Fixing a Remotely Controlled HVAC Unit Circuit Board (Essentially a Thermostat?)

[Curmudgeon10]

This might be a longshot…

I have a ductless ("minisplit") HVAC unit in my shop.  It is controlled via a hand held remote.  The receiver for these remote commands is a small circuit board in the front cover of the unit with two sensor lamps.  While cleaning the filters behind this cover a few weeks ago (cover raised up) the cover slipped from its mounting lugs and fell on the floor.   In so doing, five small wires were pulled from the terminal block on the circuit board as the cover crashed down.  These wires originate in the guts of the wall unit and it’s not clear I can access them there to determine their order.  The order in which they must be reinserted in the terminal block on the circuit board in the cover is the current mystery. 

I can post pictures of everything if I can figure out how to do it on this forum.  So far none of my tries have been successful --- I keep getting the message that the images are too big, but they are all about 2Mb.

I'm competent enough to measure voltage and resistance --- and if instructed how, how to repin those wires in what is an awfully small block.  The point now is to see if there is anyone out there with some experience in this area that might help me, through further posts and pictures, get this unit back in operation.  A replacement has been quoted for $5K!  Oh, BTW – none of the HVAC tradespeople I contacted wants anything to do with this project.  They just want to sell me a new one.

[james_s]

It would help if you posted the model number of your specific unit, somebody may have the same or similar unit. Does the owners or installation manual show any photos of the unit with the cover off? If you're getting errors about images being too large, use a program like Irfanview to resize them before uploading.

[fzabkar]

Imgur is an alternative image hosting site.

[Curmudgeon10]

It would help if you posted the model number of your specific unit, somebody may have the same or similar unit. Does the owners or installation manual show any photos of the unit with the cover off? If you're getting errors about images being too large, use a program like Irfanview to resize them before uploading.

I'll try answering the question about the unit with a photo.  File Properties says the image is 3.86MB.  As backup, in case thie .jpg doesn't load, the unit model number is Chigo QC1H18A/QCOH18A.

First, to head off some other suggestions that might be offered, please be advised this is an old unit, probably dating from 2003.  It was made in China, but the tradename and its products are still active in the United States.  I have attempted to contact the manufacturer --- no replies after several attempts --- AND I've reached out to several stateside distributors of Chigo products, with no luck.  I've even corresponded with some overseas distributors in South Africa, and one in Australia.  Yes, there are two very brief schematics in the owner's manual, but neither deals with this part of the circuitry.

I resized all images before posting to well below 5Mb.

There are no photos in the owners manual with the cover "removed," as it is never intended for the cover to be removed during normal operations.  I suppose it would be possible to do so by first removing the circuit board installed in the cover that receives the remote control signals, which would leave the wire connections to that board intact, just not sure why one would want to do that.  Thanks for your interest.

[Curmudgeon10]

I see my latest attempt at loading a photo was successful.  I'm going to load some more.  First, two pics showing the front and back of the circuit board.  The "front" is the surface facing the user when they operate the remote.

[Curmudgeon10]

The final photo is of the wires that have become disconnected.  I'm pretty sure the red wire is power.

[themadhippy]

with nothing more to go on id be tempted to go black brown red orange yellow

[sokoloff]

With the wires disconnected, measure the DC voltages across pairs of them. I'd start with the red probe on the red, orange, and yellow  (in turn) and the black probe on the black (first) and brown (second). The goal is to find a pair (hopefully exactly one) that has about +5V from one to the other.

The negative of those will connect to the pin farthest from the CN4 silkscreen designation.

The one that carries power will connect to the pin closest to the CN4 silkscreen.
The next two farthest from the CN4 will be drive signals for the two LEDs (coming from the control unit to that satellite PCB), green being in the 4th position, and yellow in the 3rd position. (This might help you figure out which is which if you know what LED is lit [if any] when the unit is initially powered on.)
The last [2nd from the CN4 designation] will be the IR signal back from this board to the main controller.

The two LED drives probably have a current limiting resistor (that the positive supply will not have). If you have some spare discrete LEDs, you could connect LEDs across the various lines that show positive supply to the positive pin and put the negative pin to the negative wire and see which wire has the LEDs burning too brightly (and quickly burning out): that one will be power, going to the one closest to CN4. If the other one or two light the LED but don't burn it out, the worst that will happen is you'll get the LEDs reversed and have to fix that later.

[Ian.M]

As a first step it would be worth trying to identify the ground wire by checking continuity to chassis ground (with the mains power off).

Simple voltage measurements are likely to be ambiguous if the LEDs are driven from CMOS logic, as the outputs are typically rail to rail, and the IR signal wire may well have a pullup resistor to the +supply which will make it measure almost exactly the same voltage as the +supply wire.  Repeating each measurement with a 10K resistor connected in parallel to the multimeter and noting the resulting voltage drop will rapidly show which wires are fed via resistors on the main board, so, by elimination, locate the power wire.  Also the voltage drop will tell us the approximate resistor value on the main board, and therefore if that wire is probably safe to connect to the IR sensor.

