Searching \ for '[EE] USB to RS232 converter' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/io/serials.htm?key=rs232
Search entire site for: 'USB to RS232 converter'.

Exact match. Not showing close matches.
PICList Thread
'[EE] USB to RS232 converter'
2005\09\25@104400 by olin piclist

face picon face
A couple of weeks ago Wouter asked if I would test out a particular USB to
RS-232 converter to see if it worked with the EasyProg and ProProg.  I said
I would, Wouter sent me the adapter, and here is my report.

I performed the tests on a normal desktop PC running Windows 2000 service
pack 4.  This is one of the operating systems claimed to be supported on the
outside cover of the adapter packaging.  There is no brand name or model
number on the packaging I received so I can't really say what adapter this
is.

The software installation went reasonably smoothly, although there were some
discrepancies between what the documentation said and the exact wording
displayed by Windows.  I think the documentation assumed a form of Windows
9x (98SE and ME are listed as supported) without mentioning difference that
could be encountered on other Windows variants.  The differences were mostly
minor except the pathname to enter for the installation files.  The name of
the top level directory was a little different, and they only mentioned and
showed the WIN98 path.  It wasn't hard to browse the disk to find the right
(at least I think so) path, but this could be rather intimidating for
someone not comfortable with poking around on their own.  Once the software
was installed, the adapter showed up as COM4 in device manager.  I rebooted
the machine at that point, even though the documetation didn't say anything
about that.

As a quick go/nogo test I connected the adapter to an EasyProg and ran
PIC_PROG.  It couldn't communicate at all.  It acted just like the serial
port wasn't connected or the programmer was powered down.  I tried it with a
real serial port and a normal extension cable and everything worked fine.
However since Wouter took the trouble to send the adapter, I wanted to
provide more than the short answer "It doesn't work".  This also delayed my
response a few days before I could get some time to investigate further.

The next test was to look at the lines with a scope.  With nothing
connected, the adapter was putting out about -10V on pin 3 (TX), 4 (DTR),
and 7 (RTS).  All other pins showed 0V on the scope.  Of course all these
measurements were relative to pin 5 (GND).

I brought up Hyperterm on COM4 at 9600 baud, 1 stop bit, no parity, no flow
control.  It should only be using RX, TX, and GND in this configuration, and
I have used it that way before talking to PICs with only those lines.  No
matter what I typed in Hyperterm, TX stayed solidly at -10V.

To see if receive works, I connected the adapter to another COM port on the
same machine via a cable that crosses RX/TX and RTS/CTS and passes GND.  The
adapter could recieve characters from the other COM port and display them in
the Hyperterm window, but sending didn't work.  I verified with a scope that
the TX line from the adapter never moved.

Next I thought that maybe the adapter still required some of the handshaking
lines to be driven even though it wasn't supposed to with the COM port setup
I had specified.  I grabbed a DB9-F solder cup connector and wired pin 7
back to pin 8 (RTS back to CTS) and pin 4 back to pins 1 and 6 (DTR back to
DSR and DCD).  This should allow the adapter to send even if "hardware
handshaking" were specified.  I then ran RX, TX, and GND to another DB9-F
with RX/TX crossed over and connected that to another COM port on the same
machine.  Again the adapter could receive, but would not transmit.  The TX
line from the adapter stayed at -10V as verified with a scope.

That's as far was I went.  In the end this adapter never sent a single byte
out, despite all the various things I tried to make it do so.  I suppose the
particular unit I have could be broken, but I've seen this from other USB
adapters too in the past.  It seems like such a simple and obvious test case
that it's hard to imagine anyone shipping adapters that behave in this way,
but at this point I've run out of ideas (and patience) trying to make it
send anything.

So the long answer is "It doesn't work".


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\25@112208 by Robert Young

picon face
----- Original Message -----
From: "Olin Lathrop" <spam_OUTolin_piclistTakeThisOuTspamembedinc.com>
To: "Microcontroller discussion list - Public." <.....piclistKILLspamspam@spam@mit.edu>
Sent: Sunday, September 25, 2005 9:43 AM
Subject: [EE] USB to RS232 converter


> A couple of weeks ago Wouter asked if I would test out a particular USB to
> RS-232 converter to see if it worked with the EasyProg and ProProg.  I
said
> I would, Wouter sent me the adapter, and here is my report.
>

-- snip --

>
> So the long answer is "It doesn't work".
>

Olin,
I've build a handful of custom RS232 to USB adapters and dealt with some of
the silly software issues associated in getting "legacy" equipment to run.
Not that I'm calling the EasyProg "legacy".  Just referring to the
disappearance of real RS232 ports on new PCs.

Anyway, I am going to order an EasyProg kit (I need a spare programmer,
loaned my spare to a client and haven't gotten it back).  I have a box of
commercial USB adapters I've bought over the years with different chip sets
as well as some spares of custom models.  It will probably be a couple of
weeks but I can sift through what I have on hand.

By the way, is the PIC18F6525 (via the ICD pins or course, hard to fold up a
TQFP and fit in a DIP socket) in the supported list for the EasyProg?  I
didn't see it specifically, I did see the PIC18F6520.  I do mostly surface
mount stuff now so I'd be using the RJ12 socket or header anyway.

