Hello all!

One of the three-aspect signals that I got just recently has become defective, and it has burned the 5211 decoder to which it was connected! Both the signal and the decoder had been working for a few weeks now, and I don't know what actually happened. But here is the sequence of events. I am afraid this is going to be long, so bare with me.

When I turned the power on this morning, the delta unit, through which I feed all the 9 pcs of 5211 decoders that I have, started to keep a strange sound, and none of the solenoid devices worked. I guess the delta booster was tripping due to a short circuit somewhere along the line. So I started to disconnect the 5211 decoders one at the time to find out which one was causing the problem, and found it. When this particular 5211 was disconnected, all the rest worked fine, but when it was connected again, nothing worked.

The next step was that I disconnected all the devices that were connected to the outputs of this 5211. They were actually two three-aspect signals. But still the 5211 unit alone caused the delta booster to trip. You know that Viessmann 5211 decoders have a separate power input socket to supply the power for the outputs. This can be connected to a separate trafo, or with a short piece of wire it can be connected to the digital feed coming to the unit. I am using this second option with all my Viessmann decoders. So next I disconnected this small piece of wire from the "E" socket, thus disconnecting the power from all outputs. After that the unit didn't trip the delta booster anymore, and everything else worked, too.

So based on this diagnostics, it seems that the defective signal has fried something in the printed circuit board of this 5211 on the output side, and when connected to the rest of the circuit, it brings the whole accessory circuit down.

And how do I know that it is the signal which is defective, and that it burned the 5211? Well, I didn't know it first, but after I did the steps described above, I wanted to test the two signals that were connected to the burned 5211 were OK. As I am not any genious in electronics, I simply first connected one of the two signals to another 5211, and it worked fine. Then I connected the other one, and after just a few seconds, "click" comes out of this second 5211 and it is burned exactly the same way as the first one. So the real bad guy was found. It is just pitty that I have to learn my lessons the hard way. So I now have two fried 5211's and one somehow defective signal, which was causing all this.

So finally, here are my questions to you experts:

1. Based on what was described above, do you think there is a chance of fixing the decoders by changing the burned components, or is it not feasible? They "only" cost 28 EUR piece.

2. If you think they could be fixed, any suggestions what components may have been burned?

3. Does anyone have a schematic of the printed circuit board of Viessmann 5211, and list of components?

4. What could be causing the problem with the signal (I haven't had time to get it out yet)? I guess it must be shorted somehow, and therefore shorting the output ports of the 5211 as well, causing it to fry?

5. And yes, I disconnected the power between each step above, so I didn't burn the second unit just because of disconnecting and reconnecting things alive.

6. And yes, I know that I was still stupid to burn the second decoder.

I will probably order a couple of new decoders anyway to replace these ones, but if there is a chance of fixing them, they could always be used as spares, if I manage to burn more.

And I don't know yet if I need to order a new signal, too? Or if it is simply that the studs to which the wires are soldered in the bottom of the solenoid cylinder are bent and touch each other or something like that which can be fixed. Or can it be shorted inside in the solenoid maybe?

Any suggestions will be appreciated.

Pff...

PS. The both fried unit are in this picture:

http://personal.inet.fi/...2004%20Signals%20004.jpg

Thanks, Ron!

Ron, Lars, Jorge, everybody, considering the sequence of events above, any hints of what exactly may have been burned inside? And what about the signal? Any hints that you may be able to give can be helpful.

Thanks again.

Ron,

All signals which are directly driven by 5211's are form signals, Viessmann 4500, 4502, and 4531. The defective one is 4502.

The light signals which I have (4011's) are all driven through Viessmann 5221 module, which takes the pulses from 5211 to change state. I guess 5211+5221 is pretty much like 5213, but it gives smooth transition of the lights from one state to the other.

And the lights in all signals, even in the form signals, are directly fed by a separate light feed, not through 5211's.

