Reverse engineering a salvaged pcb and brushless DC motor

Thread Starter

JunkieNL

Joined Mar 19, 2020
54
Hi,

I got hold of a mainboard and fan motor of an inside unit of an airconditioner. These parts got replaced under warrenty. I don’t know in which way the unit was malfunctioning, but replacing these apparently fixed the problem. Since I don’t know much about these motors I thought it would be fun to see if I could get it working.

I could not find any datasheet for this motor and there is very little information on the website of the manufacturer, so I tried to reverse engineer the mainboard.
1640702100763.png1640702112646.png
The motor is connected to the white 6-pin on the bottom left.

  • pin 1: Vm goes straight to another connector that is labeled 310 Vdc. There is an AC filter on the bottom right of the board, but the rectifier is missing.
  • pin 2: not connected
  • pin 3: GND
  • pin 4: Vcc is behind a 15V regulator.
  • pin 5: Vsp is behind an optocoupler. When the optocoupler is off, Vsp is pulled to ground. When the optocoupler is on, a voltage divider should give 6V. My guess is that 6V PWM can be used to control the motor speed.
  • pin 6: PG is front of an optocoupler, so this is used as an output. It is pulled up to 15V.

I hooked it up without the mainboard, using Vm = 92 Vdc (instead of 310 Vdc). Vcc = 15 Vdc. Vsp = 6V PWM @20 kHz with variable duty cycle. The motor runs fine and I can control the rpm by chaning the duty cycle.

1640702140289.png

Next step is to see if I can replace the 3 power supplies and the function generator. If I am ever going to use this motor I want a closed box solution, which I feed mains AC and can control through an optocoupler. My first thought is that maybe I can salvage the original PCB for that purpose, by just adding a mains rectifier and external controls. It turns out that the pcb also needs an external 19V to power Vcc, so I don’t think it is worth the hassle to salvage this pcb.

A few questions remain:

  • There is way more electronics on this part of the board, which seem to be for motor control. For example a dual OpAmp, 2 transistors and even more connector pins. Is any of this needed for reliable operation or could I do without? (It seems to run just fine without.)
  • The PG pin does not show any interesting information on the oscilloscope, with or without the pullup resistor. How is this supposed to work? What information should it provide? Is this the part that is broken maybe?
  • If I ever want to operate this at full power, is that as simple as rectifying 235V AC mains into an electrolytic capacitor (400V, 100uF) with bleeder resistors? I could also salvage the AC filter that is on the pcb. Adding the low voltage part should be rather straighforward.
 

MaxHeadRoom

Joined Jul 18, 2013
28,619
It is marked Brushless DC, (BLDC), so you will have 3 power conductors and other conductors for the means of electronic commutation.
 

MisterBill2

Joined Jan 23, 2018
18,179
I see a total of ten multi-pin connectors on this circuit board, as well as a CPU with a very large number of pins. It may be that the motor driver is a part of this board, the photo is too blurry to see if there are driver parts or not, although drivers may be on the other side. My point being that it is not likely that there is any way to use this circuit board to make that motor run.
 

Thread Starter

JunkieNL

Joined Mar 19, 2020
54
What I have learned so far:

  • The BLDC motor has an internal controller and has a 6 pin connector:
    • 310 Vdc
    • nc
    • GND
    • 15 Vdc
    • 6,1 V PWM (dutycycle controls motor speed)
    • PG pin as output (function still unknown)

  • The driver part of the board is actually pretty isolated from the rest. It receives 310 Vdc and 19 Vdc and has 2 digital inputs and 1 output. The inputs and output are through an optocoupler.
    • Input 1: turn the controller on and off.
    • Input 2: Control motor speed through PWM.
    • Output: follows PG pin (function still unknown)

1640861588511.png
 

MisterBill2

Joined Jan 23, 2018
18,179
OK, I guess. I was not able to read the tag on the motor in the previous photo, so I had no clue. Looking at this photo I do not see parts that would be a controller for a brushless motor with an external driver. There is a question now about the supply voltages, which is do they have one common negative connection? 310 volts DC could be from rectified 220 mains or from a voltage doubler from the 120 mains supply. And that 310 volts will need some currrent capability to spin a motor that size, at least an amp and probably 2 or 3 amps. So either there is a supply that we have not seen or the supply was not replaced and so you did not get it.
If you can show a clear photo of the while label on the motor then we can provide additional advice. Of course, if the motor was the failed part it may still be a failed part.
 

Thread Starter

JunkieNL

Joined Mar 19, 2020
54
To fully understand what is needed to drive this motor, I want to find all the other functions of this driver board. This is the circuit to the right of the white connector. I left out the capacitors. The OpAmp is a Unisonic LM393G.
1640864274895.jpeg

I believe I understand what this is supposed to do. It turns the PWM signal to the motor (Vsp) off when the motor voltage (Vm) is not available or too low. How does this actually work?

