Thermozyklus ZE5 Heating Controller. Bus error

[edugimeno]

Hi! I was given this board. Its supposed to be a heating controller. It can be programmed over othernet from a computer, and can manage different outputs via a "BUS" connector (there's suppossedly an external module where this BUS wires enter with some relay outputs that power on/off heating valves for different rooms

The owner stated something like "I reversed the polarity and although the manual states that's its made to withstand reverse polarity, and would wanr the used onscreen, it ended up broken and would say "shortcut in output" all the time, or something like this

He said that he sent it to the manufacturer and they said they changed "maybe a capacitor" and sent it back. Only component I can see that looks like it might have been replaced is the bus mosfet (FQP 17N40), but he tested the unit at arrival and it would still not work.

When I plug it into +18V it seems to work fine but he says that it shows the error when he plugs the "BUS" wires. Again he confirmed that the BUS wires are not shorted as he has another unit now working on them

He has a bunch of these controller and he tried to swap different BUS outputs to this unit and this same BUS to another unit and it was clear that this unit was the faulty one, there was no shortuc in the BUS connection.
I've disassembled it and not knowing that protocol that "BUS" is, all I could check is that there are +18v constant at the BUS, the 3.3v regulator has 5.0v at input and 3.3v at output, and that's basically it...not sure what else I could test not undesrstanding the BUS protocol.

Has anyone worked with these systems?

Thanks!

[stj]

it says "CAN" so probably CANBUS, look at ic's 10 & 11

[edugimeno]

it says "CAN" so probably CANBUS, look at ic's 10 & 11

You are right in that, but being so close to the DB9 port, I assume there is a canbus connectivity thru that port, but the electrovalve output is a the opposite side and I believe there are completly different systems. I would like to be able to emulate the presence of that valve bus (probably just a given value resistor across) to make it believe it has the bus connected and see the error onscreen.,..,

[stj]

but you dont know what bus is bad.
i also see MODBUS, that's RS485 so look for an 8pin chip for that too.
the software is a bit shit not saying which bus is locked-up!

[stj]

btw, i think your underestimating this unit - it looks like all the conectivity is serial.
even the ethernet port is a lantronics-module - those convert ethernet to serial for cheap mcu's.
not a cheap part btw.

[edugimeno]

No the guy said he doesn't connect anything to the DB9 port, he never used it. All he connects is power (18v) and the orange port labeled BUS which goes to another device that has relays that turns electrovalves on/off.
The unit starts OK and shows normal behaviour until you connect the "BUS" orange port to the wires that feed the valves.

[redwire]

It looks like a proprietary one-wire bus. So that is power+data on the line. I think the big mosfet is a current-source so it can supply power, the electrovalves are powered buy this bus or not?
Check T4 it might be the RX transistor and looks not great. The jumpers match between the boards?

[edugimeno]

It looks like a proprietary one-wire bus. So that is power+data on the line. I think the big mosfet is a current-source so it can supply power, the electrovalves are powered buy this bus or not?
Check T4 it might be the RX transistor and looks not great. The jumpers match between the boards?

Thanks. That could be...I measured voltage while board was on and and it was putting out 18V (same as PS+). And while board=off it has 58Kohm across
I checked T4 and it looks ok, I mean it has 0,6V across leads 1&2 and 1&3 in one polarity, same as most of the other T''s on this board (others are PNP or NPN whatever the opposite)
Yes T4 is directly connected to BUS socket...
Also I checked all small C's and I can't find any shorted one

I can't compare jumpers but Im sure this guy hasn't moved them. I don't have another board with me

I don't know if the relay modules are self powered or powered by this bus but I assume they have their own power as some are stated to manage 30 electrovalves, too much for this 3A power supply

Kind of lost now...

[redwire]

The bus has two aspects - outputs power, and send/receive serial data.
I don't think it should output full 18V (same as PS) at idle? The power mosfet should have some voltage drop, and the MCU turns it on and off I think.
Try put a load resistor, say 500Ω-1kΩ across the bus and see if it's still 18V or if it drops down. Not sure how much power you can get on the bus. The valves might do a sequence whereby only one moves at a time.
If the error message shows up only with a valve connected, then bus power might be collapsing or the RX data seems stuck.

You might have to sketch a partial schematic.

