How hard is it to implement a 1984 "simple" hardware communication protocol?
-
Flow control lines (RTS/CTS, DTR/DSR)? It's worth pinning them permanently high just to be on the safe side.
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony "Common sense is so rare these days, it should be classified as a super power" - Random T-shirt AntiTwitter: @DalekDave is now a follower!
This isn't RS232, it's just a simple UART. There is only TX and RX.
To err is human. Fortune favors the monsters.
-
Please don't take it as mansplaining :)
honey the codewitch wrote:
A) It's what the spec says
If you are referring to MIDI baud rate, we both agree on that.
honey the codewitch wrote:
B) I have reference code - multiple projects by different authors actually, that use this baud rate
Are they on ESP32? A quick search on available baud rate for ESP32 didn't give any definitive results. Not all baud rates can be obtained from an oscillator. See this list of oscillator frequencies and search for "UART" in it. On the transmitter side, the frequency is divided by an integer number to generate the baud rate. On the receiving side, the oscillator frequency is divided to a multiple of the baud rate (x8 or x16) and, after the initial high to low transition of the start bit, successive bits are sampled in the middle. If divided frequency doesn't match the baud rate, the last bits will be sampled improperly. (yes, this part smells of mansplaining :) ) The MIDI baud rate is a divisor of 1MHz while most communication baud rates are multiples of the original 75 bauds teletype rate.
honey the codewitch wrote:
C) I'd be getting garbage instead of nothing
Are you referring to my suggestion of sending a repetitive character? This is just a test to see you have the correct baud rate. You don't need to have anything connected to the transmitter.
Mircea
The baud rate on the ESP32 end is configurable and must match the baud rate of MIDI
Mircea Neacsu wrote:
The MIDI baud rate is a divisor of 1MHz while most communication baud rates are multiples of the original 75 bauds teletype rate.
That's interesting, and if the reference code did not exist (some of it is for the ESP32, some for Arduino, I've tried both) I'd be really concerned by that. As it is, there are known goods that are doing what I'm doing. I suspect a bad module, so I've ordered a different, simpler module and hopefully that will work.
To err is human. Fortune favors the monsters.
-
MIDI is a two wire serial protocol. It's a UART with a 5 pin DIN on the end. Literally I should be able to simply type Serial1.begin(31250); // MIDI baud rate And start talking to it. I've been through A USB host controller an ESP32 S2 an ESP32 S3 a MIDI breakout none of the hardware is giving me *anything* from MIDI, neither USB nor standard. I'm getting *another* MIDI breakout today. I hope the old one was defective. If that doesn't work I'm hardwiring a raw 5 pin DIN to an optocoupler and doing the whole elephanting thing from scratch. Nothing like waiting on hardware in order to code something you're pretty sure won't work.
To err is human. Fortune favors the monsters.
Hey HTCW that seems like a rather basic and general question. Does the "processor powerhouse" have trouble recalling one it`s fundamental lessons or that`s actually an exotic question but I fail to see the subtility. what`s so special about 1984 why not 1985, 2000 or some other year?
-
Hey HTCW that seems like a rather basic and general question. Does the "processor powerhouse" have trouble recalling one it`s fundamental lessons or that`s actually an exotic question but I fail to see the subtility. what`s so special about 1984 why not 1985, 2000 or some other year?
Well, I had the year wrong. Should have been 1983 - forgive my shoddy memory. 1983 was when they standardized the MIDI protocol. The reason it's relevant is simplicity. These were the days before stuff like Very Large Scale Integration in chip manufacturing, and solid state digital circuits were still pretty primitive so protocols had to be simple and of limited throughput.
To err is human. Fortune favors the monsters.
-
MIDI is a two wire serial protocol. It's a UART with a 5 pin DIN on the end. Literally I should be able to simply type Serial1.begin(31250); // MIDI baud rate And start talking to it. I've been through A USB host controller an ESP32 S2 an ESP32 S3 a MIDI breakout none of the hardware is giving me *anything* from MIDI, neither USB nor standard. I'm getting *another* MIDI breakout today. I hope the old one was defective. If that doesn't work I'm hardwiring a raw 5 pin DIN to an optocoupler and doing the whole elephanting thing from scratch. Nothing like waiting on hardware in order to code something you're pretty sure won't work.
To err is human. Fortune favors the monsters.
-
What happens if you connect the midi out directly with the midi in (loopback or however one call this)?
I don't send MIDI out with this device yet, so I can't check that. I did however, connect the keyboard to the computer successfully using MIDI DINs to USB so I know the keyboard is good. I think I have narrowed it down in this case to a hardware problem. I think my previous problems were software, but as it usually goes with the most vexing troubleshooting, I was probably dealing with issues on multiple fronts. I have a replacement piece of hardware coming today and we'll see how that goes. It's even simpler than the module I currently have, and should eliminate some variables from the equation if nothing else. It's also smaller and not one of those insipid Arduino shields I hate so much.
To err is human. Fortune favors the monsters.