What I just read about positive feedback in an OpAmp is that when V+ > V-, then you get full positive saturation and when V+ < V-, you get full negative saturation. In this case the reference voltage on V- is 4.2V.

Lets say that initially the 310V is not connected. Then V+ is about 12V. This is more than the 4.2V on V-, so the OpAmp would fully saturate to about 15V and this will pull Vsp to ground. What I expected.

Is this about right so far?

Next I will try and apply Kirchoff and calculate V+ when the 310V is connceted.
 

Thread Starter

JunkieNL

Joined Mar 19, 2020
54
OK, I guess. I was not able to read the tag on the motor in the previous photo, so I had no clue. Looking at this photo I do not see parts that would be a controller for a brushless motor with an external driver. There is a question now about the supply voltages, which is do they have one common negative connection? 310 volts DC could be from rectified 220 mains or from a voltage doubler from the 120 mains supply. And that 310 volts will need some currrent capability to spin a motor that size, at least an amp and probably 2 or 3 amps. So either there is a supply that we have not seen or the supply was not replaced and so you did not get it.
If you can show a clear photo of the while label on the motor then we can provide additional advice. Of course, if the motor was the failed part it may still be a failed part.
I believe the actual driver is indeed internal. The motor can be controlled by just feeding DC voltages and a PWM signal. Here's a clearer image of the previously posted photo:
1640871116951.jpeg

The motor is only 27W, so that is less than 100 mA at retified 235 Vac. So I was thinking of using a DB107 rectifier and a 100uF, 400V capacitor.
 

MisterBill2

Joined Jan 23, 2018
18,179
To fully understand what is needed to drive this motor, I want to find all the other functions of this driver board. This is the circuit to the right of the white connector. I left out the capacitors. The OpAmp is a Unisonic LM393G.
View attachment 256418

I believe I understand what this is supposed to do. It turns the PWM signal to the motor (Vsp) off when the motor voltage (Vm) is not available or too low. How does this actually work?

What I just read about positive feedback in an OpAmp is that when V+ > V-, then you get full positive saturation and when V+ < V-, you get full negative saturation. In this case the reference voltage on V- is 4.2V.

Lets say that initially the 310V is not connected. Then V+ is about 12V. This is more than the 4.2V on V-, so the OpAmp would fully saturate to about 15V and this will pull Vsp to ground. What I expected.

Is this about right so far?

Next I will try and apply Kirchoff and calculate V+ when the 310V is connceted.
The LM393 is a dual comparator, not an op-amp. The positive feedback is to provide hysteresis so that it does not oscillate as the switch threshold voltage is crossed. It may be that the other half of the LM393 serves as the duty cycle generator, if you trace that part you may find the circuit that you need.

And with the photo readable I see that the motor power is not as much as I had guessed.
Now the big challenge will be the power supplies, because without some isolation they will be connected to the mains directly, at least the 310 volts. And "PG" might possibly be "power ground". For that part you will need to find a circuit for the air conditioner to understand how the power connections must be done.
 
Last edited:

Thread Starter

JunkieNL

Joined Mar 19, 2020
54
The LM393 is a dual comparator, not an op-amp. The positive feedback is to provide hysteresis so that it does not oscillate as the switch threshold voltage is crossed. It may be that the other half of the LM393 serves as the duty cycle generator, if you trace that part you may find the circuit that you need.
I thought all OpAmps were also comparators, but now I get that there might be a difference between an actual comparator and an OpAmp as comparator. I will look into this later.

The input of the other comparator is just connected to ground. I believe the PWM has to be sourced external to this part of the board through one of the optocouplers.
 

Thread Starter

JunkieNL

Joined Mar 19, 2020
54
The LM393 is a dual comparator, not an op-amp. The positive feedback is to provide hysteresis so that it does not oscillate as the switch threshold voltage is crossed. It may be that the other half of the LM393 serves as the duty cycle generator, if you trace that part you may find the circuit that you need.

And with the photo readable I see that the motor power is not as much as I had guessed.
Now the big challenge will be the power supplies, because without some isolation they will be connected to the mains directly, at least the 310 volts. And "PG" might possibly be "power ground". For that part you will need to find a circuit for the air conditioner to understand how the power connections must be done.
I don't think PG stands for "power ground", because motor is working without the PG connected. The PG pin feeds the input of an optocoupler, so I believe this might be a sensor output. I have not yet figured this out. I did not see anything usefull on the oscilloscope.
 

MisterBill2

Joined Jan 23, 2018
18,179
OK, that is reasonable. On some power supplies, "PG" means "Power Good", indicating hat the processor reset command can be removed.
If you have the motor running already, I am wondering what you are using for the 310 volt source. That is a bit high for common experimenter power supplies. And it does require quite a bit more caution.
 

