Repair of 2413S PLM When the Power Supply Fails

[Brian H]

C7 and C13 are in the switching power supply. Low ESR.

C3 120Hz DC ripple, C11 60 Hz half wave DC ripple. I would think ripple current.

C8 5VDC from regulator. I don't have an opinion.

 

I believe the ones in the thread are low ESR, 105 degree C or higher and 2000 or more hours rated time.

[BobMic]

While waiting for the parts I drew up a partial schematic showing the power supply which may be of interest to some. 

2413X schematic.pdf

 

Please use caution if making these measurements, especially C3 and C11 which are attached directly to AC line voltage.  I use an AC line isolation transformer when working on line connected devices such as this.  Be Safe!

 

Thank you to all the others that have contributed to this post.

 

Thank you for the schematic in post #282.  I thought I was going to have to draw it myself. One question...it looks like D7 is backwards?

 

I repaired a 5010 which of course failed at 2y 3mo and it has been working now for about 4 months.  Yesterday my other 5010 which is 2y 7mo old failed so I dug back into my bag of caps and swapped them out only to find it completely dead after the repair.  I expect I have a problem with board damage as I had trouble getting a couple of the caps off this time.  I will do some tracing with the help of your schematic tonight.  I will confirm tonight and report back with what I find and also what I screwed up so that the next guy who makes the same mistake can fix it.

 

I also wanted to mention that i tried to use the board from a 2413S with the 5010 daughter board and the firmware must not be compatible.  Anyone have a work-around for that?

 

Thanks to everyone that has contributed to this thread.  You can put me down as 1 5010 EzFlora fixed and 1 more soon to be fixed.

 

Bob

[Brian H]

The LNK354GN  specification sheet's {post 302 here} typical standby reference schematic. Not exactly the same as Smartlabs design is using.

In our D7 position in the feedback circuit. It is a Zener Diode. So it is possible the band is correct for a Zener type diode and not shown correctly in the diagram here.

The schematic here appears to have the chip number mislabeled. As the FCC Photos show an LNK354GN and is what I found in mine.

 

A few Smartenit modules base number is 5010.

5010K Powerline Modem maybe?

Since it can also have an optional RS485 communications added to the Daughter Board. It is possible it had to have different firmware to handle RS232 and the optional RS485 signals also.

 

I too messed up a 2443 hardware V2 with the same base PLM main board. Found I had pulled out a plated through holes core while removing a capacitor.

Had to signal trace and add a insulated wire jumper to fix the busted run.

[BobMic]

Thanks Brian, I got the spec sheet. 

 

That makes sense that D7 is a Zener and R7 is its load.  That would mean that if the voltage at the cathode of D6 exceeds the Zener voltage by ~ 1.2v the transistor in the opto drives a current into the feed-back of U1 to regulate the output.  That would put the Zener at about 18.8v for the readings others have reported (20v).  R6 may have some influence depending on its value but probably there to limit the current into the opto.  Thanks for the tip on the through-hole.  Probably what I need to do.

 

I didn't get a chance to work on it last night and not sure when I will get another chance with Christmas almost here but I will report back once I get a look.

[Harold]

I had an unsuccessful repair a while back. My thoughts have returned to the 2413 issue. A bit of looking around shows that an enclosed 5/12 VDC power supply with line cord can be found for under $10.

 

That is around what I paid for parts in my repair effort.

 

It there any reason not to simply cut the traces/wires from the internal power supply to the PLM active parts and attach the external supply? Simple and easy.

[mwester]

It's not just a matter of supplying DC power -- the Insteon devices also require access to the AC signal as well in order to receive/inject the power-line communications signal, and for the RF side to coordinate its transmission.  So, power-supply repair is required; replacement with an external source is not a solution.

[Brian H]

The unregulated +12 Volts DC. That usually is around 20 volts. Is used in the Insteon power line transmitter.

You also still need the 120VAC on the input. To power the small power line derived DC supply for the Zero Crossing detection circuit.

[Harold]

I think that the posts above are not following my thoughts on the subject. Perhaps I was not clear.

 

There are 2.5 sources of internal power in the unit. From the power line, a bridge rectifier generates 12 VDC. The 12 VDC then feeds a regulator that generates 5 VDC. These two voltages are used to pretty much provide everything on the board with power. Providing these two voltages from an external source rather than the internal components changes nothing. Assuming. of course,  you cut the appropriate existing power traces.

 

The zero crossing detector is a separate circuit fed from the 5V supply.

 

The .5 source I noted above is a circuit that converts 12 VDC to 30 V to operate the interface of the line transmitter and receiver. This is a voltage converter driven by the chip controlling the unit. The 30V source is from a half wave rectifier and regulated by a Zener diode. So the chip is probably not doing anything important on this supply. There is one 470 uf 50V cap in the 30V circuit.