Question Ron: If it is output transistor, could that one output transistor trip the whole digital circuit? I mean it is not just that output which is blind, it is that when this fried 5211 is connected to the booster (delta unit), together with all other 5211's, none of them work, and the delta unit gives a clicking sound every few seconds like if it is tripping from over-current or something. As soon as the feed to this 5211 is disconnected, everything else works again. Can one burned output transistor in one 5211 cause this? Even when there is nothing connected to that output, or to the whole 5211?

Hello Kimmo, I agree with Ron and Lars, the BD680 (could be BD682) are normally fried under short circuits. Strange, however, that the protection didn't work. You need some soldering skills to replace the damaged transistors and an Ohmmeter to find out the damaged ones. Each transistor has 3 pins: one of them is connected to the printed track that comes from the "E" input thru a big resistor and a diode. Another pin goes to the output connector. You should measure in the Ohm scale between these two pins. The damaged transistors will show nearly 0 (zero) ohms. Invert the ohmmeter wires (black and red) and will find 0 ohms again, indicating a fried transistor. Have them replaced by new ones.
Hope this can help you to recover your decoders.
In regards to the problem signal, you'd search for a bad connection or short circuit in it.
To avoid to get your decoders fried again, insert a 10 Ohm resistor on the yellow wire before testing the signal. This will limit the current fed to the signal under the limits of the transistors.
Regards,

Jorge,

Should the 10 Ohm resistor be left in permanently, or just during testing? And what is the color code of such resistor? In fact, I just noticed that I have several 50 Ohm resistors. I quess such a resistor does the job, too?

Thanks again,

I think I have found the burned component in the failed decoders.

When I measure the resistance between the E-socket and any of the yellow DC-outputs in the fried units, it shows 0 Ohms resistance. When I measure the same in any of the working units, it shows high (infinite) resistance.

When I measure the first component in the trace coming from the E-socket (I think it is a diode), the fried units show 0 Ohms here as well, but working units show high resistance. So it seems that this component (diode?) is burned.

I also think that I know where the short circuit is in the troubled three-aspect signal. As you know, in a three-aspect signal there are four wires which are connected to 5211 decoder: yellow DC-input, and red, green and yellow pulse inputs. When I measure any of the working three-aspect signals that I have, there is no connection between any of the three pulse input wires when measured against each other, the meter shows infinite resistance in all cases. But in the trouble signal it shows a low resistance of 40 Ohms between the red and yellow pulse input wire, but infinite resistance between all others. So in this signal the red and yellow pulse input wires must be in contact somehow. I don't know how this could have burned the decoders, though, but that seems to be the only difference between working signals and the trouble one. Maybe this has caused two solenoids to pull constantly against each other, and thus drawing constant power from the DC-output causing an over-current situation in the decoder output? I don't know if this is possible.

Any comments on these findings, Jorge, Lars, Ron, others?

Thanks!

It is D12, the first one after the E-socket, and the only diode between the E-socket and all DC-outputs (yellow output ports). Its type is 1N4936.

Diodes D3-D10 are the ones associated with each pulse output as you say (red/green output ports), but these diodes are all OK, or at least they show infitite resistance also in the fried units, unlike D12.

Hi Kimmo, I'm at the office now but tonight when I arrive home I'll take a still-not-assembled 5291 and try to tell you something about components, tracks and possible measuremets you can take.
See you later...

Hi Kimmo, it's me again... Your D12 is fried, no doubt, but you could have more components to replace...

10MOhms when testing a diode is not important. You know, you are 4 hours ahead of mine, but I need to look at the unassembled PCB before giving you more info and this I'll be able to do at about 22:00 today. Will try to explain you how to test all candidate components. If you are awake at 2:00 tonight, you can contact me at jorge_vilarrubi@msn.com for a direct chat in an attempt to shorten the time for having this problem resolved. At your choice.
Regards,

Kimmo:

Sounds like a plan. I don't have a disassembled one to look at, so if Jorge does, that is the best approach. Good luck.

Let me know what you find out.

