OTGW DIY - AWB ThermoMaster 2HR, doesn't seem to work

[Drak0z]

Hi there,

Let me start by saying I have no experience whatsoever with OpenTherm or the OpenTherm gateway. I am however a well rounded IT professional and a software engineer by profession.
I have received my Opentherm Gateway DIY kit (+USB and LAN/WLAN) today, soldered it together following the specs, connected the USB cable and plugged it into my laptop. Fired up otmonitor (for Windows, as Administrator) and connected. Without the thermostat nor my boiler connected, it seemed to do nothing, which was the way I expected it to behave.
Then I unplugged it, plugged my thermostat's cable in on one end and my boiler's cable in on the other end. After a few minutes (seconds? wasn't downstairs to check) my thermostat went into error and my boiler kept heating up the place.
I unplugged the OTGW, put my thermostat directly to the boiler and it all went smoothly again.
When I fired up otmonitor with the OTGW plugged in properly (with boiler and thermostat connected) I noticed a flood of messages:
The message "<timestamp> T907C0233 Write-Data OpenTherm version Master: 2.20" was logged about 2 times per second.
That's also the only message that shows up in my log (apart form the occasional Error 03 message).

I have tried updating the firmware (both 3.4 and the 4.0b5), both without success. Firmware 3.4 seemed to be flooding my log with Error 03 messages, but I'm not too sure about the cause of that (because it didn't do it constantly).
Now my question is: Did I do (or am I doing) something obviously wrong? Do I maybe have incompatible hardware? If it is an incompatible boiler, that's too big a problem, I'm getting a new one in two weeks time (Intergas HReco 30 will be installed here). If it's something else, well, I just have to figure out how to fix it

My hardware:
OTGW DIY kit (from opentherm-gateway.com ), connected to USB
Laptop running Windows 7, but will eventually connect it to a Raspberry Pi
Honeywell Chronotherm Modulation T8851M
AWB ThermoMaster 2HR (from 2002)

Thanks in advance!

Cheers,
Peter

[pwhooftman]

When you put the OTGW together and it still was loaded with the diagnostic firmware, did you run a diagnostic test (via File, Diagnostics in OTMonitor) ?

Furthermore, see http://otgw.tclcode.com/debugging.html

I remember i had to raise the reference voltage to 6 (VR=6) altough diagnostics where fine

[Drak0z]

Many thanks for the quick response.

I did run diagnostics when it wasn't connected. Since I didn't get much of an output, I thought it'd be okay.
When I had it hooked up to both the boiler and the thermostat, I remember trying to run diagnostics again, getting output that looked sane on #2, a whole list of 4 digit numbers followed by a newline char on #3 and sane looking output on #5. Because I couldn't make much of these outputs, and because otmonitor didn't show anything, I decided to flash the firmware. I will however re-flash the diagnostic firmware and retry these tests when I get home this evening (late unfortunately) and will post results here.

Very newbie question about raising the reference voltage to 6; according to the documents this is normally 1.250, does it actually change anything physically with the voltages, or is it just a divisor/variable used in on screen calculations?

[hvxl]

Your trouble appears to be on the boiler interface, so especially the output of test #3 should be interesting.

The reference voltage setting does actually change the voltage on pin 1 of the PIC, which is used as the threshold for the comparators. Changing it mainly influences the thermostat interface, which seems to work OK for you. So I don't have high expectations that changing it will help much in your case.

[Drak0z]

Thought I'd post a little update here. I actually haven't been able to test anything just yet, but I went to the shop and bought myself a voltage meter (because my old one appeared to be broken). I'll do some measurements over the weekend and will try running the diagnostics after going through the troubleshooting page. Thanks so far for thinking along with me

[Drak0z]

Well, after doing some measurements I am (even more) confused
Power supply topic appears correct: 1.6kOhm on X3, ~5.03 Volts on both pins on IC1 and IC2
OT Interfaces is where I'm getting confused: 26 Volts on X1, but strange measurements in the Amps. Most of the time I'm measuring between 20 and 20.5 mAmps, and only very seldom I'm measuring the (expected, according to the documentation) 6 mAmps. I was thinking of continuing measurements, but figured I'd best stop and ask for directions first. My initial thought was that there could be a short between pins 3 and 5 of IC1, but that gave me a resistance of about 1.4kOhm, so I expect that this is not the case.
When I tried (re)connecting to the gateway via otmonitor I got a long list of "Reads" even though nothing is connected at the moment.I tried flashing in the debugging firmware. The main difference I saw in this screen was that neither the 'Code memory' nor the 'Data memory' bars were filled (before they were at about 60%). I also had to power off and on the OTGW by hand before it started the flash.
Anyway, it looks like it is still (somewhat?) working, as I am able to run diagnostics.