Robert Young
YR Consulting
rwyoungspamKILLspamieee.org

2005\09\25@121828 by olin piclist

face picon face
Robert Young wrote:
> By the way, is the PIC18F6525 (via the ICD pins or course, hard to
> fold up a TQFP and fit in a DIP socket) in the supported list for the
> EasyProg?  I didn't see it specifically, I did see the PIC18F6520.

The 18F6520 is supported but the 18F6525 is not.  The 18F 6525, 6621, 8525,
and 8621 have a unique programming spec that I haven't looked at yet.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\25@152106 by Wouter van Ooijen

face picon face
> So the long answer is "It doesn't work".

Let's start stating that I certainly don't think you would write
anything but what you observe.

But the funny thing is that I produce all my Wisp628's (chips, kits,
assembled) with a Wisp628 connected to exactly the same adapter! So it
surely works under some circumstances (I use XP).

If I find a W2k somewhere I will certainly try it, maybe there is
something special with the combination of this adapter and w2k.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\25@152106 by Wouter van Ooijen

face picon face
> I've build a handful of custom RS232 to USB adapters and
> dealt with some of
> the silly software issues associated in getting "legacy"
> equipment to run.
> Not that I'm calling the EasyProg "legacy".  Just referring to the
> disappearance of real RS232 ports on new PCs.

Olin reported that Hyperterm could not succesfully use the virtual port,
so I don't think it is Olin's PC software that is the culprit. More
likely the win2k drivers.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\25@155802 by Stef Mientki

flavicon
face
Generally speaking USB under windows is bull shit !!

Win98SE doesn't work at all, winXP gives numerous problems.

Unfortunatly, we need to get used to these kind of problems.

Are you talking about your FTDI module Wouter ?
We've just discovered a new problem with the FTDI chip,
which we've dicussed with FTDI and hope to solve next week,
tweaking some USB parameters and using a yet unreleased driver.
From my experience with FTDI, D2XX-drivers are much more stable then
VCP-drivers.

Stef





Wouter van Ooijen wrote:

{Quote hidden}

2005\09\25@160219 by olin piclist

face picon face
Wouter van Ooijen wrote:
> But the funny thing is that I produce all my Wisp628's (chips, kits,
> assembled) with a Wisp628 connected to exactly the same adapter! So it
> surely works under some circumstances (I use XP).
>
> If I find a W2k somewhere I will certainly try it, maybe there is
> something special with the combination of this adapter and w2k.

I have a Win XP Pro and Win XP Home at two other places, although less
ability to diagnose problems.  I'll try to check out the adapter on those
systems.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\25@161651 by Wouter van Ooijen

face picon face
> Are you talking about your FTDI module Wouter ?

no, about http://www.voti.nl/shop/p/M-USB-SERIAL-1.html

> We've just discovered a new problem with the FTDI chip,
> which we've dicussed with FTDI and hope to solve next week,
> tweaking some USB parameters and using a yet unreleased driver.
>  From my experience with FTDI, D2XX-drivers are much more stable then
> VCP-drivers.

But VCP is so damn easy!

Can you give details about your problems?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\25@164142 by Stef Mientki

flavicon
face

>But VCP is so damn easy!
>  
>
Yes, you don't have to rewrite the existing programs ;-)

And I do have to admit that with the lower baudrates upto 115k,
I never have seen problems with the VCP drivers !!

>Can you give details about your problems?
>
>  
>
On a few PCs, the device disconnects spontanuously.
Then the only solution is to unplug / replug the USB device,
so we cann't do the reset in software,
which is normally possible (one of the advantages of D2XX).

We also know that the problem can be influenced
by using different cables USB2.0 certified or not.

An extra problem we've is that it's very hard to define which drivers
windows should use.
WinXP and "managed workplaces" both have mechanism to restore a driver
which they find is better :-(

The suggestions from FTDI where
- increase  FT_SetResetPipeRetryCount
- use D30000BX-driver (not yet released)

I'll post the solution, if we find one.

Stef

>  
>

2005\09\25@170408 by Bob Blick

face picon face
On 25 Sep 2005 at 22:16, Wouter van Ooijen wrote:
> > Are you talking about your FTDI module Wouter ?
> no, about http://www.voti.nl/shop/p/M-USB-SERIAL-1.html

It's a Prolific PL-2303 chip, I bet. Hyperterminal does not work
properly with it under 98SE and XP as verified by me.

However I use these adapters(or their ilk) with Bray Terminal and
also in Delphi programs I've written using the tpapro library and they
works fine. Perhaps Hyperterminal and Olin's program try low-level
stuff that doesn't work with a virtual com port.

Olin, the Prolific chips are the most, well, "prolific" of the USB-
RS232 chips on the market, and although I agree there are issues, if
you are writing an application and trying to hit the largest part of the
population... maybe you should look at how you access the serial
port and switch methods to something more generic.

Cheerful regards,

Bob

2005\09\25@172109 by olin piclist

face picon face
Olin Lathrop wrote:
> I have a Win XP Pro and Win XP Home at two other places, although less
> ability to diagnose problems.  I'll try to check out the adapter on
> those systems.

I installed the adapter on a Win XP Home system.  This time I ignored the
instructions (which are Win 98 specific anyway) and let Windows pick the
driver from the installation CD.

The low level tests worked in both directions when fed back to another COM
port on the same machine.  This was using a short F-F adapter I made earlier
that loops back all the handshaking lines at the USB adapter end and only
connects TX, RX, and GND to the other end.

