>>> .. How to fix it if the 0.33uF capacitor has failed .. <<<
The Watchman Sonic power supply unit/ display meter in the kitchen threw a wobbly.
It was blinking in a funny way – see video.
(Just in case it confuses you the calibration table taped on top of the watchman sonic on the video is my own)
So I tried re-pairing the part that goes on top of the tank with the power supply/display meter a number of times.
The first time I tried this a few months ago it worked.
Then I had to do it again a month or so later and it worked. Then a month or so after that it would not pair at all!
(I had also tried replacing the battery in the part that sits on top of the oil tank. A new battery made no difference)
It had failed.
So I cut all around the edge of the power supply gently with a hacksaw and got inside.
See attached photos below.
>>> No obvious component burn out <<<
So I bought a new complete Watchman Sonic as a replacement.
Then paired the old part that goes on top of the tank with the new wall power supply/display meter. Those parts, old and new, are now working together perfectly.
So I know the problem is with the old wall power supply/display meter.
Since this consists of
- power reduction electronics from 240v a.c possibly to 5 v d.c. TTL
- other electronics including display electronics
can anyone suggest which component(s) have failed (lots photos below) ?
Bear in mind the parts that have failed made a slow failure as described above. Capacitor? Still not sure, and can remember little of my electronics.
Looking forward to your replies !
Thanks 🙂
>>> UPDATE 1 , Thu 09/10/2014 <<<
Turns out that this component (photos below), after searching http://uk.rs-online.com/web for “B32922” (as shown on component) is a 305V polypropylene film capacitor.
- The film will have failed (slowly, which fits with what was observed above).
The capacitor part search is
and the choices of capacitance offered on the RS site are 0.047µF, 100nF, 150nF, 220nF, 330nF, 470nF, 680nF.
However the manufacturer site :
seems to suggest (?) this capacitor part may no longer be made. So the offerings on the RS Components site may be incorrect as far as a replacement goes.
I cannot identify the component properly from the three above photos. Anyone who understands the codes on the device able to do this ?
>>> Can anyone identify this capacitor properly, so we can identify a replacement <<<
>>> Since I reckon this is the failure component that everyone else is getting can anyone recommend a robust non failing replacement ? <<<
🙂
>>> UPDATE 2 , Fri 10/10/2014 <<<
According to page 5 of the following doc
http://www.epcos.com/inf/20/20/db/fc_2009/X2_B32921_928.pdf
the capacitors in the EPCOS beginning part number B32922 all vary slightly in size.
There are only 2 capacitors that are close in size. These are :
- 0.22uF, 7.0 mm × 12.5 mm × 18.0 mm, B32922C3224+*** ◆, 3320 3600 4000 2
- 0.33uF, 8.0 mm × 14.0 mm × 18.0 mm, B32922C3334M*** ◆, 2920 3000 2000 2
The capacitor device size I measure on the Watchman sonic, as best I can is : 17 mm x 7.5 mm x 13 mm, with a lead spacing between the two pins underneath the device of 15mm.
While the top of the capacitor is clearly marked “33 M 305V” indicating a possible 0.33uF device the size is closer to the 0.22uF. The devices are cheap.
The farnell web http://uk.farnell.com/ lists these as
Brand = EPCOS, Part No = B32922C3334M000, Description = CAP, FILM, PP, 330NF, 305VAC, RAD, Farnell order code = 1112844, cost each = £0.40.
or a
Brand = EPCOS, Part No = B32922C3224M, Description = CAP, FILM, PP, 220NF, 305VAC, RAD, Farnell order code = 1112842, cost each = £0.37.
an exact size comparison with existing device should give away the correct replacement.
I will get some ordered up, try them out, and post an update 🙂
Farnell web was £20 minimum order and couldn’t think of anything else to order to bring the price up.
