Who is CJMCU? Can they be reached?

[Peabody]

As described in other threads, I'm working on embedding a USB-to-serial
adapter in one of my MSP430 projects, and I've found the perfect adapter for
this purpose. It's smaller than the other adapters, has the right pinout, and
even comes with a micro-USB socket onboard.

The chip used in the adapter is the CP2102 from Silicon Labs. But there's a
problem. The output labeled in silkscreen as "3.3V" actually measures
4.25V. And it's not just the one I received:

https://www.youtube.com/watch?v=8qcGdB2yVd8

It seems that all modules of this design are like that. And the CP2102 is
definitely a 3.3V part. So it's either something wrong with the board
layout, or something the manufacturer has done intentionally to get that
output up close to 5V.

This is a deal killer if it can't be corrected because all the MSP430 parts
would likely be destroyed by voltage that high. Oh, and it's not just the
3.3V output, it's DTR and TXD too.

I've been on the Silabs forum, and so far nobody has come up with an
explanation for how this could be done. If everything is connected as it
should be, which appears to be the case, this shouldn't be possible.

So what I'd like to do is contact the manufacturer and try to get some
information about how this was done in the hope that it will be something I
can reverse easily. The markings on the board are:

CJMCU

BQ09615H2 2925a

I've seen CJMCU elsewhere in relation to drones and other stuff, but Google
gives me no home page or any other info about them.

Does anyone here know anything about CJMCU? Where would you go to find them?

[Jim Thompson]

On Sun, 02 Apr 2017 12:34:22 -0500, Peabody
<waybackNO...@yahoo.com> wrote:

>As described in other threads, I'm working on embedding a USB-to-serial
>adapter in one of my MSP430 projects, and I've found the perfect adapter for
>this purpose. It's smaller than the other adapters, has the right pinout, and
>even comes with a micro-USB socket onboard.
>
>The chip used in the adapter is the CP2102 from Silicon Labs. But there's a
>problem. The output labeled in silkscreen as "3.3V" actually measures
>4.25V. And it's not just the one I received:
>

[snip]

Is that a loaded measurement? Or unloaded? It's probably an LDO and
"leaks" up to rail without any load.

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| STV, Queen Creek, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

Thinking outside the box... producing elegant solutions.

[TTman]

On 02/04/2017 18:54, Jim Thompson wrote:
> On Sun, 02 Apr 2017 12:34:22 -0500, Peabody
> <waybackNO...@yahoo.com> wrote:
>
>> As described in other threads, I'm working on embedding a USB-to-serial
>> adapter in one of my MSP430 projects, and I've found the perfect adapter for
>> this purpose. It's smaller than the other adapters, has the right pinout, and
>> even comes with a micro-USB socket onboard.
>>
>> The chip used in the adapter is the CP2102 from Silicon Labs. But there's a
>> problem. The output labeled in silkscreen as "3.3V" actually measures
>> 4.25V. And it's not just the one I received:
>>
> [snip]
>
> Is that a loaded measurement? Or unloaded? It's probably an LDO and
> "leaks" up to rail without any load.
>
> ...Jim Thompson
>

I'd agree with that... Seen it before.

---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

[Tim Wescott]

On Sun, 02 Apr 2017 10:54:56 -0700, Jim Thompson wrote:

> On Sun, 02 Apr 2017 12:34:22 -0500, Peabody
> <waybackNO...@yahoo.com> wrote:
>
>>As described in other threads, I'm working on embedding a USB-to-serial
>>adapter in one of my MSP430 projects, and I've found the perfect adapter
>>for this purpose. It's smaller than the other adapters, has the right
>>pinout, and even comes with a micro-USB socket onboard.
>>
>>The chip used in the adapter is the CP2102 from Silicon Labs. But
>>there's a problem. The output labeled in silkscreen as "3.3V" actually
>>measures 4.25V. And it's not just the one I received:
>>
> [snip]
>
> Is that a loaded measurement? Or unloaded? It's probably an LDO and
> "leaks" up to rail without any load.
>
> ...Jim Thompson

If so it's easily tested by putting a load on it.