-
I don't send MIDI out with this device yet, so I can't check that. I did however, connect the keyboard to the computer successfully using MIDI DINs to USB so I know the keyboard is good. I think I have narrowed it down in this case to a hardware problem. I think my previous problems were software, but as it usually goes with the most vexing troubleshooting, I was probably dealing with issues on multiple fronts. I have a replacement piece of hardware coming today and we'll see how that goes. It's even simpler than the module I currently have, and should eliminate some variables from the equation if nothing else. It's also smaller and not one of those insipid Arduino shields I hate so much.
To err is human. Fortune favors the monsters.
-
Quote:
connect the keyboard to the computer successfully using MIDI DINs to USB
Ok, but also that sounds strange for me. What I thought is, that MIDI is 5mA current loop...
MIDI to USB cable[^] There's no loop. It doesn't work like token ring if that's what you're saying. It's basically a simple UART with a 5 pin DIN. The *only* weird thing about it, is MIDI INs are always optocoupled (or otherwise physically decoupled) from the rest of the circuit to prevent ground loops (i think) - otherwise it's a simple 2 wire UART interface (not RS232, as there are no control lines, just TX and RX) - MIDI only uses like 2 or 3 of the five pins . i say maybe 2 because MIDI ins and outs require two separate lines. It is not full duplex at least not over one cable. I hope all that made sense, and answered whatever you were wondering about. I wasn't entirely clear on what you're talking about to be honest. :)
To err is human. Fortune favors the monsters.
-
MIDI to USB cable[^] There's no loop. It doesn't work like token ring if that's what you're saying. It's basically a simple UART with a 5 pin DIN. The *only* weird thing about it, is MIDI INs are always optocoupled (or otherwise physically decoupled) from the rest of the circuit to prevent ground loops (i think) - otherwise it's a simple 2 wire UART interface (not RS232, as there are no control lines, just TX and RX) - MIDI only uses like 2 or 3 of the five pins . i say maybe 2 because MIDI ins and outs require two separate lines. It is not full duplex at least not over one cable. I hope all that made sense, and answered whatever you were wondering about. I wasn't entirely clear on what you're talking about to be honest. :)
To err is human. Fortune favors the monsters.
-
Well, I had the year wrong. Should have been 1983 - forgive my shoddy memory. 1983 was when they standardized the MIDI protocol. The reason it's relevant is simplicity. These were the days before stuff like Very Large Scale Integration in chip manufacturing, and solid state digital circuits were still pretty primitive so protocols had to be simple and of limited throughput.
To err is human. Fortune favors the monsters.
ok that`s legit. I have no further objections, I will only note that "hardware communication" and computers aren`t necessarely one and the same thing. MIDI does relate to computers but the case is at the edge of the mainstream IT (it`s not defining for IT "the beginnings") it rather belongs to the branch of music hardware. I have the impression that MIDI playback on PC as a thing belongs to the late 90`s (Age of Empires had MIDI playback I think).
-
ok that`s legit. I have no further objections, I will only note that "hardware communication" and computers aren`t necessarely one and the same thing. MIDI does relate to computers but the case is at the edge of the mainstream IT (it`s not defining for IT "the beginnings") it rather belongs to the branch of music hardware. I have the impression that MIDI playback on PC as a thing belongs to the late 90`s (Age of Empires had MIDI playback I think).
I never mentioned computers. In fact, computers just complicate this because MIDI over USB can't do MIDI pass through devices
To err is human. Fortune favors the monsters.
-
MIDI is a two wire serial protocol. It's a UART with a 5 pin DIN on the end. Literally I should be able to simply type Serial1.begin(31250); // MIDI baud rate And start talking to it. I've been through A USB host controller an ESP32 S2 an ESP32 S3 a MIDI breakout none of the hardware is giving me *anything* from MIDI, neither USB nor standard. I'm getting *another* MIDI breakout today. I hope the old one was defective. If that doesn't work I'm hardwiring a raw 5 pin DIN to an optocoupler and doing the whole elephanting thing from scratch. Nothing like waiting on hardware in order to code something you're pretty sure won't work.
To err is human. Fortune favors the monsters.
-
It sounds super frustrating, but also irritatingly fascinating. I wonder if there are any books on the subject from the era. Will you write an article about all of this when you get it all figured out? I would love to read that.
I fully intend to. I'm a little ways away from it now though. I have big plans. :)
To err is human. Fortune favors the monsters.
-
ok that`s legit. I have no further objections, I will only note that "hardware communication" and computers aren`t necessarely one and the same thing. MIDI does relate to computers but the case is at the edge of the mainstream IT (it`s not defining for IT "the beginnings") it rather belongs to the branch of music hardware. I have the impression that MIDI playback on PC as a thing belongs to the late 90`s (Age of Empires had MIDI playback I think).
I'm replying here because your new comments are in moderation. This site is about programming, whether it's computers, or MCUs. In this case it's an MCU which acts as a MIDI controller. It is not a PC. It is little different than the circuitry you find in today's MIDI controllers. I am coding the firmware as well as building the hardware - they're both part of the very same project, so it is within the scope of CP. I have a substantial reputation for coding IoT here, and most folks in the lounge know me, so I didn't think it necessary to elaborate on all the details of what I was doing. This is simply part of the project. Edit: I should add, the lounge is for general commentary on the life of the people that post here. Lots of people post things that aren't code related. In fact that's probably 80% of the stuff here, if you count Wordle and such.