I then tried an EasyProg but that didn't work.  I don't have the equipment
here to diagnose it, but one difference is that the handshake lines are not
looped back in this configuration.  When I get back to my real office, I'll
make a straight thru M-F adapter that loops back the handshake lines on one
end and transfers only RX, TX, and GND.  I'm also thinking of removing the
driver on the Win 2000 system and letting Windows pick what it wants from
the disk.  The pathnames in the directions didn't exist, and it seemed there
might be more than one choice.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\25@180946 by Dave Tweed

face
flavicon
face
Stef Mientki <.....s.mientkiKILLspamspam.....mailbox.kun.nl> wrote:
{Quote hidden}

Thanks so much for describing this! I've been chasing what sounds like the
exact same problem on one of my projects. Let me know if you find a good
solution.

I have a board (FPGA + DSP) that collects data and sends a steady stream
of status information (a few hundred bytes/second) via an FT245BM to a
host PC for monitoring. We use two different Sony Vaio laptops. One is
slightly older than the other, but they are otherwise running the same
version of Windows XP and the FTDI D2XX drivers. The older laptop
experiences these disconnects significantly more often than the newer
one, and I'm wondering if there's some subtle difference in the USB host
implementation (motherboard chipset?) between the two laptops.

-- Dave Tweed

2005\09\26@025905 by Chen Xiao Fan

face
flavicon
face
The ultimate solution is to use USB and not RS232 IMHO. RS232 is
already a legacy port in modern PCs. Of course it still has
its place in quite a lot of product. However just look at
the development tools from Microchip and others: USB
(PCIKit 1/2) is the way to go. USB/RS232 (MPLAB ICD2 and Promate
III) are even better but the cost will be higher. They even start
to offer USB-to-Parallel adapter for ICE2000 customers.

Then there will no issue with FTDI an Prolific and so on.

The cost of USB MCUs are getting lower and lower. Wouter is
using a PICKit 1 clone for his class since it is cheaper
than a Wisp628 plus a power supply. The only problem is
that USB is more difficult to program on the firmware
side and the host side (both Windows and Linux and other
OSes) and the learning curve is pretty steep. I just start to
learn USB under Linux/Windows and I have already met plenty
of problems under Linux/Windows. Now I have two threads in
the libusb and libusb-win32 mailing lists respectively regarding
PICkit 2. However the leagues of PIC programmer developers like
Wouter and Olin and others are all very good at programming and I
am sure they will have no problems with USB.

As for problems related to HyperTerminal, I think WH Tan has
some good findings in one Microchip Forum post. It is built-in
with Windows but it does create some problems.


Regards,
Xiaofan

{Original Message removed}

2005\09\26@030728 by Vasile Surducan

face picon face
On 9/25/05, Wouter van Ooijen <EraseMEwouterspam_OUTspamTakeThisOuTvoti.nl> wrote:
> > So the long answer is "It doesn't work".

I have produced for my own requirements a small numbers of similar
board (only the same principle) like Wouter does. No one from the
people I sold the boards does report it don't work.
The board picture is on the bottom of the page:

http://www.surducan.netfirms.com/module.html

So I have doubt's Wouter's board is not working. Maybe you have done
mistakes in installing the windows driver for the USB which is a pain
in the ass on some Win systems.

cheers,
Vasile




{Quote hidden}

> -

2005\09\26@032142 by Chen Xiao Fan

face
flavicon
face
Definitely a lot of difference with the host USB implementation.

I have a no-name USB gamepad (HID device) which only works with
INTEL chipsets. The two desktops (one is Nvidia NFORCE 2 based and
the other is Nvidia NForce 3 based) do not work with it. NVIDIA
blames the gamepad vendors for that. My brother works for
Creative and they had some hard time to get their USB sound
cards and MP3 players working with VIA chipsets. VIA blames
them but later VIA chipsets got better with the USB implementations.

Regards,
Xiaofan

-----Original Message-----
From: Dave Tweed [picspamspam_OUTdtweed.com]
Sent: Monday, September 26, 2005 6:10 AM
To: @spam@piclistKILLspamspammit.edu
Subject: Re: [EE] USB to RS232 converter
...
The older laptop
experiences these disconnects significantly more often than the newer
one, and I'm wondering if there's some subtle difference in the USB host
implementation (motherboard chipset?) between the two laptops.

-- Dave Tweed

2005\09\26@033655 by Wouter van Ooijen

face picon face
>> So the long answer is "It doesn't work".
>
> I have produced for my own requirements a small numbers of similar
> board (only the same principle) like Wouter does. No one from the
> people I sold the boards does report it don't work.
> The board picture is on the bottom of the page:
> http://www.surducan.netfirms.com/module.html

You are referring to my FT232 boards. They use the FTDI drivers, which
are different from the no-brand/prolific drivers that Olin tsted.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\26@064808 by Alan B. Pearce

face picon face
>> So the long answer is "It doesn't work".
>
>Let's start stating that I certainly don't think
>you would write anything but what you observe.
>
>But the funny thing is that I produce all my Wisp628's
>(chips, kits, assembled) with a Wisp628 connected to
>exactly the same adapter! So it surely works under
>some circumstances (I use XP).
>
>If I find a W2k somewhere I will certainly try it,
>maybe there is something special with the
>combination of this adapter and w2k.