--
Tim Wescott
Control systems, embedded software and circuit design
I'm looking for work! See my website if you're interested
http://www.wescottdesign.com

[Peabody]

Tim Wescott says...

>> Is that a loaded measurement? Or unloaded? It's
>> probably an LDO and "leaks" up to rail without any
>> load.

>> ...Jim Thompson

> If so it's easily tested by putting a load on it.

It drops very slightly at 20 ma. So essentially the
same loaded or unloaded. It's basically a diode drop below
the 5V USB level. So not really regulated at all.

And I do have a different adapter using the exact same chip,
even the same date code, and it's 3.3V loaded or unloaded,
and works fine.

I think they did this to make it compatible with 5V
Arduinos. I just wish I knew how they did it. I've been
through both adapters, and the connections and readings are
the same on both. So something is going on inside the chip.

[Jim Thompson]

Have you checked the board for different resistor values? Maybe
setting the "regulator" voltage?

[Peabody]

Jim Thompson says...

> Have you checked the board for different resistor
> values? Maybe setting the "regulator" voltage?

The only resistors on the board were the LED resistors, and
I removed both the LEDs and the resistors just to make sure
they weren't the problem. There are no resistors left on
the board, only capacitors and the CP2102 chip.

[Tim Wescott]

On Sun, 02 Apr 2017 15:29:55 -0700, Jim Thompson wrote:

> On Sun, 02 Apr 2017 16:54:05 -0500, Peabody
> <waybackNO...@yahoo.com> wrote:
>
>>Tim Wescott says...
>>
>> >> Is that a loaded measurement? Or unloaded? It's probably an LDO and
>> >> "leaks" up to rail without any load.
>>
>> >> ...Jim Thompson
>>
>> > If so it's easily tested by putting a load on it.
>>
>>It drops very slightly at 20 ma. So essentially the same loaded or
>>unloaded. It's basically a diode drop below the 5V USB level. So not
>>really regulated at all.
>>
>>And I do have a different adapter using the exact same chip,
>>even the same date code, and it's 3.3V loaded or unloaded, and works
>>fine.
>>
>>I think they did this to make it compatible with 5V Arduinos. I just
>>wish I knew how they did it. I've been through both adapters, and the
>>connections and readings are the same on both. So something is going on
>>inside the chip.
>>
>>
> Have you checked the board for different resistor values? Maybe setting
> the "regulator" voltage?

The 3.3V is coming from the USB-to-serial chip? In these degenerate
times, you probably set the voltage via USB. Maybe there's info on how
to do it on the web?

If there's a 3-terminal regulator setting the voltage, then it could be a
4.3V part (which sounds weird), or they screwed up the footprint and
you're just getting your current from the protection diodes.

[Lasse Langwadt Christensen]

https://wiki.kewl.org/dokuwiki/interfaces:cp2102

which could also explain you problems with pins when not powered even though
the datasheet indicates it shouldn't be a problem

[Jasen Betts]

On 2017-04-02, Peabody <waybackNO...@yahoo.com> wrote:
> As described in other threads, I'm working on embedding a USB-to-serial
> adapter in one of my MSP430 projects, and I've found the perfect adapter for
> this purpose. It's smaller than the other adapters, has the right pinout, and
> even comes with a micro-USB socket onboard.
>
> The chip used in the adapter is the CP2102 from Silicon Labs. But there's a
> problem. The output labeled in silkscreen as "3.3V" actually measures
> 4.25V. And it's not just the one I received:
>
> https://www.youtube.com/watch?v=8qcGdB2yVd8
>
> It seems that all modules of this design are like that. And the CP2102 is
> definitely a 3.3V part. So it's either something wrong with the board
> layout, or something the manufacturer has done intentionally to get that
> output up close to 5V.
>
> This is a deal killer if it can't be corrected because all the MSP430 parts
> would likely be destroyed by voltage that high. Oh, and it's not just the
> 3.3V output, it's DTR and TXD too.
>
> I've been on the Silabs forum, and so far nobody has come up with an
> explanation for how this could be done. If everything is connected as it
> should be, which appears to be the case, this shouldn't be possible.

could be one of the chips is counterfeit?