Searched B32922C3334M000 and B32922C3224M on http://www.ebay.co.uk and got two of each for £7.72 delivered. Awaiting delivery ! 🙂
>>> UPDATE 3 , Tue 14/10/2014 <<<
This part, B32922C3334M000, just arrived by post. The 0.33uF capacitor. It’s the same size as the existing.
Here is where I got it www.ebay.co.uk/itm/EPCOS-B32922C3334M000-CAPACITOR-CLASS-X2-0-33UF-305VAC-/181228519169?pt=UK_BOI_Electrical_Components_Supplies_ET&hash=item2a320fbf01
So I got the soldering iron out, removed the old one, fitted this new one, taped the unit back together, switched on, and IT WORKS perfectly !!!
Paired it with the part that sits on top of the tank and tested it by moving it to various distances above the floor and the meter reads correctly 🙂 🙂
WATCHMAN SONIC FIXED !!
>>> UPDATE 4, Mon 20/10/2014 <<<
If Ur interested to know the purpose this capacitor serves on the circuit board have a look at this simple explanation from St Andrews Uni.
http://www.st-andrews.ac.uk/~www_pa/Scots_Guide/audio/part5/page1.html
As to why so many PSU’s (eg phone chargers), computers and other electronic devices have failed in recent years, this page
http://en.wikipedia.org/wiki/Capacitor_plague
suggests that “A researcher is suspected of having taken the secret chemical composition of a new low-resistance, inexpensive, water-containing electrolyte when moving from Japan to Taiwan”, thus explaining all the early device failures.
EPCOS, http://www.epcos.com, the manufacturer of this capacitor is a TDK company and whether the original, or even the newly purchased replacement capacitor, is of an old batch made in Taiwan with a defective electrolyte formula is not known.
If the new replacement fails I will source a different : much higher spec, upped voltage capacitor, replacement.
ps .. pls click on images for higher res to identify components
Hm, thanks for the photos. I have a working one of these and I was pondering how to hook it up to a home automation system. From your photos I see it’s using a Si4320 FSK receiver. Probably in ‘standalone’ mode where it’s fairly dumb. Have you done any more investigation?
Do you have any use for the no-longer-broken one? I was baulking at buying a whole new one just to play with it but if you’ve got a spare one you don’t need and have already disassembled…?
Hi David,
I was going to put it on ebay as I managed a full repair and re-assembly; I never got round to it. I have no idea how to hook it up to home automation.
I do however run an arduino ATMega2560 with ethernet shield, connected to the home network (and via dydns to everywhere else) that allows me to control much of the house lighting via the arduino served web page.
I have intended to replace the Honeywell heating controller and the radiator thermostats with auto wireless TRV’s though time and other projects have beaten me there.
The watchman sonic; I think it is just a matter of decoding the received FSK signal from the TX on the top of the fuel tank, after sorting out what the pairing does.
I imagine the pairing simply allocates to both the TX(tank measurement) and RX device(meter) the same 8-bit code from 256 anti collision possibilities (avoiding confusion between neighbours watchman sonics).
I don’t know that chip. Is the TX/RX freq 433.92MHz?
If you use a software oscilliscope (simple RX receiver wired to audio in jack with Audacity) and get the code sequences pls post up.
🙂
Hi, thanks for the quick response.
Firstly, there’s another teardown at https://hackaday.io/project/1548-rocket-receiver-teardown with a little more information.
As it happens, I have just this moment bought a spare receiver from eBay to play with.
Yes, the Si4320 is capable of 433.92 MHz. It is controlled from a PIC over a SPI connection. It should be fairly simple to sniff the serial traffic and see precisely how it’s configured.
You can also buy identical modules (RFM01-433-D) which can easily be connected to an Arduino or something else. In fact I’ve just bought four of them because I couldn’t buy only one at a time. If you want to email me your address, I can shove one in the post for you (for free).
Once I’ve sniffed the traffic on the new unit that will arrive soon, it should be relatively simple to hook up an RFM01 to an Arduino and listen to what the sensor is sending.