Not connected to anything I figured I could only run the Voltage levels test. These are the outputs:

Code: Select all

Enter test number: 5
Thermostat: 3.22
Boiler: 1.87, 2.18, 2.19, 2.50, 2.88
Reference: 1.21

Opentherm gateway diagnostics - Version 1.1

Then I hooked it up and got the following outputs:

Code: Select all

Enter test number: 2
529,478,521,977,538,448,1038,977,537,448,552,448,552,449,552,448,1039,478,522,478,522,478,523,477,522,977,537,449,552,448,552,449,551,449,552,448,552,449,551,449,1038,976,538,448,552,448,1039,478,522,977,536,450,1038,479,521,467,534.
907C0233  Write-Data  OpenTherm version Master: 2.20
522,486,522,976,538,448,1039,976,537,450,550,449,552,449,551,449,1039,477,522,479,521,479,522,478,522,976,537,450,551,448,552,449,552,448,552,448,552,448,552,449,1039,975,538,448,552,450,1037,479,521,977,538,448,1038,479,522,467,533.
907C0233  Write-Data  OpenTherm version Master: 2.20
504,493,523,976,537,449,1038,977,537,448,552,448,552,448,552,449,1038,478,522,478,568,478,522,478,522,976,537,449,552,449,551,448,552,449,552,448,551,450,550,450,1038,976,537,449,552,448,1038,478,522,976,538,448,1039,478,522,467,533.
907C0233  Write-Data  OpenTherm version Master: 2.20
509,493,522,976,538,448,1039,977,536,449,552,448,551,449,552,448,1038,478,524,477,522,479,522,477,523,976,538,448,552,448,552,448,552,448,552,448,552,448,552,448,1039,976,537,449,552,449,1037,479,521,977,537,449,1038,478,523,466,533.
907C0233  Write-Data  OpenTherm version Master: 2.20
528,478,522,976,538,448,1039,976,537,449,552,449,551,448,552,448,1039,477,523,490,523,478,522,479,522,976,538,448,551,450,550,449,552,448,552,448,552,448,553,448,1038,977,538,448,551,449,1038,478,522,976,539,448,1038,478,522,467,534.
907C0233  Write-Data  OpenTherm version Master: 2.20

Code: Select all

Enter test number: 3
8191.
8178.
8180.
8179.
8198.
8193.
8215.
8203.
4767.
3118.
4772.
5687.
6396.
7041.
7302.
7422.
7539.
7599.
7675.
7719.
7764.
7813.
7838.
7872.
7900.
7923.
7952.
7962.
7990.
8003.

Code: Select all

Enter test number: 4
OK1A high-to-low: 320us
OK1A low-to-high: ### Error
OK1B high-to-low: ### Error
OK1B low-to-high: ### Error

Code: Select all

Enter test number: 5
Thermostat: 0.77, 1.04, 2.23
Boiler: 0.62, 0.87, 1.24, 1.47, 1.65, 1.88, 2.38, 2.62, 2.87, 3.28, 3.51, 3.83, 4.07, 4.99
Reference: 1.21

Code: Select all

Enter test number: 6
B: 19.129
T: 19.153
B: 19.177
T: 19.353
B: 19.378
T: 2.068
B: 2.094
T: 2.119
B: 2.140
T: 5.237
B: 5.260
T: 5.285
B: 5.310
T: 8.401
B: 8.427
T: 8.448
B: 8.473

Did I break something?

[hvxl]

There is definitely something wrong with the boiler interface. You should normally see only two values for the boiler in test 5: 0V and something around 3.4V. When the boiler is not connected you should only see 0V. Is your R4 the correct value and making proper contact to GND and pin 18 of IC1?

[Drak0z]

Thanks for the quick reply. I have measured my resistors (just wanted to be sure and measured everything). All the resistors are as specified on the Gateway schematic at http://otgw.tclcode.com/otgw-wt.gif
Resistances between the R4 and GND and pin 18 of IC1 were either around 4 Ohms or around 334 Ohms depending on which end of the resistor I measured (duh )

Just to be sure, I took a picture with my phone.
http://goo.gl/pwJmbh

The IC2 part has bridges from 11 to 14 and from 12 to 13 as per instructions for USB. I have the feeling I am missing something obvious.

[hvxl]

The resistors all look OK. Next, make sure your transistors are the correct ones. Q1 should be a BC558A (Q4 is the odd one out, being a BC548A. Some people assume the 4 transistors in a row are the same type and so they mix up Q1 and Q4).