I know that other equipment (not related to this list) had problems with
ordinary serial ports on W2k, and it required pins 8 and 20 to be linked (on
a 25 pin connector, not sure of the equiv. on a 9 pin). I don't recall Olin
mentioning these ones in his report - do you link these signals on the
Wisps?

It seems to driver in W2k changed compared to earlier versions. I assume XP
has the same problems.

2005\09\26@070728 by Wouter van Ooijen

face picon face
> I know that other equipment (not related to this list) had
> problems with
> ordinary serial ports on W2k, and it required pins 8 and 20
> to be linked (on
> a 25 pin connector, not sure of the equiv. on a 9 pin). I
> don't recall Olin
> mentioning these ones in his report - do you link these signals on the
> Wisps?

no, and I never heared of a problem that was cured by connecting 8 and
20 (I assume you refer to a DB25, not to the now standard DB9).

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\26@074610 by Alan B. Pearce

face picon face
>no, and I never heared of a problem that was cured
>by connecting 8 and 20

Well I know that the specific item in question had problems with the
software opening the port until that was done in the hardware.

>(I assume you refer to a DB25, not to the now standard DB9).

Yes, that is why I mentioned the 25 pin and 9 pin connectors (without
specifically calling then DB ...)

2005\09\26@081138 by olin piclist

face picon face
Alan B. Pearce wrote:
> I know that other equipment (not related to this list) had problems with
> ordinary serial ports on W2k, and it required pins 8 and 20 to be
> linked (on a 25 pin connector, not sure of the equiv. on a 9 pin). I
> don't recall Olin mentioning these ones in his report - do you link
> these signals on the Wisps?

I have used ordinary serial ports on Windows 2000 systems many times without
problems.  The main machine in my Embed office is a Win 2000 system, and my
programmers and many other things run fine on it using only RX, TX, and GND.
My driver software for these devices specifically sets up the COM port so
that RTS/CTS, XON/XOFF, etc, are disabled.  My software and hardware work
fine using a real COM port on every Windows system I've ever tested them on.

I can think of no reason that it shouldn't be possible to emulate a COM port
over USB.  I think the problems are most likely in the Windows driver
implementation.  You actually have to wake up and know what you're doing and
read a lot of Microsoft documentation to create a good Windows driver,
whereas the rest of a USB to serial adapter isn't that difficult.  Given the
high volumes of these things, it's surprising to me that the manufacturer
can't spend the extra $10-30K to have the driver written properly and tested
in many different configurations.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\26@112915 by William Chops Westfield

face picon face
On Sep 26, 2005, at 5:12 AM, Olin Lathrop wrote:

> My driver software for these devices specifically sets up the
> COM port so that RTS/CTS, XON/XOFF, etc, are disabled.

Perhaps this is the problem?  If the USB driver/frob, in the interest
of working within USB's bursty nature, refuses to let drivers turn off
(for instance) HW flow control, then you might never get to the point
where looping back signals fools things into working.  When you were
playing with hyperterminal and the USB frob, did you try HT with the
assorted signals ENABLED and looped back, or just with the signals
looped back and still supposedly disabled in HT?

BillW

2005\09\26@114002 by Bob Blick

face picon face

> My driver software for these devices specifically sets up the COM port so
> that RTS/CTS, XON/XOFF, etc, are disabled.  My software and hardware work
> fine using a real COM port on every Windows system I've ever tested them
> on.

That's fine, but two extra routes on your PC board cost you nothing, and
it insures that if there is forced hardware hadshaking, you are covered.
That's why I always loop back 4 to 6, and 7 to 8 when using a three wire
connection.

Cheers,

Bob


2005\09\26@120520 by Roy J. Gromlich - PA

picon face
Likewise - it is always safest to loop back 7 - 8 and 4 - 6 -1 on a DB9 serial port. I usually make them solderable jumper pads so they can be opened or closed as needed. As has been noted here, some serial ports on some machines can be set to TOTALLY ignore the handshaking pins (RTS/CTS DSR/DTR) and others can not - simpler to have the hardware jumpers there is you need them.

RJG
 {Original Message removed}

2005\09\26@124145 by olin piclist

face picon face
William Chops Westfield wrote:
> When you were
> playing with hyperterminal and the USB frob, did you try HT with the
> assorted signals ENABLED and looped back, or just with the signals
> looped back and still supposedly disabled in HT?

I only tried the USB converter with all handshaking disabled in software.

However this is a normal mode of operation for RS-232, is supported by the
Windows calls, and works fine on real RS-232 ports.  I haven't done a
project in a long time that used hardware (RTS/CTS) flow control.  I have
occasionally used XON/XOFF flow control because it works with the simple 3
wire interconnect.

I'm hoping to try again on the Win XP Home system this evening with a
loopback adapter between the USB serial port and the EasyProg.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\26@124423 by olin piclist

face picon face
Bob Blick wrote:
> That's fine, but two extra routes on your PC board cost you nothing, and
> it insures that if there is forced hardware hadshaking, you are covered.
> That's why I always loop back 4 to 6, and 7 to 8 when using a three wire
> connection.

Yes this is easy enough to do in the future, but of course too late for
existing hardware.  I've never had this problem before, so I didn't think of
it as an issue.  In fact we don't know that this is the problem at hand yet.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\26@132158 by Wouter van Ooijen

face picon face
> Perhaps this is the problem?

no, no, no: as I said, I use these converters (or others, or self-build
FT232 ones) with my Wisp628 all the time, without any problem.