[Swake]

Found a pic with a similar IR / LED combo. Source:
https://www.ptacsolutions.com/CHIGO-QUIETSIDE-TURBO-AIR-331400074-PRINTED-CIRCUIT-CONTROL-BOARD-With-2-LED-Lights-FOR-MODELS-QSHE-181-TAS18V-AND-OTHERS-Important-Confirm-number-of-LED-light-on-existing-before-ordering-_p_201232.html

[Swake]

There are Chigo boards for sale On aliexpress. An example in this link, but there are more:
https://www.aliexpress.com/i/1976370038.html?gatewayAdapt=glo2nld

Different models probably, but an important constant: that connector is having the wires in the same order on both sides. This means that if you can find out the order on the 'unbroken' mainboard you have a solution.

[sokoloff]

Impressive searching, Swake!

[Ian.M]

The pin order at the main board end in Swake's photo goes yellow, red, ?, ?, ?  so *if* its the same at each end, we can be fairly certain the red wire is *NOT*  +Supply.

[TheMG]

Looks pretty straightforward to figure out based on the components present on the board.

One wire will be ground/common, to the negative side of that electrolytic capacitor.

Another wire will be likely +5VDC, to the positive side of the capacitor

Another one will be data from the third pin of the infrared remote sensor which is not ground or VCC.

That leaves only two wires, one for driving each of the LEDs.

With a multimeter, you can identify the ground and 5V pretty easily and connect them to their respective terminal. Once you get those figured out, very likely the rest of the wires will go in the same color order as they are on the main board end. If not, then trial and error would work at that point (connect one at a time to the data pin of the IR sensor until the remote works).

[fzabkar]

I would think that the IR sensor's output would be pulled up to Vcc. The LEDs appear to be grounded, so no pullups. However, that should still be useful because simply measuring the resistance between Vcc and each of the three signal pins should differentiate the LEDs from the IR sensor.

[Curmudgeon10]

I want to say very much appreciate all the suggestions for troubleshooting!  It will take me some time to digest.  Practically speaking, I'd like to extend those wires via some temporary splices to I can troubleshoot on my workbench, several feet away.  And how do I go about getting the bare wire ends back in that terminal block?

[sokoloff]

If the existing connector is not obviously repairable, I would desolder it [remove it entirely] and either solder the wires directly to the PCB or replace it with a new connector of suitable pitch and that you can get the pins for.

[Curmudgeon10]

The wiring to the connector measures 1mm in diameter.  What gauge should I look for to splice to it?

[Swake]

There is not much current going over those wires, so about any type or diameter of wire will be good enough.
The length of the wire will always have an impact, keep it as short as possible.

[Curmudgeon10]

It may take me a few days (lots of other things going on), but I'll report back on how it went.  Again, many thanks to those who offered advice.

[james_s]

A good source of wire like that is old computer cables. Pick up a 25 pin serial cable or a parallel printer cable at a thrift store and cut it up, you'll get several feet of wire in a rainbow of colors. Some cables have nicer insulation than others but the wire is generally similar.

[Ian.M]

... or hack up an Ethernet patch-lead.  Patch-leads are stranded wire, but fixed Ethernet wiring is (or rather should be) solid core.  You wont get all the nice colors, but it is an expedient source of minimally color coded hookup wire.

[Curmudgeon10]

One final question before I take this on: is there a risk of killing something permanently on the board (or even on the other end, where the wires dive into the wall unit itself) if I mess up a trial connection order?

[sokoloff]

There is. It’s small, but not zero. Sending a voltage in to a pin that’s not intended to receive it, or connecting two outputs together and having one try to drive low and the other high can cause problems. Or connecting a 5V supply directly across the LED could burn it out.

Connecting boards together without knowing the wiring is not zero-risk.

I think using reasonable care to figure out what’s ground and power goes a long way to reducing the risk.

[Ian.M]

Yes.  If you mis-identify +Power and/or Ground, and connect them to the front board in the wrong place, you could blow the IR receiver, one of the LEDs, or possibly even something on the main board.  There's also a much smaller risk that a mixup between one of the LED wires and the IR signal wire could damage the IR receiver, if its using low enough series resistor values for the LEDs.

If there's a part no. on the underneath of the IR receiver, it could be replaced, and LEDs are cheap and easy to swap, so it wouldn't necessarily be game over, but would certainly cause delays and expense.  If you damage the main board its anyone's guess how hard it could be to fix - anything from replacing a single small cheap part to a complete write off requiring a new board , if its badly designed and the power supply burns up

Make the suggested measurements, and if you aren't 100% certain you've positively identified the power and ground wires don't experiment!  If in any doubt, post your measurements here and wait for us to analyse them.

The safest option would be to dig into the unit enough to get decent photos of both sides of the main board, which would let us figure out the wire colours and functions from what they connect to, only leaving the possibility the LED wires might need swapping.  Sure its a pain in the butt, to get to it, but you are gambling with at least $50 of boards, and that's if you are willing to wait for them to come on the slow boat from China!