--
This email has not been checked by half-arsed antivirus software

[Mike Perkins]

On 02/04/2017 18:34, Peabody wrote:
> As described in other threads, I'm working on embedding a USB-to-serial
> adapter in one of my MSP430 projects, and I've found the perfect adapter for
> this purpose. It's smaller than the other adapters, has the right pinout, and
> even comes with a micro-USB socket onboard.
>
> The chip used in the adapter is the CP2102 from Silicon Labs. But there's a
> problem. The output labeled in silkscreen as "3.3V" actually measures
> 4.25V. And it's not just the one I received:

I know someone who spent a day trying to work out a similar error.

In their case the battery indicator was on, and of course with a new
battery all the voltages mysteriously returned to normal.

Either way, check the accuracy of your measurement.

--
Mike Perkins
Video Solutions Ltd
www.videosolutions.ltd.uk

[Peabody]

Lasse Langwadt Christensen says...

> https://wiki.kewl.org/dokuwiki/interfaces:cp2102

> which could also explain you problems with pins when not
> powered even though the datasheet indicates it shouldn't
> be a problem

Thanks very much for finding that. Of course that's one
possible explanation. But the CP2102 has a number of
settings that can be programmed into the chip. And I
believe one of those allows you to turn off the regulator.

These chips are so cheap to begin with that I didn't think
anyone would bother to make fakes ones. Well, in the end it
may not matter why it doesn't work if I can't find a way to
make it work.

Since posting, I *may* have found software that will let me
reset the chip to factory defaults. If it still misbehaves
after that, then a fake chip would be the explanation.

However, the problem with the I/O pins not being hi-Z when
the device is off is also there in a diferent module that
does output 3.3V, and which I have no reason to believe
isn't a genuine chip. So that's a separate issue.

[Tim Wescott]

Huh. Just looked at the part, and the 3.3V regulator is fixed. Now I'm
thinking it's either a wiring problem or maybe you got some counterfeit
parts?

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

I'm looking for work -- see my website!

[Peabody]

Tim Wescott says...

> Huh. Just looked at the part, and the 3.3V regulator is
> fixed. Now I'm thinking it's either a wiring problem or
> maybe you got some counterfeit parts?

Yes it's fixed, but I think it may be possible to disable it
in some way. But it's also possible it's a bogus chip. It's
not a routing problem. I've confirmed everything goes where
it's supposed to, and doesn't go where it shouldn't.

Thinking about this, my circuit limits what's passed through
to the processor to the voltage at Vcc of the 50A buffer.
And it will tolerate higher-than-Vcc inputs. So even at
4.25V on all outputs, if I run the 4.25V supply through an
IR LED to the 50A Vcc, that would let me use this adapter,
but would limit voltage to about 3V or a little higher. The
gate of the blocking MOSFET would still have to go back to
the 4.25V, which would put it 1.25V above source, but that
should be ok.

Have you ever used an LED to drop voltage? Of course the
drop would vary a bit with current, but I'm looking at under
10 mA, so it should be pretty stable. And the typical
voltage drop on an IR LED is about 1.2V, max 1.5V at 20 mA.
IR204 for example.

What do you think? Of course there are other adapters, but
this one is tiny.

[rickman]

I think you have a crap chip on your board and you should drop it rather
than try to mess with it.

To see what is wrong, you can try doing one thing I thought of. I found
a web page where some guys building a board with one of these had
trouble because they shorted the USB VBUS to their 3.3 volt line. This
would give different symptoms depending on what your USB power source
is. First, measure the 5 volt rail. If it is also 4.3 volts try
disconnecting it and measure the resistance between the 3.3 and 5 volt
lines. If this is not your problem I expect your board has a cloned
chip and will never work right.

https://www.rcgroups.com/forums/showthread.php?1595761-AfroFlight32-now-we-re-talking-progress/page34

One other point. When I look at the Silabs site I find the CP2102 and
the CP2102N. The CP2102 is over $2 in 1000 piece quantities. The
CP2102N is just over a dollar. Which of these do you think could be
used on a board that sells for $1.25 assembled and shipped? NEITHER!