In fact I could never get *any* programmer to work reliably with the
on-board serial port of my PC ....

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\26@132158 by Wouter van Ooijen

face picon face
> As has
> been noted here, some serial ports on some machines can be
> set to TOTALLY ignore the handshaking pins (RTS/CTS DSR/DTR)
> and others can not

That's not a hardware issue (except mayebe on some esoteric multi-port
cards), its the (in this case windows-) software and/or OS.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\26@132200 by Wouter van Ooijen

face picon face
> That's fine, but two extra routes on your PC board cost you
> nothing

On my Wisp628 PCB they would not have been free.

> it insures that if there is forced hardware hadshaking,

But exactly who could force this, if it is you yourlself who writes the
PC software?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\26@140559 by Roy J. Gromlich - PA

picon face
Agreed - drivers are the most likely cause.  I have several different USB - RS232 adapters and all work in 3-wire mode with anything I have tried which doesn't actually REQUIRE hardware handshaking. OTOH - I have an old PC clone which requires the hardware loop-back before it will initialize the serial port regardless of how the port is configured.

Actually, if I don't have the jumpers on the PCB I use a DBp solder-cup in a a shell and solder the jumpers on the plug.

RJG
 {Original Message removed}

2005\09\27@044937 by Alan B. Pearce

face picon face
>> That's fine, but two extra routes on your
>>PC board cost you nothing
>
>On my Wisp628 PCB they would not have been free.

?? surely they would if put on the PCB master?

>> it insures that if there is forced hardware hadshaking,
>
>But exactly who could force this, if it is you
>yourlself who writes the PC software?

The problem I heard is not that the software enforces handshaking, rather
that the port driver in the op sys enforces detecting the return signals to
know that there is a device "out there" to connect to, even when the port is
being requested with no handshake.

And HyperTerminal seems to be the worst bit of software to use to verify
this. I always use the "terminal" program from Windows 3.x - it is a much
saner piece of software.

2005\09\27@073721 by Xiaofan Chen

face picon face
Problems with Hyperterminal:
http://forum.microchip.com/tm.asp?m=96946

Some better ones:

1) TeraTerm
http://hp.vector.co.jp/authors/VA002416/teraterm.html

2) Tiny application by Ramon de Klein
http://home.ict.nl/~ramklein/Projects/Serial.html

3) Bray Terminal BRAY
http://bray.velenje.cx/avr/terminal


On 9/27/05, Alan B. Pearce <KILLspamA.B.PearceKILLspamspamrl.ac.uk> wrote:

> And HyperTerminal seems to be the worst bit of software to use to verify
> this. I always use the "terminal" program from Windows 3.x - it is a much
> saner piece of software.
>

2005\09\27@100654 by Roy J. Gromlich - PA

picon face
That is correct - Hyperterminal does its own serial port configuration during initialization - only after it is satisfied with what is out there does it default to the configuration you have selected in System/Ports or wherever your OS puts it.  Using TeraTerm I do not have this problem - using Hyperterminal I do on MOST (not all) computers.  And H-T really does not like USB-RS232 dongles.

I discovered this the hard way many years ago, and have had the 1-6-4 & 7-8 hardware jumpers on every PCB with RS232 that I have done since then.  Cheap insurance.

RJG
 {Original Message removed}

2005\09\27@102419 by olin piclist

face picon face
Roy J. Gromlich - PA wrote:
> That is correct - Hyperterminal does its own serial port configuration
> during initialization - only after it is satisfied with what is out
> there does it default to the configuration you have selected in
> System/Ports or wherever your OS puts it.

This doesn't agree with my experience.  I have used Hyperterm many times to
talk to simple devices that use only RX, TX, and GND without any handshaking
loopback.  Hyperterm definitely does not send anything to the device on
startup, only when a key is pressed.  Maybe you have one of the fancy
terminal emulation types specified?  Maybe some of those do cause it to try
and find what's out there.  In any case, I've never seen the behavior you
describe.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\27@102701 by Spehro Pefhany

picon face
At 10:06 AM 9/27/2005 -0400, you wrote:
>That is correct - Hyperterminal does its own serial port configuration
>during initialization - only after it is satisfied with what is out there
>does it default to the configuration you have selected in System/Ports or
>wherever your OS puts it.  Using TeraTerm I do not have this problem -
>using Hyperterminal I do on MOST (not all) computers.  And H-T really does
>not like USB-RS232 dongles.
>
>I discovered this the hard way many years ago, and have had the 1-6-4 &
>7-8 hardware jumpers on every PCB with RS232 that I have done since
>then.  Cheap insurance.

Yup. Me too. Just copy and paste the DE-9 and connections. ;-)

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
RemoveMEspeffTakeThisOuTspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2005\09\27@104801 by Alan B. Pearce

face picon face
>Roy J. Gromlich - PA wrote:
>> That is correct - Hyperterminal does its own serial port configuration
>> during initialization

And Olin replied

>This doesn't agree with my experience.  I have used Hyperterm
>many times to talk to simple devices that use only RX, TX, and
>GND without any handshaking loopback.  Hyperterm definitely
>does not send anything to the device on startup, only when a key
>is pressed.

It does not send a character on the TX line, but it does play with the
DTR/DSR lines to detect if there is a device it can talk to.