[edugimeno]

I understand. Ok, I will leave this in stdby until 2 wk from now, when I've scheduled a visit to this place to have some measurements. Will probably be able to measure a working bus voltage too

Thanks

[CapLeaker]

A “CANBUS” needs to be terminated to work proper. Problem is you don't now which BUS is bad. I wonder If you put a TTL on that thing and let it boot, if she will show a more exact error.

[stj]

the ethernet is intersting,
some of those lantronics ports have an mcu hidden in them that serves a webpage and/or a telnet server

[redwire]

I wonder- if the owner is "sent it to the manufacturer" for repair and yet it doesn't work? "maybe a capacitor changed" Seems dodgy.

The DB-9 connector has CANBUS and RS-485 MODBUS. JP9 is for CAN terminator. JP4, JP5, JP6 are for RS-485 bias and terminator.
The "BUS" appears to be another serial port, 1-wire. The jumper by the conn I can't see what it's for. "Uart3" JP7 might be interesting to snoop.

FQP17N40 N-ch 16A 400V seems a bit crazy. If it was P-ch I would expect full 18V at the BUS, but no way with an N-ch high side switch. You'd have to sketch things out a bit to see who receives and who pulses power.
Reverse-polarity, I would expect D2 to cover that. T4 connects right to the BUS (its emitter?) so that is why I suspected it got hit.

That jumper might be for a master or slave config, as far as power.

[edugimeno]

A “CANBUS” needs to be terminated to work proper. Problem is you don't now which BUS is bad. I wonder If you put a TTL on that thing and let it boot, if she will show a more exact error.

I don't think a terminator would change much. Actually right now, leaving the BUS wires open (no terminator, no nothing), I get no error. The guy got SHORTED BUS ERROR when he connected it to the bus wires, but the bus wires were actually not shorted as he put a working controller on those and it works fine

[stj]

look for a low value series resistor used for current-sense then.

[CapLeaker]

Hmm… look at the canbus with an oscilloscope. Look at a working one vs the non working unit.

[edugimeno]

look for a low value series resistor used for current-sense then.

Mmmm I checked a bunch of them, and then I found this one. R45 on the attached picture.
It shows "100" which according to tables I found should be 10 Ohm but it measures 1 K ohm in one direction and 5K ohm in the oppostite direction (maybe a D o C is affecting the polarity measuring).
So this does not make sense right?
One of this resistor leads has 100% continuity with of the BUS leads...

[edugimeno]

Ok, this is what I did.
I first tried to see what happened when I put a random value resistor across "BUS". 1st resistor I found was 10K so I tried. I instantly got "BUS IS SHORTED" error on screen.
Then I removed the 10 Ohm SMD resistor and soldered 2x4,5 Ohm resistor in series. Then applied the 10K resistor across bus and got no error.
For now I have no way to fully check board until I take it to he building where it was working and connecting it to its real BUS.
Will be back when I can get these results

Thanks everyone!

[edugimeno]

Ohhhh FAIL!! I went to this place today, plugged my device in the bus, and "SHORTED BUS" came up right away! I measured the bus with no device and it got no voltage (so the voltge is powered by this device) and it showed 200 Kohm.
I could bring home a workind device, as the guy has some spares, so I can compare voltages...

[edugimeno]

Ok, I have some advances here. Take a look at both boards. Left if the board to be repaired, that was already sent for repair to the official "shop". Right is a working board, although the guy doesn't want to use it because this variant doesn't have a LAN port so he can't configure it from his computer.

All seem the same except the LAN port barely visible at the bottom left
Then you will see at my board (left) that I replaced a, supposedly marked "100" resistor R45, by 2x4.5Ohm resistors in series (thru hole but mounted above). Then I go to the new board and this R45 resistor measures 0,1Ohm, not 10 Ohm, ... and I realize it actually shows "R100", which matches the reading.
But ven more, if you look at the main bus mosfet, which a highlighted its marking, one is a P mosfet and the other is N mosfet!!! How could this be?? And I don't see changes in circuit traces to modify this board to use a different mosfet type... So Im guessing that whoever serviced (or tried to service) used the wrong mosfet in it right?
This is something that redwire already rised a flag for it.... That it wouldn't make sense a a N mosfet there....
So first thing I would do is get a P mosfet there.
Also would I change the R45 by 0,1Ohm too? Or let this one go.?

Thanks everyone!
PD: Im also attaching a picture I took before I swapped R45 by 2x4,5Ohm, I can now see the marking was 100% clear: "100", not "R100"