These boards output 4.3 on the Vdd rail because someone screwed up in
designing the chip and they aren't going to pay to fix them. Before you
buy any of these boards, try asking the vendors if the V3.3 is actually
3.3 volts. See if any say that is correct. When you buy on ebay, you
only need to start a "not as described' complaint and they will refund
your money and you get to keep the item. Otherwise, you can try your
fixes. Or you can get device that works.

You could get an external USB to serial cable that plugs into a header
in the box. Then you don't care so much how small it is and you can get
one that works properly.

--

Rick C

[Peabody]

rickman says...

> I think you have a crap chip on your board and you
> should drop it rather than try to mess with it.

> To see what is wrong, you can try doing one thing I
> thought of. I found a web page where some guys building
> a board with one of these had trouble because they
> shorted the USB VBUS to their 3.3 volt line. This would
> give different symptoms depending on what your USB power
> source is. First, measure the 5 volt rail. If it is
> also 4.3 volts try disconnecting it and measure the
> resistance between the 3.3 and 5 volt lines. If this is
> not your problem I expect your board has a cloned chip
> and will never work right.

The 5V rail on mine is 5V. It's the USB 5V line. Those
guys had the 3.3V pin routed directly to VBUS, which is not
the case in mine. I've traced everything out, and mine is
routed correctly - exactly the same as another adapter of a
different design, but using the same chip, which works
perfectly, 3.3V and all.

> One other point. When I look at the Silabs site I find
> the CP2102 and the CP2102N. The CP2102 is over $2 in
> 1000 piece quantities. The CP2102N is just over a
> dollar. Which of these do you think could be used on a
> board that sells for $1.25 assembled and shipped?
> NEITHER!

I understand your point, but the other adapter - the one
that works - cost me $1.45, delivered, and it's the CP2102,
or if it's a counterfeit, it's one that works. And that
adapter has a full size USB connector on it, plus a six-pin
header, all for $1.45. Did I mention delivered from China?
So I don't know that the price tells us anything.

> These boards output 4.3 on the Vdd rail because someone
> screwed up in designing the chip and they aren't going
> to pay to fix them. Before you buy any of these boards,
> try asking the vendors if the V3.3 is actually 3.3
> volts. See if any say that is correct. When you buy on
> ebay, you only need to start a "not as described'
> complaint and they will refund your money and you get to
> keep the item. Otherwise, you can try your fixes. Or
> you can get device that works.

I think it's entirely possible that this chip is a fake, or
maybe a run that Silabs screwed up and this guy bought them
for a song. But I just want to finish running down
the possibility that it's an EEPROM setting. A guy on the
Silabs forum has promised me a command line method to reset
the chip to the factory default. If he comes through, then
I'll have the answer. If the reset gives me 3.3V, then I
can make this adapter work. If not, then the chip is
defective, and I think rather than risking the IR LED
method, I'll use an adapter that works right even if it's a
bit larger. The LED method would also require a pulldown
resistor for when USB isn't connected, so that's two extra
parts to make it work, and at some point I need to stop
adding parts. If the forum guy doesn't come through, then
I think it would be time (maybe past time) to pick another
adapter.

> You could get an external USB to serial cable that plugs
> into a header in the box. Then you don't care so much
> how small it is and you can get one that works properly.

Since I've gone to the trouble to do this, including writing
a Windows program to do the flashing and a custom BSL for
the processor, I hoped to make it useful to anyone who wants
to do this. And the idea was to make it so no Launchpad is
needed, just a regular USB cable. I'd like to stick with
that if I can.

[rickman]

The other parts are very, very inexpensive, but how can they buy genuine
CP2102 chips and sell the module for $1.49? It's simple math. Neither
of these modules are using a genuine CP2102, or at least no legal chips.
I've heard of overseas fabs (and they are pretty much *all* overseas)
selling overrun production or testing rejects on the grey market. So
you get what you pay for.