 

If an external 5V and 12V feed is provided (disconnecting the on-board 5V and 12V) , only the components for the charge pump generating the 30V are left. I am not sure that these are among the capacitors that fail. It is simpler if this charge pump is left intact. I do not see any information on how the chip manages this voltage. It is the voltage used for driving the interface to the power line. But it is harder to get a 5V, 12V, 30v external power supply.

 

So, replacing 5V and 12V internal supplies leave only a possible failure prone cap in the 30V circuits.

 

I need to do a bit more work and go back to see if the 30V circuit cap is on the list of points of failure. But, overall, this would seriously simplify the repair process. Minimal dealing with tiny wires and plated through holes. The cost should be similar to buying (plus shipping) all the caps, but seriously reduce the aggravation of the soldering and handling of replacing a bunch of caps.

 

I believe the schematic I am looking at does not completely match the current product; but it is probably close.

[Brian H]

Where did you find a full 2413 schematic.? There is one for the older 2412 PLMs.

There is the power supply area schematic in post #282 of this thread.

 

There is no charge pump in a 2413.

The 2412 did have one. I had two with the self destruct firmware in it. Factory reset and you got a 50% chance.The hidden magic smoke was released. As the Charge Pump control FET overheated and failed. Then the coil in the pump burned up.

 

The Zero Crossing has a separate power line derived supply so it has to stay.

 

Keep posting your updates. Interesting theory. I personally would just change the caps and get on with it.

[Harold]

You are right Brian. I was using a schematic for a 2412. I appear to have mislabeled it when I saved it some time ago.

 

I agree that D7 has to be a Zener. The parts above T1 may be wrong. It looks like a snubber, but it can't snub without different wire destination.

 

Without seeing where everything goes I don't know if using an external source will be useful. If J6 is 12 V, pretty much everything from the bridge rectifiers to the end of the drawing to the right could be bypassed. And that gets rid of most of the problematic caps. If that switcher is making 30V, it gets harder. The drivers for the line and radio interfaces may need that 30V.

[Breezyken]

I revised the schematic in post #282 to indicate that D7 is a zener.  My repaired 2413S has been running for one year now, and I'm not about to open it up again to check any of the parts in question.  If anyone has specific corrections or additions to the schematic for the 2431x V1.B 1351 or similar that contribute to understanding the power supply I will consider revising the drawing again.  T1 actually has multiple windings.  I simplified the drawing of the transformer since I had no way to determine the actual windings and I wanted to concentrate on the caps and their place in the power supplies.

 

As for adding an external power supply, I don't think it is worth the trouble considering you have to account for the signal insertion/detection onto the AC line as well as the zero crossing detector.  Just replace the caps with ones with conservative ratings and designed for a switching application (low ESR)!

[Harold]

I would not mess with those other things. I suspect they are similar to what they did on the 2412. As long as the other processes get power, they should work the way they are supposed to. They don't care where they got it. In the 2412 they are both direct connections to the transformer.

[slcasner]

I'm about to do this capacitor replacement on a 2413S V1.7 1222 that didn't fail all the way (the LED still goes green) but programs would fail to execute.  Power cycling seemed to help, but the problems repeated.  I replaced it with a spare in 2016.

 

Not all of the original replacement capacitors are still available.  Mouser shows the ESX106M400AH4AA 10uF 400V in stock, but KEMET does not show that as a current product (the ESX series does not go up to high voltages) so DigiKey does not have it and I can't find a datasheet to compare.  (I prefer DigiKey because I have an account.)  Is there a recommended alternative?  DigiKey has KEMET's has ESU106M400AH4AA at 10000 hours for "electronic lighting and power" applications and ESG106M400AH4AA at 5000 hours for "electronic ballast, power supplies and long life" applications.  Does either application suggest a preference?  The ripple current at 120Hz is 90ma for the ESU and 100ma for ESG; the ESU has no ESR spec while ESG is 2.9ohms, while both specify and impedence ratio 5. The ESG is rated 350V / 400V surge rather than just 400V like ESU. The physical size for both is 10mm diameter, 20mm length, and 5mm lead spacing.  This is 1mm longer than the ESX, but it looks like there is room.  I see several posts upstream that cdragon chose the Nichicon UPW2G100MHD1TO from Mouser instead.

 

Also, is it a good idea to replace the original 10uF C7 and C13 in an older 2413 with 100uF as in the newer revisions, and if so, is the recommended part still in the KY series (EKY-500ETD101MHB5D)?  I wondered if there was any other component value change that would need to be coupled with the capacitor change.  I just bought another spare 2413 (prompted by the 30% off offer from Smarthome) and I see it has the 100uF capacitors still mounted lying down on the board even though it looks like the board layout was changed to allow at least C7 to be installed standing up.  I can't tell if the inductor is the same, though; it looks the same as in the old unit but there are no markings I can see on either one.

 

[Brian H]

The original capacitor in the line powered side of the supply was a 6.8uF/250 Volt. The 10uF/400 Volt was chosen as the replacement. Higher voltage rating for a better safety margin and capacity for better operation.