2005\09\27@115443 by Wouter van Ooijen

face picon face
> >> That's fine, but two extra routes on your
> >>PC board cost you nothing
> >
> >On my Wisp628 PCB they would not have been free.
>
> ?? surely they would if put on the PCB master?

The PCB is single-sided, with edge-mounted connectors. Not all required
pins are on the copper side.

> >But exactly who could force this, if it is you
> >yourlself who writes the PC software?
> The problem I heard is not that the software enforces
> handshaking, rather
> that the port driver in the op sys enforces detecting the
> return signals to
> know that there is a device "out there" to connect to, even
> when the port is
> being requested with no handshake.

Might be true, but that would give problems with maybe 50% of the
existing serial peripherals!

> And HyperTerminal seems to be the worst bit of software to
> use to verify
> this. I always use the "terminal" program from Windows 3.x -
> it is a much
> saner piece of software.

I fully agree. I used to copy that program to each new PC, but I am not
sure I still have it. I simply use something like 'xwisp com2 term
19k2'.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


>
> --

2005\09\27@120556 by olin piclist

face picon face
Alan B. Pearce wrote:
> It does not send a character on the TX line, but it does play with the
> DTR/DSR lines to detect if there is a device it can talk to.

But most of my devices have nothing connected to those lines and HyperTerm
still talks to them fine.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\27@123953 by Roy J. Gromlich - PA

picon face
This is interesting - we have the same software (HyperTerminal) exhibiting two mutually exclusive behaviors.  Certainly not the first time something like this has been observed in the WWW (Wonderful World of Windows).

But, of course, it is NOT the same. We likely have different releases of the software running on many different versions of the OS, which is in turn running on different hardware. If one wanted to document the behavior of different release versions of the program, with different PC hardware and different versions of Windows, a pattern would probably emerge.  This, while interesting, is probably not worth the trouble, since the hardware loopback jumpers, when needed, are the universal "fix".

This type of inconsistent behavior is one of the (supposed) reasons for changing over to USB and killing RS232.  As those who have tried to implement USB in a device can tell you, it is not the answer either.
RJG
 {Original Message removed}

2005\09\27@145525 by Aza D. Oberman

flavicon
face
<Alan B. Pearce questions Olin's observation that Hyperterm does not require
hardware handshaking>

<Olin>
> >This doesn't agree with my experience.  I have used Hyperterm
> >many times to talk to simple devices that use only RX, TX, and
> >GND without any handshaking loopback.  Hyperterm definitely
> >does not send anything to the device on startup, only when a key
> >is pressed.
>
> It does not send a character on the TX line, but it does play with the
> DTR/DSR lines to detect if there is a device it can talk to.

The hardware handshaking is likely a requirement of your hardware, not of
Hyperterm. [That's not to defend Hyperterm, which in my experience is about
as reliable and easy to work with as a politician.]

Aza D. Oberman

2005\09\27@151644 by Wouter van Ooijen

face picon face
> The hardware handshaking is likely a requirement of your
> hardware, not of
> Hyperterm. [That's not to defend Hyperterm, which in my
> experience is about
> as reliable and easy to work with as a politician.]

I agree on Hyperterem and/or the politicians, but it is unlikely that
requiring the handshake (if that is the case) is just caused by the
hardware: I use the same hardware without any hardware handshake (just 3
wires, GND, TxD, TxD).

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\27@155519 by Harold Hallikainen

face picon face
When designing 3 wire EIA232 devices that don't need handshaking, I just
connect RTS to CTS and DTR to DSR. That SEEMS to solve many problems
(though Hyperterminal continues to have various problems). I've seen
Hyperterminal throw a bunch of garbage on the screen with no handshake
9600bps 8N1 data over a COMM port (not even USB to 232). If I run Tera
Term, it works great!

Harold


{Quote hidden}

--
FCC Rules Updated Daily at http://www.hallikainen.com

2005\09\27@190639 by Aza D. Oberman

flavicon
face
<Wouter van Ooijen observs on the question of RS-232 hardware handshaking>

>...it is unlikely that
> requiring the handshake (if that is the case) is just caused by the
> hardware: I use the same hardware without any hardware handshake (just 3
> wires, GND, TxD, TxD).

So your particular hardware does not insist on hardware handshaking.

I've got a Rockwell UART here (R6551AP) that get's its knickers in a knot
without handshaking.  I believe the original IBM PC RS-232 adapter also
required hardware handshaking or the UART would simply not pass data.  Both
work fine with the magic 6-4 & 7-8 jumpers.

I still submit that there are variations in hardware that may or may not
require strapping the handshaking before any software can receive or send
data through them.

Aza D. Oberman

2005\09\28@014954 by Wouter van Ooijen

face picon face
> >...it is unlikely that
> > requiring the handshake (if that is the case) is just caused by the
> > hardware: I use the same hardware without any hardware
> handshake (just 3
> > wires, GND, TxD, TxD).
>
> So your particular hardware does not insist on hardware handshaking.

I was the guy that sent a usb-to-serial converter to Olin and I am using
exactly the same converter. So my particular hardware and Olin's
particular hardware are the same!

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\28@041828 by Alan B. Pearce

face picon face
>[That's not to defend Hyperterm, which in my experience is
>about as reliable and easy to work with as a politician.]