I have some 100 watt LED modules that seem to light up all the LEDs.
But if you throttle the current back you see that some of the LEDs are
brighter and others don't light at all. They are factory rejects that
cost nearly nothing. I only paid a dollar for some of them and they are
barely worth the price.

I'm not clear. This is just a generic serial port, right? What special
software is needed to use it? Are you doing the trick to drive reset on
your target from a handshake line via a capacitor so it can be rest
after flashing? How would using the CP2102 be any different from using
the serial cable? Just the fact that it's the cable that's special
rather than the CP2102?

--

Rick C

[Peabody]

rickman says...

> I'm not clear. This is just a generic serial port,
> right? What special software is needed to use it? Are
> you doing the trick to drive reset on your target from a
> handshake line via a capacitor so it can be rest after
> flashing? How would using the CP2102 be any different
> from using the serial cable? Just the fact that it's
> the cable that's special rather than the CP2102?

The idea is to build the CP2102, or a module containing one,
into the circuit board of an MSP430G project. In the case
of a G2553, it already has code supporting this in ROM, and
with a special sequence on Reset and one other pin, it will
boot up into BSL. And TI provides a Windows command line
program that carries out the flashing. The /Reset capacitor
won't work with the special signalling pattern. It has to
be a direct connection.

The G2231 however has no such code, or a serial port. TI
provides one version of a software UART "custom" BSL that
would exist down in INFO memory, and it must occupy no more
than 248 bytes. I made some changes and optimizations so it
would work better. That code will go into BSL if the chip
is powered up from USB, or after it is reset if self powered
but connected to USB. Since I'm using the 74VHCT50A, I
don't need the capacitor - it can be a direct connection
from DTR to /Reset.

But TI provided no program on the Windows side to carry out
the flashing - the protocol is completely different from the
G2553. So I had to write that. Fortunately, Silabs
provides an excellent app note with all the API calls to
talk to the CP2102.

In either case, aside from having the software and the drivers
installed, all the use would need is a regular USB cable, which
he almost certainly already has.

But again, nobody needs this if they are willing to buy a
Launchpad for about $15 delivered and then mess with the
connections from the Launchpad to the MSP device. But I
just wanted to make it cheaper and easier on any ultimate
user in cases where the device ends up being used by lots of
civilians. And it seems that the cost of embedding all this
in the device is pretty low. There are lots of modules
under $2, and of course it doesn't have to be a CP. I
picked that only because it was used in this tiny CJMCU
module that even has the right USB socket built in. So it
would work perfectly for the G2231, but not the G2553
because it doesn't bring out RTS.

The 50A is 40 cents. And if needed, a mini USB socket is a
dime or less. And the MOSFET is 49 cents if that's needed.
So something around $3 in total, not counting the cost of
board space.

So I'm just trying to create a complete, integrated system
for doing it this way in case anyone ever wants it.

[Peabody]

Peabody says...

> I think it's entirely possible that this chip is a fake,
> or maybe a run that Silabs screwed up and this guy
> bought them for a song. But I just want to finish
> running down the possibility that it's an EEPROM
> setting. A guy on the Silabs forum has promised me a
> command line method to reset the chip to the factory
> default. If he comes through, then I'll have the answer.

He did come through, but things took a strange turn. I was
set to test the power mode, but when I plugged in the
adapter, nothing happened, and Device Manager showed nothing
there. So apparently when I was probing this and that to
figure out the circuit, I must have damamged something since
this thing is deader than four dollars. So I can't test
a CJMCU device until the second one I ordered from the same
source arrives next week.

But, in the mail today I received two of the same type of
adapter that I ordered two weeks ago from an entirely
different source (but also in China). So I was going to
test them instead, but it turned out that the 3.3V output
on these is actually 3.3V. Yay!!!

This batch is the identical board to the CJMCU parts, but
they do not have that label on the back even though the
picture in the Ebay auction shows it as CJMCU. In fact
there's nothing on the back. So apparently there's a second
source for these parts, and they work correctly. I've
written to the seller asking if he will be a reliable source
for the non-CJMCU parts.