It is on a full wave rectifier. I don't think it has to have a low ESR but a good ripple rating at the 120 cycle DC from the rectifier.  The voltage on the capacitor should be around 180 volts DC. So I would think 350 volts would be fine.

 

C7 and C13 being changed to 100uf/50 volts was an update by Smartlabs. I did not find any other obvious other components when they where changed.

I would use the 100uf/50 volt ones to replace the original 10uF/35 volt ones. They are the two in the switching supply output. So they should be a low impedance/ESR type. I have an original V1.0 with the reworked board to add the coil and second capacitor to make it a Pi filter from a capacitor only one.

It has 10uF/25 volt ones. 25 volts is only about a 4-5 volt safety margin as the supply is around 20-21 volts.

[eduardo_garcia]

guys

 

My prepaired PLM died 18 months after I repaired it. I think it is time to use a new one. I got a new one version 2.2. Lets see how much time for a new failure

 

regards

 

Eduardo

[Teken]

Are you going to be sad if I tell you they are on hardware 2.3 now?!?    

[larryllix]

They like to hide the revision number inside a sealed box and you take pot luck when you buy them unfortunately.

[Brian H]

I believe there was a 2.4 mentioned in a thread here.

My 2.3 was installed on 04/21/2017 and is the revision Smarthome indicated a fix was implemented.

[slcasner]

Brian, thanks for the advice.  Here's a picture showing my old 1.7 and new 2.4.

 

[Brian H]

My 2.3 looks the same as your 2.4.

C7 and C13  mounted against the main board. With long leads bent at 90 degrees.

Mine had a small glob of component glue holding them to the main board.

 

On the Serial Daughter Board. You should find the 2.4 has a TI MAX232EI RS232 chip on it. It meets a stricter ESD specification. Over the chip in the earlier revisions.

[JimS]

Amazon sourced 2413S 1/16/2018 with Rev 2.4 3817. Old V1.5 1122 bit the dust.

[JimS]

For the record, not a "newbie" just don't post a lot and live on the forum.

[sun-sense]

Product = ISY 994i/IR Pro (1100)

Firmware = Insteon_UD994 V4.6.2 (2017-06-21-13:52:37)

UI = Insteon_UD994 V4.6.2 (2017-06-21-13:52:37)

PLM #2413S V1.C 1438     -     Main PCB GF-01 94V-O E246338 1418 P/N 200-8790200-023-9  date 20091217

MobileLinc Pro V2.0.7 on Android

 

 

Well I'm pissed !!!

 

I had a few flaky events a month or so ago and then a week or so ago my outside PIR's, all 8 of them, stopped turning on my external floods.  They flashed at me when I walked around but when I checked the ISY event window, sometimes they would appear, and the ISY would send the turn on response but nothing happened.  At other times, the ISY event window shows nothing when I activated them.  I also have MobileLinc Pro on my Android phone by the way.

 

So is it the ISY or the PLM I wonder.  The ISY was giving me lockups when browsing windows so I initially suspected that, but then thougth if it was having communication issue with the PLM, that could be delaying things.

 

Then after I duckduckgo'd, I came across the PLM and capacitor issues.  This is my second PLM by the way.  The other one failed outside the 2-year warranty as well.

 

So I find this thread and having seen that some people had v2.4 in October and JimS just rx'd a v2.4 via Amazon, I thought what the hell.  I don't have the parts and want my external security lights back up, so Amazon it is.

 

So i just received it today.   v2.3 date 4616 which I believe is the 46th week of 2016.  Over a year old...nice !!!

 

Irrespective, I unplug the old one, plug it in and JUST tell the ISY to restore, ignoring the prompt to put the PIR's in active mode as they are high up all around the house.

 

The one PIR I had brought in to test showed up in the ISY event window.  When I walked around the exterior, each and every PIR appeared in the MobileLinc Pro app as well.  When I came back in it was all recorded in the ISY event window.

 

OK - problem solved.  But I am a little pissed to receive a 13+ month old unit with a version that was replaced over 3-months ago so I have initiated a Amazon return.

 

Now, I can try and order a set of capacitors, but if anyone has an extra set they are willing to sell to me please message me.

 

My old PLM #2413S V1.C 1438

 

Thanks.

 

ADD:  This isn't the first issue I have had with Insteon devices.  Three plugins have failed with melted cases.

[Brian H]

If you purchased from Amazon. There are more than one source of the 2413S PLM.

 

If sold and shipped by Amazon it should be the latest.

 

If you got it from one of the Amazon Store Merchants. Then it is possible their stock is not as recent and you got an older hardware 2413S.

[sun-sense]

Direct from Amazon.  "choose One-Day Shipping at checkout"

 

"Ships from and sold by Amazon.com. Gift-wrap available. "

 

"buy from ‘ships from and sold by Amazon’ to ensure you receive the latest revision "

 

 

So much for that.

 

 

ADD:   The PIR's have just sent the Dusk/Dawn On state.  Haven't seen that for about 10-days.

 

Next, the lights will come on when one of the local Raccoons wonders onto the property.  I hope