Total Frustration with TI

Oh joy. I feel for you.

Part of the reason that I do nearly all of my development with ARM Cortex parts using 3rd-party open-source tools is because the tools seem to have fewer really fatal flaws, and when they do there's someone out in the community who's flogging things to make the flaws go away.

It was a stroke of genius on ARM's part to make a common debug interface

-- it means that one tool chain will work on anybody's ARM Cortex parts.

I found another way to download to the target. Seems TI has a web page to support this. Really? A web page to control a target board on my PC? Ok...

I gave it a try and of course there are drivers to download. But this is handled automatically. Well, mostly. The first driver downloaded and installed ok. In the process it seemed to install the second driver. When I clicked the final button to refresh the page I am still prompted with a message that I need to download and install the second piece of software. Ok, I guess it didn't get loaded before, it must have been something else with a similar name. I click to download and install the second package. Refresh. Same message.

I tried this a couple more times before I realized, it was downloading, but not installing! So I manually ran the installation and refreshed the page. Now it seems responsive. I asked it to verify the firmware on the target. It takes some time because it has to install yet more stuff. Finally it starts running and gives an error... :-*

Seems this board is old enough the emulator firmware is not up to date. It wants to update that... do I have an emulator for the emulator? I guess the flash emulator emulates itself? :-/

Ok, some time and lots of blinking red LEDs later, the board seems to be updated and it completes my request to verify the "out of box" firmware. You got it! An error! It fails confirmation. I guess that firmware is also out of date. Do I dare update the firmware? In for a penny...

That worked, the image updated and now verifies and the board continues working the same as far as I can tell. I wonder what was changed in the OOB firmware?

I still need to get the FET USB driver working. Can anyone else download the ti_msp430driver_setup_1_00_01_00.zip file from

If I know it is hosed for everyone and not just me I will know better how to get help from TI. I've tried two of my four browsers. I really don't want to go on a browser hunt. I'm pretty sure it's TI.

Until then I'll work on getting things running with the rPi I suppose.

A long time ago, when the MSP430 first came out, I bought one of their little USB development boards, and found the process seamless.

But -- that was a long time ago.

I hope things get better. Sometimes you wonder if companies really want to sell their products.

Download worked fine for me Rick, it was on my machine within 30 seconds of clicking your link.

Yes, John Temples wrote me that it worked for him as well. Looks like I picked the wrong two browsers to try, Opera and Chrome. It worked for me when I tried FF. What are the odds?

What browser did you use?

I'm actually a bit frustrated with it all now. I'm going to quit for the day and work on financial stuff. I can't seem to get a CPA to return my call or worse they don't seem to respond to emails. I'm looking for a new one and they all seem to be living in the 1900's.

The irony here is that on one of my projects based on a Freescale Kinetis, the debugger I am using (with Linux) is a small board that came free with a TI demo kit!

It must be one of the ARM core parts from Luminary -- their debug hardware is compatible with OpenOCD, and works pretty well with other manufacturer's stuff (although the early boards didn't treat reset as well as the 3rd-party dongles do -- I don't know if this was fixed in later versions or not).

Firefox (on linux)

It was indeed a Luminary Micros part - well done!

As long as I am griping, I'm trying to read the schematics for the MSP-EXP430FR4133 and they are a royal PITA. First they use a light grey for the text on the pages. I have no idea who thought light grey is an acceptable color for anything on a drawing that you actually intend people to read. IDIOTS!

Secondly, the only text anywhere on the drawing that can be treated as text by selecting, copying or searching are the bloody pin numbers on the parts. There is no way to search for net names to see what goes where. IDIOTS!

Finally the text labeling the pins is incredibly microscopically small! It is less than half the height of the other text on the page. IDIOTS!

Anyone know what schematic package TI is using for their drawings? I want to make sure I never use it.

I will say they appear to have an interesting design for measuring the power of the target. It appears they are using a DCDC converter controlled by an MCU to measure the current by counting the on time of the pass transistor. I wonder what the benefit of this is over just using a series resistor and measuring the voltage delta across it. They are dropping 5 volts to 3.3, so the resistor could be between the regulator output and the sense point and not cost any extra voltage drop.

They have some patents on this method and even on a compiler feature that lets you measure the power of code functions and optimize for speed the sections that use the most power. The patent abstract isn't clear on how automated this is. One thing I noticed is that the patents all very specifically mention using an inductor. So if this is instead done with a switched capacitor bank, it would not violate any patents except maybe the compiler patent.

I think contract web designers designing for their portfolios, not for the product. Everything looks like a high-end restaurant menu.