Funny you should mention the Honeywell heating system. I’ve just this month installed an Evohome system and wireless TRVs everywhere. And then hooked it up to the Domoticz home automation software…. which is why I’m now looking for what *else* I can connect 🙂
I was originally planning to run wires for 1wire temperature sensors and wired radiator actuators in every room, but the Honeywell system seems to do the job relatively well. A friend has it, and he convinced me that battery-powered actuators and wireless communication *weren’t* the constant pain the the arse that I had suspected they would be.
Dammit, I was impatient and way trying to use simple probes instead of waiting for a proper SOIC probe that clips on. In the process, I managed to blow up both my new receiver *and* the Bus Pirate I was trying to use to sniff the SPI traffic. Now I’m back to looking for another receiver on eBay but there don’t seem to be any cheap options at the moment. How much would you want for yours (assuming I throw in a RFM01 unit as part of the payment)?
Hi David,
I just saw http://www.domoticz.com/wiki/Lua_-_Oil_Tank_Monitor
Do you have any info on packets that come off the Watchmansonic 433Mhztank transmitter?
Sure. It’s supported by rtl433 now. See https://hackaday.io/project/1548-rocket-receiver-teardown for all the fun details.
Replying with a different account so it doesn’t sit in moderation… yes, rtl_433 supports it out of the box now and it’s been feeding tank levels into Domoticz for me for a over a year now.
Thanks David
I got the RTL-SDR (newest one I think) out the other evening and tried it. I got it so I could at some point get round to sniffing the packets off a Honeywell wireless CM927 central heating controller, so I could replace it with a PID one with web interface. Never had time.
Anyway I hooked it up to sniff for the watchman sonic packets on 433.92Mhz using various commands eg
rtl_433 -d 0 -R 42 -A -q
I did not see anything. Perhaps I have the command wrong or am too far away and need to bring the transmitter beside the computer for testing.
Eventually I hope to get it all working with one of these http://www.ebay.co.uk/itm/5x-433MHz-RF-Transmitter-Receiver-Kit-Module-Arduino-ARM-WL-MCU-Raspberry-PI-/162102463039?hash=item25be0f4a3f:g:xikAAOSwnNBXXm2E
and Arduino. If and when I do I will post instructions.
I use one of those 433MHz TX’s already with an arduino mega and ethernet shield as a web server to set various sockets and wall lights on/off as I click the link on the page.
I tried Domoticz and couldn’t get it going on Lubuntu; looks good.
Ok. I got this. How to interpret? Any brief pointers would be useful.
I don’t see the tank level/ acoustic sensor encoding.
ps think I’m better with Manchester pulses off a software oscilloscope. I can meaningfully read those.
$ rtl_433 -d 0 -R 42 -A -q
Detected FSK package
2017-01-03 12:58:23 : Oil Watchman : 8bb0bb8 : 80 :17 : 20.000 : 0 : 70
Analyzing pulses…
Total count: 50, width: 35110 (140.4 ms)
Pulse width distribution:
[ 0] count: 1, width: 769 [769;769] (3076 us)
[ 1] count: 32, width: 257 [255;264] (1028 us)
[ 2] count: 16, width: 516 [516;521] (2064 us)
[ 3] count: 1, width: 7 [ 7; 7] ( 28 us)