Thread Starter

JunkieNL

Joined Mar 19, 2020
54
OK, that is reasonable. On some power supplies, "PG" means "Power Good", indicating hat the processor reset command can be removed.
If you have the motor running already, I am wondering what you are using for the 310 volt source. That is a bit high for common experimenter power supplies. And it does require quite a bit more caution.
Form my original post:
I hooked it up without the mainboard, using Vm = 92 Vdc (instead of 310 Vdc). Vcc = 15 Vdc. Vsp = 6V PWM @20 kHz with variable duty cycle. The motor runs fine and I can control the rpm by chaning the duty cycle.
The 92 volts are coming from 2 power supplies in series (31V+61V).

Next thing is to rectify mains in a safe enclosure and then control the motor speed by low voltage PWM through an optocoupler.
 

MisterBill2

Joined Jan 23, 2018
18,179
OK, I had lost that part from the original post. Probably you can recover the opto from the original board, since if the motor is OK the board probably had a fault of some kind. Or possibly not. I saved a client a lot of money once by resetting a washermachine computer that the service guy said needed to be replaced for $375 + labor. REading the instructions plus unplugging it for 5 minutes was not that hard! I got a good tip on that service call. This happened in my post-retirement handyman career.
 

MaxHeadRoom

Joined Jul 18, 2013
28,619
What I have learned so far:

  • The BLDC motor has an internal controller and has a 6 pin connector:
    • 310 Vdc
    • nc
    • GND
    • 15 Vdc
    • 6,1 V PWM (dutycycle controls motor speed)
    • PG pin as output (function still unknown)

  • The driver part of the board is actually pretty isolated from the rest. It receives 310 Vdc and 19 Vdc and has 2 digital inputs and 1 output. The inputs and output are through an optocoupler.
    • Input 1: turn the controller on and off.
    • Input 2: Control motor speed through PWM.
    • Output: follows PG pin (function still unknown)
So it appears to be similar to the modern HVAC motors where the commutation circuit is internal to the motor, the 3ph is also produced internally.
 

atferrari

Joined Jan 6, 2004
4,764
I am really enjoying your posts.
When doing the pre-surgery to replace the micro in a printer (no schematic availabe) found useful to identify every single pin connected to common.

What I learnt also is that expressions like that PG could explode in your face when reading the user manual.

Buena suerte
 
Last edited:

Thread Starter

JunkieNL

Joined Mar 19, 2020
54
Previously I had the motor running without the PCB and it seemed to work at these low speeds. I cannot test at full speed because I only have 92 Vdc available instead of 310 Vdc. I wanted to see if the motor could run with the PCB.

This is what I did:
-connect motor to white connector on pcb.
-feed 92 Vdc to pcb for Vm
-feed 19 Vdc to pcb for Vcc
-feed 20 kHz PWM signal to optocoupler for Vsp
-short output of other optocoupler to turn the pcb on

This is what happened:
-the motor is running
-speed does not respond to duty cycle (although this works at 20 kHz without the PCB)

So I used my oscilloscope to see what is happening and it looks like the optocouler has a rise and fall time of about 100 ms. So it cannot follow the 20 kHz signal and ends up being stuck at an intermediate voltage and speed.


Turns out the motor is responding to the duty cycle, but only between roughly 20% and 50%. The PWM is probably charging a capacitor and the motor sees a constant voltage between 0 and 6.1 V. Between 50% and 100% duty cycle, the voltage does go up, just not the rpms. Maybe for higher loads, this is different.
 
Last edited:

Thread Starter

JunkieNL

Joined Mar 19, 2020
54
I had a look at the Vsp output again. In my first attempt (without the PCB) I only saw noise, but now it is much cleaner. Please note this is a 0.5 V signal on top of a 15 V signal. One thing I had to do to increase Vpp and make this visible is to drive the optocoupler a bit harder.

IMG_20211231_001655.jpg

The probe is directly connected to the motor output and I am not using the PCB. I will try that as well and see if that changes anything. The frequency observed is exactly 12 times the frequency of the motor. With no load it will go up to 800 rpm.
 

MisterBill2

Joined Jan 23, 2018
18,179
OK, and now you have the feedback to control the speed precisely. Now to run it at somewhere near 310 volts. A full-wave doubler with reasonable capacitors, off the 12 volt mains, will give about 270 volts until you apply a load. It will be better if you use higher value capacitors, possibly 470 mfd at 450 volts. But at that point neither the positive or the negative side is near neutral. So to be safe, or isolated, you need a transformer. And evidently I was thinking the motor was quite a bit bigger than it really must be.
 
Top