To err is human. Fortune favors the monsters.
-
I never mentioned computers. In fact, computers just complicate this because MIDI over USB can't do MIDI pass through devices
To err is human. Fortune favors the monsters.
"hard" "music" this is site where posts usually have some bearing to computers/IT industry. So it is the "this is about computers" presumption that governs. When it`s not about computers express mention should be made that isn`t not about computers. When you use the word "hard" and start talking about music you move things in a totally different registry. That`s not something that can be made tacitly. The word "hardware" has one very clear meaning on this site, different from the one it might have on a music industry site. Is CP now suddenly a "rock and roll" website? that`s what it begins to look like if you neglect what I said above.
-
MIDI is a two wire serial protocol. It's a UART with a 5 pin DIN on the end. Literally I should be able to simply type Serial1.begin(31250); // MIDI baud rate And start talking to it. I've been through A USB host controller an ESP32 S2 an ESP32 S3 a MIDI breakout none of the hardware is giving me *anything* from MIDI, neither USB nor standard. I'm getting *another* MIDI breakout today. I hope the old one was defective. If that doesn't work I'm hardwiring a raw 5 pin DIN to an optocoupler and doing the whole elephanting thing from scratch. Nothing like waiting on hardware in order to code something you're pretty sure won't work.
To err is human. Fortune favors the monsters.
Doesn't MIDI come in 5v and 3.3v? You might need a level shifter. E.g. ESP32 is 3.3v and won't appreciate 5v input.
-
Doesn't MIDI come in 5v and 3.3v? You might need a level shifter. E.g. ESP32 is 3.3v and won't appreciate 5v input.
My module has a level shifter and can operate at either. I found the problem eventually. My TX and RX lines were crossed because I was confused about the orientation: I thought TX was labeled like "I go to the TX pin" rather than "I am the TX pin" - don't know where my head was at.
To err is human. Fortune favors the monsters.
-
MIDI to USB cable[^] There's no loop. It doesn't work like token ring if that's what you're saying. It's basically a simple UART with a 5 pin DIN. The *only* weird thing about it, is MIDI INs are always optocoupled (or otherwise physically decoupled) from the rest of the circuit to prevent ground loops (i think) - otherwise it's a simple 2 wire UART interface (not RS232, as there are no control lines, just TX and RX) - MIDI only uses like 2 or 3 of the five pins . i say maybe 2 because MIDI ins and outs require two separate lines. It is not full duplex at least not over one cable. I hope all that made sense, and answered whatever you were wondering about. I wasn't entirely clear on what you're talking about to be honest. :)
To err is human. Fortune favors the monsters.
The "current loop" is an old standard (originally used for teletypes I believe), it is still used in many industrial apps, as it's good for noise rejection. Instead of using voltage (such as +/-12V) to communicate, it uses a 20mA on, 4mA off signal. The receiver has a resistor that the current flows through, so it can monitor the current by measuring the voltage. The 4mA off state is used so that no current indicates a broken wire. The standard is also used for sending analog values from sensors. Sorry for the rambling mansplainage.
-
The "current loop" is an old standard (originally used for teletypes I believe), it is still used in many industrial apps, as it's good for noise rejection. Instead of using voltage (such as +/-12V) to communicate, it uses a 20mA on, 4mA off signal. The receiver has a resistor that the current flows through, so it can monitor the current by measuring the voltage. The 4mA off state is used so that no current indicates a broken wire. The standard is also used for sending analog values from sensors. Sorry for the rambling mansplainage.
Ah, thanks for the explanation. MIDI is just simple UART. In fact, care is taken to avoid ground loops that can transfer noise or hum from one piece of equipment to the next, which means by the standard, MIDI In ports are always optocoupled (or equivalent). They aren't in a ring, either. The cables are half-duplex UART meaning one cable carries either RX or TX depending if it's an In or Out port it's plugged into. Weirdly, they use 5 pin DIN to connect. My guess is the physically large connector is good for stage equipment for a number of reasons.
To err is human. Fortune favors the monsters.
-
MIDI is a two wire serial protocol. It's a UART with a 5 pin DIN on the end. Literally I should be able to simply type Serial1.begin(31250); // MIDI baud rate And start talking to it. I've been through A USB host controller an ESP32 S2 an ESP32 S3 a MIDI breakout none of the hardware is giving me *anything* from MIDI, neither USB nor standard. I'm getting *another* MIDI breakout today. I hope the old one was defective. If that doesn't work I'm hardwiring a raw 5 pin DIN to an optocoupler and doing the whole elephanting thing from scratch. Nothing like waiting on hardware in order to code something you're pretty sure won't work.
To err is human. Fortune favors the monsters.
I did build my own MIDI card for the IBM PC ISA bus in the 80's. The main difference between MIDI and RS232 serial comms is the baud rate needed. I used a 2MHz crystal for the MIDI. RS232 uses 1.8something MHz crystal. The reason is that the baud rate needed for MIDI is not possible otherwise.