So this is really better outcome than I had hoped for.
Well, I guess you get lucky sometimes.

Oh, and it's $1.37 with free shipping. Could still be
counterfeit, I guess, but at least is works.

[rickman]

Good news on the modules. So how do we tell which vendors sell the good
board and which sell the dodgy board?

> Oh, and it's $1.37 with free shipping. Could still be
> counterfeit, I guess, but at least is works.

Let me ask you to explain why this device is so important. I read your
other post but I'm still not completely clear. Let me restate what I
perceive is going on.

You are building a board with an MSP430 processor. As a means to
program the MSP430, you want the CP2102 device to be built onto your
board so it is always there, but it will be added as a chip, not the
module. That is the primary reason you are doing any of this, right?

For prototyping you are buying these very small modules with the chip
and micro-usb connector. I don't recall just what your requirements are
to interface these modules to the prototype MCU board (which I assume is
one of the launchpads or other TMS340 prototyping boards), but it
requires that you use a buffer chip to isolate the CP2102 when it has no
power via the USB connector. So your board will need the CP2102 and the
buffer chip to be built in.

On top of that there is a problem with some unknown number of these
micro-USB CP2102 modules that it outputs 4.3 volts instead of 3.3 volts
both on the Vcc pin and the signal pins. In addition, this module does
not bring out to an I/O connector the handshake signal needed to reset
the MCU. I assume you have a way of soldering to the handshake pin or
something.

So the part I still don't get is why you are so wedded to this module
and why it needs to be so small and so inexpensive. You talk as if this
will be something others will be able to use, but other than the devices
built into your board, this will not be an easy solution for others.
They will have to buy the CP2102 module, test it to see if it is the
wrong voltage, buy the buffer chip and then find a way to connect these
components together.

Since this will only be used with the MSP430 and nearly all the
prototyping boards available with these chips also contain a programmer
USB interface chip, why is this even useful? Why doesn't it make more
sense for you to just get your prototype done using whatever boards you
can find that work and then focus on the board you are fabricating?

--

Rick C

[Peabody]

rickman says...

> Good news on the modules. So how do we tell which
> vendors sell the good board and which sell the dodgy
> board?

Well, I know of one seller for the good boards. Other than
that, a buyer would have to ask the seller whether his stock
actually has the CJMCU legend on the back, or is blank, blank
being what you want.

> Let me ask you to explain why this device is so
> important. I read your other post but I'm still not
> completely clear. Let me restate what I perceive is
> going on.

Well, I would admit that it makes less sense to me now than
in the beginning.

> You are building a board with an MSP430 processor. As a
> means to program the MSP430, you want the CP2102 device
> to be built onto your board so it is always there, but
> it will be added as a chip, not the module. That is the
> primary reason you are doing any of this, right?

Not necessarily. If someone is fluent in surface mount,
then both the CP2102 (QFN) and the buffer (TSSOP) would just
be additional parts to provide space and route. But the
two MSP430 processors I'm working with also come in DIP
packages. So someone could actually do a through-hole board
and use the module instead of the QFN, and the buffer in
SOIC, which even I can solder by hand, although it's not
pretty to watch.

This started with a friend who thinks there may be a lot of
interest in an MSP project he's doing, but is concerned that
flashing new firmware versions would require a user to buy a
Launchpad. I was just trying to figure out if it's
feasible to build the USB function into the project so BSL
could be done with just a USB cable which most already have.

And having started on this, I just thought I would finish it
up and provide a complete system that anyone could use if
needed. That includes the hardware side, but also software
at both ends.

Of course TI makes MSP processors with built-in USB, but I
think those would be more expensive, and only surface mount.
And an Arduino with built-in USB could also be used, but at
those would require much higher power than MSP, so batteries
wouldn't last as long. But I would say at this point that
the requirement for the buffer makes this a marginal idea at
best. Anyway, the software is done, so I just want to make
sure the hardware actually works, then write it up and be
done with it.

