Horizon T303 treadmill, MLH0910PD control board electrical repair

crwdns2853023:0crwdne2853023:0

1) Belt stops for no reason, motor control board debug LED produces 4 pulses.

2) Belt does not start, motor control board debug LED produces 4 pulses.

3) Motor control board produces clicking sounds.

At least in California technical support for Horizon products does not exist for any practical applications. They just respond with a copy-pasted text demanding a proof of purchase and a description of the problem over and over again but never offer any actual help. Also the replacement parts are astronomically expensive.

crwdns2853024:0crwdne2853024:0

My treadmill started acting up. Sometimes it just stopped spinning the belt while in use, sometimes it didn't want to even start the belt. And occasionally it was making clicking noise.

The motor itself has been tested with a pair of AA-size batteries connected in series. Motor terminals were unplugged from the control board and attached to the batteries. As soon as it was moving smoothly and with no visible irregularities, the motor was declared good. It's resistance varies from 0.9Ohm up to 2.5Ohm depending on the shaft position.

After that it was time for the motor control board itself. For reference it is MLH0910PD 110V version for US market. Reverse engineering is somewhat difficult because of 3 things:

1) The motor control board is located in a very inconvenient place, right next to the front panel post so that probing it with an oscilloscope is almost impossible.

2) The board is fully covered with glossy transparent material like clear coat. It makes it very difficult to identify electrical components and use either multimeter or an oscilloscope as it is hard to push through the layer of this material.

3) In some places the board is additionally covered with non-transparent hot glue-like material.

Due to aforementioned only partial reverse engineering has been done. Suggested repair disables one of two existing board protection mechanisms.

The board has 4 voltage rails: 160V, 15V, 12V, 5V. All 4 voltages turned out to be good, no spikes, no ripple.

Big voltage rectifier is powered through 2 relays connected in parallel 1A black relay and 30A white relay. Both relays are controlled by SOT23 transistors. Based on connection, those must be NPN bipolar type. Coils are connected to 12V rail. And when the relay is on, transistors are open and one sides of the coils are tied to ground. It turned out that when the board acted up, the relay coil control transistors closed and both sides of coils were at 12V.

The board has 2 hardware protection signals:

1) I-LIMIT which is the current through the motor measured as a voltage drop across large white resistors. It should stay low (0V) and turn to high (5V) when the motor is shorted.

2) MOS-D which monitors the motor N-MOSFET drain voltage. When it is low (motor disconnected), MOS-D signal should turn to logical high (5V).

The treadmill apparently has decided to requalify into the deep space communication dish.

It turned out that N-MOSFET drain voltage signal instead of being in low state was receiving Voyager-2 signal.

MOS-D signal goes to the microcontroller to let it turn error signal of 4 LED pulses and disable the relay. Also it goes to the hardware override circuit which is supposed to instantly disable the 30A relay. Noise on this line caused the relay to sporadically click. Extra wide noise pulses (when over 30ms) make the microcontroller shut the relay down and turn error code (4 LED pulses).

The fix is to 1) connect MOS-D signal to GND and 2) remove R20 resistor.

This combination makes the microcontroller see MOS-D always in low state and disables the hardware override so that the 30A relay does not click.

Of course, this repair doesn't fix the actual cause of the problem and only masks it by effectively disabling the MOS-D sub-circuit. It is due to the fact that I couldn't identify some of the ICs which actually create both I-LIMIT and MOS-D signals which in turn caused by the fact that the board is fully covered in clear coat and hot glue substances. My thinking is that with the MOS-D signal eliminated the only thing the board cannot see is when the motor is not connected to the board at all. But this situation shouldn't cause any trouble. If the motor rotor is shorted to the ground, I-LIMIT should still detect it and shut the relay down.

crwdns2853025:0crwdne2853025:0

The replacement board is over $350, the motor is over $700. It makes it not feasible to replace as units. Repair it if you can.

In order to do the fix you need at least

1) soldering iron,

2) solder,

3) flux,

4) piece of wire.

It is recommended to have

1) tweezers,

2) multimeter,

3) oscilloscope.