And that is about the best description of hyperterm that i have seen. <VBG>

2005\09\28@072137 by Brent Brown

picon face
While watching this thread and quietly thinking I'm glad my USB/serial
converters are behaving OK...the phone rang and everything changed! One
of my customers uses USB/serial converters along with a bootloader
program I've supplied so they can update their PIC boards whenever I send
them a new program. Yesterday they had computer problems and 0 of 3
laptops would program anything! I supply them Bafo brand converters which
use the Prolific PL-2303 chip.

Curiously, one of the converters I tested here worked on my laptop but not
on either of two other Win XP machines. Conversely, another converter, but
same manufacturer/model etc, would not work on any machine. The
symptom was the RS232 would receive OK but not transmit, and the RTS
line which I can normally set/clear with Bray Terminal would not operate.
Sounds similar to what Olin(?) observed.

To cut a long story short I got hold of the latest Windows driver from
http://tech.prolific.com.tw (NOT from Bafo!). Driver is v2.0.2.1 and the file is
named wd_pl2303h-hx-x_v20019v2021.zip. It seems to have successfully
fixed all my problems! It's an InstallShield .exe file too which seems to
remove previous drivers first and is easier to use than trying to tell Windows
where to get a driver file from.

Hope this helps others!

--
Brent Brown, Electronic Design Solutions
16 English Street, Hamilton 2001, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell/txt: 027 433 4069
eMail:  spamBeGonebrent.brownspamBeGonespamclear.net.nz


2005\09\28@080541 by olin piclist

face picon face
Wouter van Ooijen wrote:
> I was the guy that sent a usb-to-serial converter to Olin and I am using
> exactly the same converter. So my particular hardware and Olin's
> particular hardware are the same!

And are you able to make the converter transmit over RS-232 without the
handshaking lines driven or looped back?  If so, please describe the details
of your setup.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\28@081708 by Wouter van Ooijen

face picon face
> To cut a long story short I got hold of the latest Windows
> driver from
> http://tech.prolific.com.tw (NOT from Bafo!)
> Hope this helps others!

Thanks, I'll certainly mention it on my website.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\28@082314 by olin piclist

face picon face
Brent Brown wrote:
> To cut a long story short I got hold of the latest Windows driver from
> http://tech.prolific.com.tw (NOT from Bafo!). Driver is v2.0.2.1 and
> the file is named wd_pl2303h-hx-x_v20019v2021.zip. It seems to have
> successfully fixed all my problems!

Very interesting.  I'll try that when I get back to my real office this
afternoon.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\28@082355 by Aza D. Oberman

flavicon
face
<The thrust of the thread about RS-232 communications is:>
> > >...it is unlikely that requiring the handshake (if that is
> > > the case) is just caused by the hardware:
> > >  I use the same hardware without any hardware
> > > handshake (just 3  wires, GND, TxD, TxD).

<I cautioned against inferring that is the general case>
> > So your particular hardware does not insist on hardware handshaking.

<Wouter van Ooijen commented>
> I was the guy that sent a usb-to-serial converter to Olin and I am using
> exactly the same converter. So my particular hardware and Olin's
> particular hardware are the same!

Alas, I dumped my copy of the thread Wouter, so let me apologize for
covering old ground.  Are you saying that when you use the converter
handshaking is required but when Olin used the same converter he did not
need handshaking?

We've had all kinds of problems with usb-to-serial converters too so I'm
quite interested in Olin's observation that there may be different drivers
running around.

Clearly, either the communications software or the hardware can determine if
hardware handshaking is required.

My hardware point slips off the thread a bit and goes to the broader
question of DCE (modem like) device design.  I'm simply saying that it is
not unheard of  for computer (DTE) RS-232 interface hardware (UARTs) to
require handshaking regardless of the communications software using the
interface.  Two examples of this behavior are the Rockwell R6551AP UART chip
and probably the original IBM PC RS-232 adapter card as specified in the
legendary "Blue Book."

It would seem then that a prudent design of a DCE device should include
strapping pins 4&6 on a 9pin (6&20 on a 25pin) to allow even DTE devices
that require rudimentary hardware handshaking to send.

My apologies if my comments on the broader question caused confusion.

Aza D. Oberman

2005\09\28@083706 by Wouter van Ooijen

face picon face
> And are you able to make the converter transmit over RS-232
> without the handshaking lines driven or looped back?

yes

> If so, please describe the details of your setup.

The hardware isn't very interesting: Three wires from the converter to
the MAX232 (GND, RxD, TxD).

I use XP, no SP's. The device manager says:

  ATEN USB to serial bridge (COM7)
     driver: ATEN 6-7-2003 2.0.0.18

  Prolific USB-to-serial comm port (COM9)
     driver: prolific 9-4-2002 1.5.0.0

The COM7 is the one I use most often, and is the same hardware I sent
you (as far as I can tell). COM9 is an older version, it works just as
well. IIRC one of the drivers is a leftover from a previous converter
which I bought in a local shop, which worked OK too. I *never* had
problems with any converter, so I never payed much attention to which
driver I used. I am quite interested to see this solved!

Olin (or anyone else who has problems on XP), can you tell me exactly
which driver you installed which caused problems, so I can try to
reproduce the problem?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\28@090610 by Wouter van Ooijen

face picon face
> Are you saying that when you use the converter
> handshaking is required but when Olin used the same converter
> he did not need handshaking?

The other way round. I use that converter, without handshake
connections, and I have no trouble. Olin has trouble, which is at least
partly cured by adding the hardware handshake loopbacks.

