that went to the case internally (negative connections are rarely fused). Oh, maybe we shouldn't reconnect that battery then they pondered.. Eehhh, heh, l don't think there's a problem with that I said and hooked it up. Nothing else seemed to be damaged.
I had a PC-supply blow up on me on the table at home a couple of days ago but that's another story. Smelled bad for quite a while and I got myself a cleaning job..

Well, over to your challenges.

I'm not sure what the 1k zone resistor is about.. How can I learn more about that? Power to motion sensors shouldn't be passing through such high value resistors.

The horn output voltage is just caused by a leakage (or tamper sensing) current and doesn't mean much unless the alarm is triggered.

Measuring currents is a challenge to most people, I'm not sure how to help you there.. Sometimes test leads plugs are to be moved.

Do you have a 9 or 12V battery or a 6-12V wall-wart? You could use that to test the sounder and keyboard.

 

[mypcbis$crewed]

LOL, glad I'm not the only one making smoke

The 1k resistors are to monitor the zones (essentially allowing the system to know if the wiring has been compromised, EOL resistors) More in depth here http://www.diyalarmforum.com/diy-alarm-faq38/

Good call on the horn voltage ouput, I will have to check what the voltage is when triggered! Wish I had thought of that before I posted

I actually used alligator clips w/ leads (insulated ) to connect between the VOM and the zone being tested for current but still came up empty handed. I will have to see if I can borrow an analog meter from someone. I used to know how to read and connect one of those

I do have the 12v 7mA lead battery that is the backup for the system for testing and have used it successfully to sound the horn on the existing wiring. It sounded without any issues. I should probably try it again and see what current is being drawn.

The smoke detector theory is interesting. Can you elaborate a little for me on how the electrons are moving so I can understand you better. I sometimes get confused on the whole particle movement issue..... the neg. particles are supposed to go to the positive so in the smoke detector situation when the positive lead grounded with the case electrons from the systems battery? raced through the wiring and out through the + lead to earth ground? Have I got that right? That would seem likely for my scenario as well, but the battery was connected to the system at the time

Thanks again for your help, it is much appreciated!

 

[Resqueline]

You mean the battery tester. And by 12v 7mA you mean 12V 7Ah. Just nitpicking!
It was an external battery being tested, connected by allogator clips and wires with banana plugs going into the tester. Then they unplugged the positive banana plug and let it drop onto the case. Since the positive alligator was still on the battery and the negative wiring was also complete and connected to the case internally, the battery got a full short-circuit, burning the weakest spot(s) along the round-trip. It had happened once before I noticed, a track being patched. The negative wiring should have been fused just as well like the positive is (internally). I measured such a 6V 3.4Ah battery to be capable of delivering 30-40A shorted..