Can RPi by taught in Microprocessor Interfacing Course?

[muhsul]

I have been teaching microprocessor interfacing course to undergraduate students of computer engineering for few years. Generally we teach them about internal architecture of a microcontroller, its busses and addressing modes, followed by timers, interrupts, serial communication, ADC, PWM, motor control etc. I have been teaching my students about 8051 and PIC microcontrollers previously but now I want to exclude one of these controllers and teach RPi in its place. But I have a few questions that I need answers to. Any help will be highly appreciated.

1. Does teaching RPi to students fall under the scope of my Microprocessor Interfacing course? I know it's like a mini computer. Most books explain setting up the Pi and designing projects with it but to the best of my knowledge I didn't come across a single book that also explains the internal architecture of RPi or its processor. Please guide me if you know about some good resource?
2. I am completely new to RPi and want to master RPi before I teach my students. Please suggest a good book for personal learning and a book that can also be used as a text book. I am comfortable with Python.

Thanks in advance.

[k-pi]

This may be better addressed to the Foundation, but I don't think they divulge the inner workings of these boards.

[hippy]

1. Does teaching RPi to students fall under the scope of my Microprocessor Interfacing course?

I don't know. It perhaps depends how deep you go, what you consider essential and not so much.

There are ARM Architecture Reference Manuals out there, the instruction set and functionality is mostly well documented apart from Jazelle or whatever it's called which was designed to handle Java efficiently and there may be other extensions no one outside of NDA World even knows about.

The Pi Soc is ARM, plus VideoCore, plus peripherals, plus glue. I am sure I saw details of the general architecture of how it all fits together somewhere. It was enough to keep me happy but I didn't want to go deep into it. But, because of the nature of SoC's being a combination of things, I would doubt there is a single document which describes it all. And if you go beyond the SoC there's a lot more to a Pi than just the SoC.

The BCM 2835 Peripherals are reasonably documented though not to the detail one would normally expect from Microchip and similar. There's not a full electrical spec for example though most things have been addressed to the level most would need them addressing

But both ARM and Broadcom are a secretive bunch who jealously guard The Precious, and the Foundation seems to be increasingly embracing that culture; not even producing full circuit diagrams of their boards, so not everything one might like is necessarily publicly available.

It therefore depends on what you want or need. I don't see why you couldn't teach "Interfacing" using a Pi but it seems you may mean something more by that than I would.

I don't think anyone could tell you if a Pi is suitable unless they've done your course. I suspect you will have to start digging for what there is, assess that and see if it meets your needs.

[ejolson]

I have been teaching microprocessor interfacing course to undergraduate students of computer engineering for few years.

I would look at what Vince Weaver has been doing with his courses at the University of Maine.

While possible, I think it will be a big project to develop suitable materials for your course. If you don't have tenure, it might be better to publish two more papers and apply for a grant instead. If you do have tenure, it would be useful to get a teaching release in order to update the old course to use the Raspberry Pi.

[muhsul]

Thank you all for your response.

[jamesh]

The BCM 2835 Peripherals are reasonably documented though not to the detail one would normally expect from Microchip and similar. There's not a full electrical spec for example though most things have been addressed to the level most would need them addressing

But both ARM and Broadcom are a secretive bunch who jealously guard The Precious, and the Foundation seems to be increasingly embracing that culture; not even producing full circuit diagrams of their boards, so not everything one might like is necessarily publicly available.

Actually, our attitude hasn't really changed much since the original Pi was released. We've never really had full disclosure. Would have been nice to be a bit more open, but unfortunately, commercial pressures and the desire to stay solvent (and not be sued) come first.

[hippy]

Actually, our attitude hasn't really changed much since the original Pi was released. We've never really had full disclosure.

That's true but there does seem to have been a notable change from desiring to be as open as possible to being less open.

Reduced circuit diagrams and disappearing component identifiers on boards are visible signs of that but probably cause few problems for users. As will not being able to identify what components are being used on a board or having datasheets for those components. That should only be problematic to those using the boards outside the official operating envelope.

Would have been nice to be a bit more open, but unfortunately, commercial pressures and the desire to stay solvent (and not be sued) come first.

Yes; that's entirely understandable. One doesn't get to stash £23 million in the bank without protecting revenue streams. Commercial reality often pushes other ambitions out the way. That's the way of the world and one cannot really expect the Pi to change that.

[PeterO]

I have been teaching microprocessor interfacing course to undergraduate students of computer engineering for few years. Generally we teach them about internal architecture of a microcontroller, its busses and addressing modes, followed by timers, interrupts, serial communication, ADC, PWM, motor control etc. I have been teaching my students about 8051 and PIC microcontrollers previously but now I want to exclude one of these controllers and teach RPi in its place. But I have a few questions that I need answers to. Any help will be highly appreciated.

1. Does teaching RPi to students fall under the scope of my Microprocessor Interfacing course? I know it's like a mini computer. Most books explain setting up the Pi and designing projects with it but to the best of my knowledge I didn't come across a single book that also explains the internal architecture of RPi or its processor. Please guide me if you know about some good resource?
2. I am completely new to RPi and want to master RPi before I teach my students. Please suggest a good book for personal learning and a book that can also be used as a text book. I am comfortable with Python.

Thanks in advance.

I think the key question is "Do you intend to run an operating system on the PI ?" , or would you be using it in a "bare metal" environment ?

If you have an OS running, then most of the hardware devices will be under the control of the drivers in the kernel.

As to "the internal architecture of RPi or its processor", well the CPUs are ARM cores, and there is plenty of materials around about programming them.

From the sound of it I doubt you would be that interested in the GPU (which is the lest documented part of the system).

PeterO

[jamesh]

Actually, our attitude hasn't really changed much since the original Pi was released. We've never really had full disclosure.

That's true but there does seem to have been a notable change from desiring to be as open as possible to being less open.

Reduced circuit diagrams and disappearing component identifiers on boards are visible signs of that but probably cause few problems for users. As will not being able to identify what components are being used on a board or having datasheets for those components. That should only be problematic to those using the boards outside the official operating envelope.

Would have been nice to be a bit more open, but unfortunately, commercial pressures and the desire to stay solvent (and not be sued) come first.

Yes; that's entirely understandable. One doesn't get to stash £23 million in the bank without protecting revenue streams. Commercial reality often pushes other ambitions out the way. That's the way of the world and one cannot really expect the Pi to change that.

We cannot release datasheets of other people chips (the suing thing), and pragmatism has overtaken desire - we've come to realise that the effort involved in making everything open is not really reflected in sales, so its not cost effective, and after all, trading are there to make profit for the Foundation.

Do we really have £23M in the bank? That shoud pay for a few of the projects we have on the go...(exciting couple of years coming up if it all pans out!)

[hippy]

Do we really have £23M in the bank? That shoud pay for a few of the projects we have on the go...(exciting couple of years coming up if it all pans out!)

That's what the RPF reported as EOY Funds for 2017 which I am assuming is in the bank rather than kept in a shoe box at Pi Towers. It seems £11 million is in an investment portfolio so maybe not actually in the bank.

[jamesh]

Do we really have £23M in the bank? That shoud pay for a few of the projects we have on the go...(exciting couple of years coming up if it all pans out!)

That's what the RPF reported as EOY Funds for 2017 which I am assuming is in the bank rather than kept in a shoe box at Pi Towers. It seems £11 million is in an investment portfolio so maybe not actually in the bank.

Eben is bound to figure out a way of spending all that. He's good at it, which is why we have a VERY expensive oscilloscope.

[bensimmo]

Was that an end of March buy?