Who is CJMCU? Can they be reached?

I think you are expecting far too much, I'm guessing the selling is one of those "happybuygoodluck" that has pcbs,ic, ladies lingerie, dog toys diy, car accessories and what ever else was available to sell all mixed together

You have two data points. I don't think that means the CJMCU logo is a solid indicator.

The whole buffer thing only makes sense if you are soldering the parts onto the board with the CPU. So your thinking is that you are laying the groundwork for someone else who may be making their own boards? That makes a little more sense, but I still don't get the importance os the module being very small or very cheap, other than some people just wanting to make stuff cheaply. In the real world, if you are only making a small number for yourself, three or four bucks difference means little compared to your time lost trying to figure out which modules will work and which won't.

Now that I understand the importance of the handshake signal for reset, I get why a USB-UART cable isn't a good idea, they seldom have more than TX/RX and power/gnd. But the nice feature of the cable is you just plug it into your board when you need it or unplug it when you aren't using it. No buffer needed. A module can be plugged into a connector at the edge of the board rather than being included on the board. Then it would work like a cable.

Aren't launchpads pretty cheap? I see the MSP-EXP430G2 for only $10 plus shipping through TI or $12 including shipping through Aliexpress. They are usually significantly cheaper than OEMs, but this may be an example of how you can only save so much money when using name brand parts. So for $12 you can have a 100% compatible programmer.

The programmer that comes with my MSP430FR4133 launchpad even has a way to measure the MCU current through the on chip power regulator.

How are you reaching the other signal(s) you need that aren't on the through-holes? Early on I thought you had said this module didn't have all the I/O pins you needed, but it would seem I misunderstood.

I guess I get it now. You have designed this into your board and are documenting it so others can design it into their boards.

I certainly learned a lot about the CP2102.

I can't figure out how they can be selling this board under $2 using real CP2102 chips. Also if they were using genuine Silabs chips, why wouldn't they be using the CP2102N chip which is about half the price?

Yes, and if you were shipping a project thing to someone, you could just throw in a module that could be used that way. And I think many would choose to do that because it doesn't require a buffer, although it would require a header. But for those who want things to be as simple as possible for the end user, with no risk of losing the module, the embedded method could still be feasible at close to the same cost.

Either way, my software for flashing the G2231 could be used, which as far as I can determine has been unavailable up to now.

Yes, but if you're dealing with someone who has no other need for a Launchpad, then that's just unneeded expense if you can flash firmware without it.

The little module would not work for the G2553 because it needs the special pattern on DTR and RTS to enter BSL, and that module doesn't bring out RTS. But it would be perfect for the G2231, which has no built-in BSL code and therefore doesn't use the special pattern. You would just need the DTR line to reset the G2231.

Yes that's right. I don't know if anyone will be interested, and I haven't figured out where to post it, but maybe someone will get some benefit from it.

Me too. Perhaps more than I absolutely needed to know.

Well maybe they are fake chips. The N would be a direct drop-in replacement, so you would think they would use it.

I don't know either.

It's $10...

Ok, I barely understand. Different versions of the MSP430 use different protocols for programming. There are versions of this board that do provide all the hand shake lines, in fact *all* the lines from the CP2102. I don't have the same preference for saving a couple of dollars on the module.

Any product I design won't be user programmable. Or if it is, I would very likely include the TI emulator circuit to provide full functionality including the current measuring functions. That is very useful. TI has a patent on this idea, but I expect this it to preclude other MCU makers from using the circuit, not people selling MSP430 boards.

I found the cause of this problem, and the solution, and wanted to report back in case anyone sees this thread in the future.

I found two versions of this module, one outputting the correct 3.3V, and the other outputting the incorrect 4.25V. The modules looked identical, but on closer inspection I found that pin 9, the /Reset pin, was left NC on the good module, but was connected to USB V+ on the bad module.

So the 5V on V+ was flowing through the protection diode on /Reset directly to the VDD output. The regulator wants to make VDD 3.3V, but it's already at 4.25V, or one diode drop below 5V. No way to tell if this was done deliberately, but I suspect it was done that way to make the module more compatible with 5V Arduinos.

Anyway, I cut the trace to pin 9 on the bad module, and it became a good module at 3.3V.

It seems the /Reset line is supposed to be either left alone or pulled low. I guess it could be tied high to VDD, but not to V+.

Fortunately, the Ebay source I found for the good modules appears to have a lot more in stock. Banggood, on the other hand, appears to have only bad modules.

Il giorno domenica 2 aprile 2017 19:34:53 UTC+2, Peabody ha scritto:

Yes I did, and by the way that link doesn't work.

But better than finding cjmcu, I found the answer to why the module outputs 4.25V, and the fix for it. So I don't need them anymore.

Yes, it does, but it's not in English. You need to use Google Translate, or a browser that automatically translates non-English pages for you.

Are you planning to write this up on a web page somewhere?

I don't have a web page, but it was discussed on the Silabs forum, and I believe I reported the solution in this newsgroup. But there's an ongoing thread on the EEVblog forum where I've described it at length:

See replies 11, 21 and 24.

And here is the Silabs forum thread:

Forums are the worst place to document anything. Often the discussion is dragged out over a range of messages which makes it very hard to get the details. A simple web page would be a great place to pull all the info together. Heck, this is worth an article on one of the web sites like instructables.com

Ok, so to summarise, disconnect pin 9 from USB+, just leave it floating.