> For prototyping you are buying these very small modules
> with the chip and micro-usb connector. I don't recall
> just what your requirements are to interface these
> modules to the prototype MCU board (which I assume is
> one of the launchpads or other TMS340 prototyping
> boards), but it requires that you use a buffer chip to
> isolate the CP2102 when it has no power via the USB
> connector. So your board will need the CP2102 and the
> buffer chip to be built in.

I don't have in mind a prototyping board. I assume a project
using this stuff would be designed and routed and purchased
from service like OSHPark. So a board designed for the
particular project. The module has six through-hole
connections, so it could be soldered to matching holes in
the main board if someone didn't want to do surface mount.
So that's why the smaller the module, the better.

> On top of that there is a problem with some unknown
> number of these micro-USB CP2102 modules that it outputs
> 4.3 volts instead of 3.3 volts both on the Vcc pin and
> the signal pins. In addition, this module does not
> bring out to an I/O connector the handshake signal
> needed to reset the MCU. I assume you have a way of
> soldering to the handshake pin or something.

Yes, the voltage issue is a problem. But if the processor
is to be powered by USB through the buffer, it will know on
boot that it should go into BSL because its RXD pin, with
pulldown enabled, will be high. In any case, the software
on the Windows side can always toggle the DTR->Reset line
low and reset the processor. So I don't think there's a
need for a reset button.

> So the part I still don't get is why you are so wedded
> to this module and why it needs to be so small and so
> inexpensive. You talk as if this will be something
> others will be able to use, but other than the devices
> built into your board, this will not be an easy solution
> for others. They will have to buy the CP2102 module,
> test it to see if it is the wrong voltage, buy the
> buffer chip and then find a way to connect these
> components together.

Well, I expect all of this to be done by whoever is creating
a project, not the ultimate users. So it's just two
additional items they have to procure and deal with.

> Since this will only be used with the MSP430 and nearly
> all the prototyping boards available with these chips
> also contain a programmer USB interface chip, why is
> this even useful? Why doesn't it make more sense for
> you to just get your prototype done using whatever
> boards you can find that work and then focus on the
> board you are fabricating?

I guess I gave the wrong impression somewhere along the
line. There is no prototyping board involved.

As to whether it will be useful, I don't really know. But
it's been interesting to work on.

[Lasse Langwadt Christensen]

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

[rickman]

On 4/6/2017 10:24 AM, Peabody wrote:
> rickman says...
>
> > Good news on the modules. So how do we tell which
> > vendors sell the good board and which sell the dodgy
> > board?
>
> Well, I know of one seller for the good boards. Other than
> that, a buyer would have to ask the seller whether his stock
> actually has the CJMCU legend on the back, or is blank, blank
> being what you want.

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

> > Let me ask you to explain why this device is so
> > important. I read your other post but I'm still not
> > completely clear. Let me restate what I perceive is
> > going on.
>
> Well, I would admit that it makes less sense to me now than
> in the beginning.
>
> > You are building a board with an MSP430 processor. As a
> > means to program the MSP430, you want the CP2102 device
> > to be built onto your board so it is always there, but
> > it will be added as a chip, not the module. That is the
> > primary reason you are doing any of this, right?
>
> Not necessarily. If someone is fluent in surface mount,
> then both the CP2102 (QFN) and the buffer (TSSOP) would just
> be additional parts to provide space and route. But the
> two MSP430 processors I'm working with also come in DIP
> packages. So someone could actually do a through-hole board
> and use the module instead of the QFN, and the buffer in
> SOIC, which even I can solder by hand, although it's not
> pretty to watch.

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.

> This started with a friend who thinks there may be a lot of
> interest in an MSP project he's doing, but is concerned that
> flashing new firmware versions would require a user to buy a
> Launchpad. I was just trying to figure out if it's
> feasible to build the USB function into the project so BSL
> could be done with just a USB cable which most already have.

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.

> As to whether it will be useful, I don't really know. But
> it's been interesting to work on.

I certainly learned a lot about the CP2102.