Gap width distribution:
[ 0] count: 2, width: 776 [773;780] (3104 us)
[ 1] count: 31, width: 259 [256;264] (1036 us)
[ 2] count: 16, width: 516 [515;520] (2064 us)
Pulse period distribution:
[ 0] count: 2, width: 1419 [1296;1542] (5676 us)
[ 1] count: 24, width: 516 [515;521] (2064 us)
[ 2] count: 15, width: 774 [771;781] (3096 us)
[ 3] count: 8, width: 1032 [1031;1037] (4128 us)
Level estimates [high, low]: 15960, 1
Frequency offsets [F1, F2]: 2612, -15311 (+10.0 kHz, -58.4 kHz)
Guessing modulation: No clue…
Detected FSK package
2017-01-03 12:58:23 : Oil Watchman : 8bb0bb8 : 80 :17 : 20.000 : 0 : 70
Analyzing pulses…
Total count: 50, width: 35109 (140.4 ms)
Pulse width distribution:
[ 0] count: 1, width: 769 [769;769] (3076 us)
[ 1] count: 32, width: 257 [256;264] (1028 us)
[ 2] count: 16, width: 516 [516;521] (2064 us)
[ 3] count: 1, width: 14 [14;14] ( 56 us)
Gap width distribution:
[ 0] count: 2, width: 772 [772;773] (3088 us)
[ 1] count: 31, width: 259 [256;264] (1036 us)
[ 2] count: 16, width: 516 [515;520] (2064 us)
Pulse period distribution:
[ 0] count: 2, width: 1415 [1288;1542] (5660 us)
[ 1] count: 24, width: 516 [515;520] (2064 us)
[ 2] count: 15, width: 774 [771;780] (3096 us)
[ 3] count: 8, width: 1032 [1031;1037] (4128 us)
Level estimates [high, low]: 15945, 2
Frequency offsets [F1, F2]: 2626, -14337 (+10.0 kHz, -54.7 kHz)
Guessing modulation: No clue…
Detected FSK package
2017-01-03 12:58:24 : Oil Watchman : 8bb0bb8 : 80 :17 : 20.000 : 0 : 70
Analyzing pulses…
Total count: 50, width: 35109 (140.4 ms)
Pulse width distribution:
[ 0] count: 1, width: 769 [769;769] (3076 us)
[ 1] count: 32, width: 257 [255;264] (1028 us)
[ 2] count: 16, width: 516 [516;521] (2064 us)
[ 3] count: 1, width: 13 [13;13] ( 52 us)
Gap width distribution:
[ 0] count: 2, width: 772 [772;773] (3088 us)
[ 1] count: 31, width: 258 [256;265] (1032 us)
[ 2] count: 16, width: 516 [515;520] (2064 us)
Pulse period distribution:
[ 0] count: 2, width: 1415 [1288;1542] (5660 us)
[ 1] count: 24, width: 516 [515;521] (2064 us)
[ 2] count: 15, width: 774 [771;781] (3096 us)
[ 3] count: 8, width: 1032 [1032;1036] (4128 us)
Level estimates [high, low]: 16004, 3
Frequency offsets [F1, F2]: 1724, -14335 (+6.6 kHz, -54.7 kHz)
Guessing modulation: No clue…
Detected FSK package
2017-01-03 12:58:24 : Oil Watchman : 8bb0bb8 : 80 :17 : 20.000 : 0 : 70
Analyzing pulses…
Total count: 50, width: 35108 (140.4 ms)
Pulse width distribution:
[ 0] count: 1, width: 768 [768;768] (3072 us)
[ 1] count: 32, width: 257 [256;264] (1028 us)
[ 2] count: 16, width: 516 [516;521] (2064 us)
[ 3] count: 1, width: 10 [10;10] ( 40 us)
Gap width distribution:
[ 0] count: 2, width: 774 [773;776] (3096 us)
[ 1] count: 31, width: 258 [256;264] (1032 us)
[ 2] count: 16, width: 516 [515;520] (2064 us)
Pulse period distribution:
[ 0] count: 2, width: 1416 [1292;1541] (5664 us)
[ 1] count: 24, width: 516 [515;521] (2064 us)
[ 2] count: 15, width: 774 [771;781] (3096 us)
[ 3] count: 8, width: 1032 [1031;1037] (4128 us)
Level estimates [high, low]: 15891, 14
Frequency offsets [F1, F2]: 4172, -13937 (+15.9 kHz, -53.2 kHz)
Guessing modulation: No clue…
This is a great hack. Thank you.
I have a question.
