This is raised repeatedly here.

This 'chinese' MIDI/USB adaptor is basically faulty. The design is potentially viable, but most versions are not made properly, i.e. components indicated on the little PCB are missing. If you follow the instructions on the web, the devices CAN be fixed, but it is not really worth the trouble, it is much better to buy a device from a reputable manufacturer (i.e. Roland) that will work correctly out of the box.

The device MIGHT work correctly in some circumstances, hence some users may report that the device is OK, but this is usually a delusion.

Geoff

Here are some pictures about the fixing: https://karusisemus.wordpress.com/2017/01/02/cheap-usb-midi-cable-how-to-modify-it/

This "product" was already a swindle when they sold it for $5. But now they raised it to $15!

I don't know much about the electrical side of things, so I can't tell how important it is to get the electrical signals to correctly follow the specifications or how acceptable it might be to skimp.

However, the buffer problems described below are more concerning to me and don't seem to be fixable. This buffer problem seems to be the likely cause of symptoms such as the following:
• When you play more than one note at exactly the same time, the MIDI interface begins to corrupt the data sent.
• When you send a System Exclusive message, the MIDI interface doesn't send all of the correct bytes.

Additionally, some keyboards send Clock messages 24 times a quarter note to tell the receiving device what tempo they are using. In this case, a Clock message will often end up close to played notes, and I suspect this might cause a poor quality USB MIDI interface with a small buffer to drop many notes.

The MIDI Implementation Chart at the back of the Yamaha PSR-E203 manual confirms it sends Clock messages. As a test, you can try disabling the Clock messages coming from your Yamaha PSR-E203 as follows: Press the FUNCTION button multiple times until you see the ExtClock setting, then press the + button to turn it ON. (See pages 48 and 49 in the manual about the function settings and page 52 describing the External Clock purpose.) This might improve the USB MIDI interface's ability to send notes somewhat, but I suspect you will still have issues if playing multiple notes at once.

Also, some keyboards send Active Sensing messages, an "I'm still here" filler message sent every 3/10 of a second to help a receiving device tell if a DIN-style MIDI cable got unplugged. I think this could also cause a poor quality USB MIDI interface with a small buffer to drop notes, but less frequently than the Clock messages would. The Yamaha PSR-E203 sends Active Sensing messages, but there's no way to disable them.

The following pages describe the problems with these poor quality USB MIDI interfaces:

Testing MIDI interfaces: This page shows
• A good quality MIDI interface that has a clean looking signal on an oscilloscope.
• One poor quality USB MIDI interface has a noisy looking signal on an oscilloscope, and also only transmits the first 3 bytes of a System Exclusive message.
• Another poor quality USB MIDI interface has the same noisy looking signal, but does transmit all the bytes of a System Exclusive message.

testing some USB MIDI adapters: In this message board thread, the same author as the page above, Frank Buss, describes some of the USB-level messages involved. At the USB level, there is an additional byte appended in front of each 3-byte MIDI message (or for each 3 bytes of a System Exclusive message). So at the USB level, these messages are 4 bytes long.

Cheap USB MIDI cable – how to modify it: This page is the same one that Pedro Lopez-Cabanillas linked to above. It describes modifications that I believe will improve the electrical signal, but I don't know much about the electrical side of things. However, the author briefly mentions you should test to make sure System Exclusive messages are correctly transmitted, because these poor quality USB MIDI interfaces have been reported to have "buffer problems" and says "[t]here is some information available that [a] Linux kernel is trying to take into account that deficiency" with a link to the following page.

found another USB midi cable with a 4 byte limit: This page contains a bug report thread for a specific Linux operating system titled "found another USB midi cable with a 4 byte limit". The number 4 probably refers to the USB-level bytes involved in a USB MIDI interface transmission.

So it sounds like these poor quality USB MIDI interfaces with buffer problems can only reliably transmit 4 USB-level bytes at once, which is the same as 3 MIDI bytes, which is the same as one note message.

Here are threads with example symptoms from other people:

Thread 1: No note ending messages when playing quickly. Non-standard note numbers which were possibly supposed to be the beginning of another MIDI message.

Thread 2: A pedalboard's notes are transmitted okay but a Yamaha keyboard's notes are often missing, unless playing very slowly.

Thread 3: Many notes missing.