--

Rick C

[rickman]

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?

--

Rick C

[Peabody]

rickman says...

> 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.

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.

> 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.

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.

> 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.

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.

> 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.

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.

> I certainly learned a lot about the CP2102.

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

[Peabody]

rickman says...

> 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?

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.

[rickman]

It's $10...

> > 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.
>
> 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.

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.

> > 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.
>
> 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.
>
> > I certainly learned a lot about the CP2102.
>
> Me too. Perhaps more than I absolutely needed to know.

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.

--

Rick C

[Peabody]

Peabody says...

> As described in other threads, I'm working on embedding
> a USB-to-serial adapter in one of my MSP430 projects,
> and I've found the perfect adapter for this purpose.
> It's smaller than the other adapters, has the right
> pinout, and even comes with a micro-USB socket onboard.

> The chip used in the adapter is the CP2102 from Silicon
> Labs. But there's a problem. The output labeled in
> silkscreen as "3.3V" actually measures 4.25V. And it's
> not just the one I received:

> https://www.youtube.com/watch?v=8qcGdB2yVd8

> It seems that all modules of this design are like that.
> And the CP2102 is definitely a 3.3V part. So it's
> either something wrong with the board layout, or
> something the manufacturer has done intentionally to get
> that output up close to 5V.

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.

[LucTa]

Il giorno domenica 2 aprile 2017 19:34:53 UTC+2, Peabody ha scritto:
> As described in other threads, I'm working on embedding a USB-to-serial
> adapter in one of my MSP430 projects, and I've found the perfect adapter for
> this purpose. It's smaller than the other adapters, has the right pinout, and
> even comes with a micro-USB socket onboard.
>
> The chip used in the adapter is the CP2102 from Silicon Labs. But there's a
> problem. The output labeled in silkscreen as "3.3V" actually measures
> 4.25V. And it's not just the one I received:
>
> https://www.youtube.com/watch?v=8qcGdB2yVd8
>
> It seems that all modules of this design are like that. And the CP2102 is
> definitely a 3.3V part. So it's either something wrong with the board
> layout, or something the manufacturer has done intentionally to get that
> output up close to 5V.
>

> This is a deal killer if it can't be corrected because all the MSP430 parts
> would likely be destroyed by voltage that high. Oh, and it's not just the
> 3.3V output, it's DTR and TXD too.
>
> I've been on the Silabs forum, and so far nobody has come up with an
> explanation for how this could be done. If everything is connected as it
> should be, which appears to be the case, this shouldn't be possible.
>

> So what I'd like to do is contact the manufacturer and try to get some
> information about how this was done in the hope that it will be something I
> can reverse easily. The markings on the board are:
>
> CJMCU
>
> BQ09615H2 2925a
>
> I've seen CJMCU elsewhere in relation to drones and other stuff, but Google
> gives me no home page or any other info about them.
>
> Does anyone here know anything about CJMCU? Where would you go to find them?

http://www.cjmcu.com
Did you search it?

[Peabody]

LucTa says...

> http://www.cjmcu.com
> Did you search it?

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.

[Michael A Terrell]

Peabody wrote:
> LucTa says...
>
> > http://www.cjmcu.com
> > Did you search it?
>
> Yes I did, and by the way that link doesn't work.

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.

[rickman]

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

--

Rick C

Viewed the eclipse at Wintercrest Farms,
on the centerline of totality since 1998

[Peabody]

rickman says...

>Peabody wrote on 12/28/2017 9:46 PM:
>> LucTa says...
>>
>> > http://www.cjmcu.com
>> > Did you search it?
>>
>> 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.
>
>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:

http://www.eevblog.com/forum/projects/adding-tx-and-rx-led-to-ch340g/

See replies 11, 21 and 24.

[Peabody]

Peabody says...

And here is the Silabs forum thread:

https://www.silabs.com/community/interface/forum.topic.0.13.html/cp2102_-
_4_24v_outpu-OK2U

[rickman]

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

[Jasen Betts]

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

--
This email has not been checked by half-arsed antivirus software