But there's really nothing you can mess up on the boiler side of the opto-couplers that could result in measurements of 2V and higher when there's nothing connected to X2. The only explanation I can come up with is that OK1 is damaged. Can you pop it out of its socket and move it over 2 pins so pin 1 of OK1 is in pin 3 of the socket and then run test #5 again without the boiler connected? And just to be sure please also run test #5 with OK1 out of its socket completely.

[Drak0z]

This might be a bit odd, but it was something that I noticed when I started installing. I did not receive 558's but 557's, and a 548B instead of an A. Since there were 4 of them I decided to just install them on their designated positions, I just assumed there had been an update in the design/specs.
BC 557B C46 are installed on Q1, Q2, Q3 and Q5
BC 548B C13 is installed on Q4

Test 5 with OK1 removed (and nothing else connected):

Code: Select all

Enter test number: 5
Thermostat: 3.23
Boiler: 1.14, 1.32, 1.57, 2.13, 2.23, 2.66, 2.87, 3.26, 3.45
Reference: 1.21

Opentherm gateway diagnostics - Version 1.1

Test 5 with OK1 shifted 2 rows (and nothing else connected):

Code: Select all

Enter test number: 5
Thermostat: 3.22
Boiler: 0.93, 1.23, 1.27, 1.83, 1.89, 2.27, 2.66, 2.81, 3.13
Reference: 1.21

Opentherm gateway diagnostics - Version 1.1

Something that I just noticed (and might be nothing), when I open up the diagnostics window, the first thing it tells me is "Invalid test" (and a typo in Voltage levels ):

Code: Select all

Invalid test
Opentherm gateway diagnostics - Version 1.1

1. LED test
2. Bit timing thermostat
3. Bit timing boiler
4. Delay symmetry
5. Voltage leves
6. Idle times

[hvxl]

Lets start with test #5 with OK1 removed and nothing connected. That limits the area to look at to only R4, pin 18 of IC1 and pin 5 of the OK1 socket. In this configuration, the test should only produce a single 0.00V measurement for the boiler. You're getting values all over the place.

Can you hook up your volt meter between pin 5 and pin 18 of IC1 and check the voltage? It should be stable at 0V. If that is the case, please run test #5 again while keeping an eye on the volt meter.

The voltages reported by the gateway and what you see on the volt meter should match and be near 0. If they don't match IC1 is probably broken. If they match, but aren't near 0, you'll have to figure out where the voltage is coming from. In that case I would start with removing IC1 (temporarily power down the gateway while you're doing this) and (with the power reapplied) check the voltage on pin 18 of the IC1 socket again.

[Drak0z]

My volt meter hooked up to pin 18 of IC1 and pin 5 of OK1:

Code: Select all

Enter test number: 5
Thermostat: 3.21
Boiler: 0.00, 0.40, 4.54, 4.80
Reference: 1.21

Volt meter itself showed 0.00 with a setting of 200mV (so that's pretty close to 0 ), before, during and after the tests.


When I disconnect my volt-meter I get these values again:

Code: Select all

Enter test number: 5
Thermostat: 3.21
Boiler: 1.24, 1.55, 1.86, 2.18, 2.47, 2.50, 2.81, 3.13
Reference: 1.21

Since I do see a positive change with the IC1-18 to OK1-5, I was thinking... Can it be some sort of grounding issue? Or might it actually have something to do with me having different transistors? If the latter's the case, I can go into town and pick up the correct types over the weekend

[hvxl]

Pin 5 of OK1 and pin 18 of IC1 are (supposed to be) connected together, so you would measure 0V there. My question was to measure between pin 5 of IC1 and pin 18 of IC1.

As I described, the area we're looking at doesn't include a transistor. So a different type of transistor is not causing this problem. It does sound like there may be a broken track between R4 and IC1, either the GND side of R4 that connects to pin 5, or the other side which connects to pin 18. Can you measure the resistance of those two print traces?

[Drak0z]

That's been a terrible misread from my end, sorry about that. Measured pins 5 and 18 of IC1, no change, everything 0 Volts.
Resistance between IC1-5 and the "OK1 side" of R4 is about 1 Ohm. Same goes for the other side of R4 and IC1-18. If I measure the OK1 side of R4 to IC1-18 I get 330 Ohms. If I measure the other side of R4 to pin 5, I get <infinite>. I think that makes sense, but I'm not sure

[hvxl]

Resistance between IC1-5 and the "OK1 side" of R4 is about 1 Ohm. [...] If I measure the other side of R4 to pin 5, I get <infinite>. I think that makes sense, but I'm not sure

Well, it doesn't make sense to me. If the resistance from one side of a 330 Ohm resistor to a certain point is 1 Ohm, the resistance from the other side of the resistor to the same point should not be more than 331 Ohm. There is something fishy there.