Hi Kimmo, Lars is right about diodes, they show low R (30-50 Ohm) when tested with the red wire connected to the marked end (white or silver band) and the black wire to the other end. When you reverse the wires, they show high R. As Lars says, D3 to D10 protect the BD680s against voltage from coils when current is shut off.
Let's talk about the PCB: current from CU RED enters the decoder at J17 (E connector) and goes thru D12 to the external track (which feeds the yellow wires of signals/turnouts). This puts a negative voltage on all yellow wires. Current from CU BROWN enters thru J2/J10 to R15 (0,39 Ohm, 2 Watt) which is expected to protect the BD680s by limiting the max. current, then goes to feed the Emitters of BD680s. Collectors of BD680s are connected to the outputs for GREEN/RED wires of signals/turnouts.
How to check BD680s: each one has 3 pins, from left to right (as seen from the lettered face, pins pointing down) they are ECB (Emitter, Collector, Base). These transistors are used as switches, connecting E to C when driven from B.
Put RED wire of your Ohmmeter to B (right side) and alternatively test E and C with BLACK wire. You should see low Resistance (40-100 Ohm). Now put the BLACK wire to B and repeat the test to C and E with RED wire. You should see high R in both pins.
Put RED wire to C (center) and Black wire to E (left), you should see high R. Reverse the wires and you'd see low R, this is an internal protective diode built inside the transistor.
A fried transistor normally shows very low R (same as diodes) in all of the previous tests.
I believe you are using the Ohmmeter correctly, since you've found 0 Ohm on D12.
Verify all diodes and transistors and tell us about your findings. We already know your D12 is bad, so you can replace it anyway. Any other bad component you can find, replace it and then try a test with no signals connected to the decoder. The best testing device I know of is a couple of Brawa lights.
Have fun and good luck, tomorrow we talk again.

Kimmo:

Hopefully you got lucky and the D12 failed before the current could affect anything else. If that is true, it is good to know.

Glad it worked.

Sounds like you have either a problem in the device or cables abraded and touching. You said this just happened suddenly. Were you doing any work in that area where you might have stressed or otherwise damaged the wires going through the board?

Try doing the measurements on the bad blue wires while moving them around as much as you can within the signal column.

Hello task force, good to know the decoders are alive again, will be good too when we discover what has caused the damage.
I don't have a Viesmann three aspect signal but I've repaired two of them.
They have a PCB included in the wiring, covered by a shrink tube.
Both signals I've repaired had problems with diodes within the shrinking tube, so this is a point to be checked. I don't know what has caused the damage on these signals.
Another signal I've repaired presented a problem in the motor, some wires were in touch with others and the isolating paint was damaged, so they had electrical contact and impeded the correct operation of the motors. To open and disassemble the motor is not an easy task, But Kimmo has prooved to be very patient...
Dear Kimmo, you have the whole weekend to investigate... Have fun!
Regards,

Happy birthday to your son, Kimmo!

Guys,

There has been some exceptionally valuable knowledge shared in this post. We need to capture this in a more explict way so that others who have a similar problem can find the knowledge and solve the problem.

I suggest that with some editing, this particular posting is a prime candidate for the material that Lars and Thanos were discussing a few months ago. We could structure as ->
- Sympton
- Test procedure
- Repair
- What else to look for..
- Some credit for the original participants in the solution.
- A link to the original posting..

What does everyone think?

Agreed. My motives are a bit selfish in that I have a few of the 5211s under our layout. Before reading this, my first reaction would have been to buy a new one at A$100.

Thanks

I agree, and I believe Kimmo could take the initial task of describing the symptoms, please feel free to ask me for adding whatever could be of interest.
As time passes, we could build an importante knowledge database.
Regards,

I can certainly do that. But in which form? As a new thread? Under which category? If such database is maintained by the means of this Forum, I think there should be own Forum category for it, named Troubleshooting or similar. In other categories the troubleshooting threads would soon get burried with other threads.

I can open this question under marklin-users.net for Johan to answer. Or what do others think?

As the only option to store this troubleshooting knowhow seems to be an ordinary thread, do you guys still think we should create a shorter and cleaner version of this thread, or leave it as it is?

I agree. Sometimes knowing HOW you arrived at a solution is as important and instructive as the solution itself.