Lately I’ve got a SFX-6 to play with. Everything is functioning well except that nothing gets transmitted through MIDI-OUT. I have verified it is not a settings issue, because there’s nothing coming through even when I attempt sysex file send. Then I decided to do some simple diagnostics with a meter. Found MIDI-OUT pin 4 shorted to GND… which should’ve been 5V->R220->PIN4 according to common midi pinout.
I opened the chasis and removed the attached capacitor but then I realized the difference – in the common pinout, PIN4 connects to pull down R220 then 5V but here PIN4 is straightly TX->inverter, and the other end PIN5 is grounded. Not sure why they chose to implement it like this…
So I conclude that, either the inverter (74HC4049) is failing – not likely, because MIDI-IN is working fine, or, the micro controller TX is malfunctioning…
Keep in mind that the 74HC4049 packs six independent level shifters into a single package. It’s entirely possible that the one that’s used for the OUT is burned out while the five others are still happily doing their thing.
Hey t thanks for your prompt reply! You’re right, can’t say the 74 chip is “totally functional” unless I probe it further. In the photo, the through hole annotated ‘2A’ is MCU TX and there’s actually a debug port on the left (not shown in the pic). Thinking about extending it with a jump wire and measure the voltage.
BTW, the circled capacitor is actually conductive (functioning as if it is a 0.02Ohm resistor… or is it really a voltage-sensitive resistor for protection?) If I later want to restore the connection, can I just short it? Sorry it’s a noob question
If I’m not mistaken, the 220Ω resistors should be there according to the MIDI specs. Regardless, I wouldn’t just take out parts that the manufacturer has put in; they probably have a reason for having it there.
I too think there should be a 220Ω resistor, but I measured 430Ω at the other PIN – when connected, they should form a closed loop. I’m shorting the connection in ‘lab condition’ and doing diagnostics with an Arduino. The debug port inside the machine is awesome… Will get back to you when I obtain data – we’ll be able to see if TX is functional.
O.k. got the results. TX is working, I get 3.3V on TX vs. GND, when I press a key, the meter displays slight disturbances – the MCU is pulling it down to transmit. However, at the inverter output, the voltage is always measured 0.15V vs. GND…
Adding 4049 to my next batch of chip purchase
Will report back when I have progress.
Can’t wait till 4049 arrive, I hacked in a 74HC04. But to my surprise it goes the same – MIDI IN working, MIDI OUT not. Measuring the output of 74HC04 I get 0.15V, but I can see disturbance in the reading when I press keys. Rolling up an Arduino monitor now.
Hooking up Arduino software serial monitor at 31250baud/s.
Disconnect external MIDI-OUT connection to PC.
Attach monitor to MIDI-OUT pin.
A lot of garbage gets through, but when I push the joystick, I can see patterns like increasing number, interleaved with garbage (0s and 1s).
Connect external MIDI-OUT, keep monitoring on the pin.
Nothing comes through. A constant 0.15V.
Will dismount the parts and insert a standard midi-out tomorrow.
hi @v-yadli, was the ‘unkown capacitor’ indeed a capacitor or a resistor instead? you would usually have a resistor for each pin on the output (and so for the thru port) as @t mentioned. i would probably also check that MIDI messages are being generated on the TX line (skipping the inverter)
Update: @guga well it’s not a capacitor after all, there’re four plates below when I remove it, so I think it is a filter for EMR purpose, not a big deal if I short it in a “safehouse”.
I’ve also tried to measure the signal on the TX port directly using my arduino analog input pin. Weird, I get different signal when I repeatedly press a key (during each key press, the arduino oscilloscope gets triggered and captures recent signal into buffer).
I will attach the signal later.