Looking at the binary interpretation of the manchester encoding,
here http://david.woodhou.se/watchman-decoded.png
on the hack here https://hackaday.io/project/1548-rocket-receiver-teardown
We have
[1] 0010 1000 [2] 1100 0110 [3] 1101 0110 [4] 0001 1101 [5] 1000 0001 [6] 0100 0000 [7] 0101 0001 [8] 1000 0011
(I have numbered the above bytes shown in square brackets [] for convenience).
This manchester encoding shown in hex is
28 C6 D6 1D 81 40 51 83
The hack describes the bytes as being as follows
——– Byte 1, 2 & 3 watchman sonic Unit ID ——–
This is the unique unit ID which is reassigned each time you pair the device
——– Byte 4, flags ———
0001 when magnet is held to transmitter, and 0110 when leak alarm occurs.
——— Byte 5 (6 bits of the 8 bits only) ———-
top 6 bits are temperature in deg C using the equation
[ -(reading x 5) + 145] / 3 = temp in degC
>>>>> Byte 5 (remaining 2 bits only & Byte 6 (all 8 bits) <<<<<<<
Depth approximately in centimetres. Temperature compensation incorporated.
Depth = 0 is a failed reading. Readings under 10cm unreliable.
QUESTION:
Using rtl_433 and linux and command
rtl_433 -d 0 -R 42 -q
I have the following output
[1]2017-01-04 18:17:58 : [2]Oil watchman : [3]28DCE63B : [4]80 : [5]26 : [6]5.000 : [7]0 : [8]107
I have numbered the above with square brackets [] for convenience. This numbering does not relate to the byte acount above.
[1] is obvious, [2] is obvious.
Does 28DCE63B use up 3 bytes as stated or 4, or does it incorporate the flag as described?
What is 80?
Is 26 the temperature before arithmetical conversion?
What is 5.000?
107 is the depth. I know that from moving the receiver around.
You’d find life a lot easier if you used the JSON output format, which labels the values 🙂
Also see https://github.com/merbanan/rtl_433/blob/master/src/devices/oil_watchman.c#L57
Great thread, thanks/ I know it’s old but my Watchman is having the exact same issues. I managed to get into it, but schoolboy error , I nicked off a ceramic capacitor (the blue one next to the one you identified as faulty)
I don”t know how to identify a replacement. It appears to have printed on it:
SC101K
SWC
K
SL8250W (cant really see this)
Photo here: – https://imgur.com/a/F0qf2Oj
Any ideas?
Hi David,
SC101K means that it is a 100pF capacitor with +/-10% tolerance.
(“101” is interpreted as 10 x 10 to the power of 1 pF, and the “K” is the tolerance, K being +/- 10%)
This video explains all
I think I see there 25V ac not 250V ac, on my capacitor, though I could be wrong.
You could try this one. After extensive searching on RS Online and Ebay this is all I could find
https://www.ebay.co.uk/itm/CAP-SUP-X1-Y1-100PF-250VAC-RADIAL-DE1B3KX101KA4BN01F-Fnl/391793026199?hash=item5b38af2497%3Ag%3AQrIAAOSwAPVZJH17&LH_BIN=1
It’s a 250V ac one, which the original may well be since the writing on it is a good deal less than clear.
(It will be plenty over capacity if it’s supposed to be 25V ac in any case)
Let us know how it goes
……..so I need a blue thing to fix mine …now just need someone who can solder. I have three broken ones 🙂 Didn’t understand most of this, other than the eBay link to “Its fixed” What I did understand was interesting. Thank you.
Thanks Chevaune,
Mostly it is the capacitor that goes and needs replacing.
See “>>> UPDATE 3 , Tue 14/10/2014 <<<" above.
You will need to check the link above, which I have copied and pasted below, is still a match for the component, though it's likely it is
http://www.ebay.co.uk/itm/EPCOS-B32922C3334M000-CAPACITOR-CLASS-X2-0-33UF-305VAC-/181228519169?
Soldering is not that tricky. There are plenty youtubes that show how.
Let us know how it goes.