But my bet is that the PC drivers make the difference.

> Clearly, either the communications software or the hardware
> can determine if hardware handshaking is required.

hardware == usb-serial converter? it should not decide that HW handshake
is needed!
hardware == the connected serial gadget? Neither my Wisp628 nor Olin's
EsayProg uses hardware handshake.

{Quote hidden}

Maybe there is hardware around that requires this, but the original PC
hardware used a UART that was fully programmable by software, the
handshake lines were simply bits in the interface that the software
could use as it saw fit. Definitely no requirement by the hardware to
use these lines!

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\28@094101 by olin piclist

face picon face
Wouter van Ooijen wrote:
{Quote hidden}

I will report the details of my Windows 2000 setup when I get back to my
Embed office this afternoon.  I'll also try installing the new drivers
someone else mentioned.  It would be nice to get to the bottom of this, and
even nicer if we can find a USB to serial adapter that can be counted on to
work with just RX, TX, and GND reliably.

The specifics of my XP setup will have to wait until I get home sometime
much later tonight.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\28@101120 by Wouter van Ooijen

face picon face
> I will report the details of my Windows 2000 setup when I get
> back to my Embed office this afternoon.

That won't help me (I have no win2k around), but maybe others.

> The specifics of my XP setup will have to wait until I get
> home sometime much later tonight.

That could be usefull.

USB specialists around: what *is* important here?
- OS
- driver
- usb/serial hardware

anything else?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\28@105350 by William Chops Westfield

face picon face
On Sep 28, 2005, at 7:11 AM, Wouter van Ooijen wrote:

> USB specialists around: what *is* important here?
> - OS
> - driver
> - usb/serial hardware
>
> anything else?
>
USB/serial hardware firmware version.  I don't even know if this
is accessible, but it might be an issue in some cases...

BillW

2005\09\28@153616 by Stef Mientki

flavicon
face


Wouter van Ooijen wrote:

{Quote hidden}

probably ...
We solved our USB problem for 80%.
The frequency of the crashes on the "bad PC" is reduced from 5 minutes
to 24 hours.
What we did: disabled a small analog part that wasn't in use and doing
almost nothing ?????????
There could be a problem on the PCB, but we see no way how it can
interfere with
either the RS232-part, nor the USB-part of the FTDI device.
In the next weeks (when I hope to have some more time) I'm still going
to try the FTDI suggestions.

Another way I was thinking of, are software USB-sniffers.
Has anyone experience with these ??
Are they usefull for logging normal running USB communication ??





Stef

>  
>



2005\09\28@164503 by Brent Brown

picon face
> > And are you able to make the converter transmit over RS-232
> > without the handshaking lines driven or looped back?
>
> yes
>
> The hardware isn't very interesting: Three wires from the converter to
> the MAX232 (GND, RxD, TxD).
>
>    Prolific USB-to-serial comm port (COM9)
>       driver: prolific 9-4-2002 1.5.0.0

FWIW, this is the exact same driver that worked on my laptop with one
converter I've had afor a year or so but not with a brand new one straight out
of the box (same brand, model...Bafo BF-810). I also use just a 3 wire serial
connection with no loopbacks. Running Win XP Pro with all latest service
packs & updates. And AFAICT the v2.0.2.1 driver fixes it.

--
Brent Brown, Electronic Design Solutions
16 English Street, Hamilton 2001, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell/txt: 027 433 4069
eMail:  TakeThisOuTbrent.brownEraseMEspamspam_OUTclear.net.nz


2005\09\28@171813 by Aza D. Oberman

flavicon
face
< Wouter van Ooijen took issue with my contention regarding the IBM PC when
I said>

> Two examples of this behavior [requiring control lines] are the
> Rockwell R6551AP UART chip and probably the original
>  IBM PC RS-232 adapter card as specified in the legendary
> "Blue Book."

I disagree, Wouter.  And although regrettably I suffer your learned
disapprobation, I enjoy the company of one of the more prolific authors
about serial communications at the time.

As though anticipating this thread, Joe Campbell in "C Programmer's Guide to
Serial Communications", Howard Sams & Co., 1978, remarks as follows on page
144:

"The IBM PC family represents the height of folly in this regard, where the
system -- DOS and ROM BIOS --  does not contain simple unadorned functions
to perform character-by-character serial I/O.  For every byte the serial
output system perfunctorily asserts the Request To Send (pin 4 [25 pin]) and
Data Terminal Ready (pin 20 [25 pin]) outputs, then tests both the Clear To
Send (pin 5 [25 pin]) and Data Set Ready (pin 6 [25 pin]) inputs are enabled
before sending this byte.  Likewise, the serial input function refuses to
read data from the serial port unless the Data Set Ready is asserted.  If
these constraints are not bad enough, the system provides no mechanism to
defeat them."

For those of us who dropped down to port calls Joe Campbell's observation
proved to be only the tip of the iceberg.  The serial port adapter was
simply not able to kick the habit of requiring handshaking.  If I recall
correctly (without checking) the problem was with how IBM chose to strap the
NS1645 UART, thus making one of the best arguments to quit engineering and
go to law school<g>.

Aza D. Oberman

2005\09\28@180237 by Rob Hamerling

flavicon
face


Aza D. Oberman wrote:

> As though anticipating this thread, Joe Campbell in "C Programmer's Guide to
> Serial Communications", Howard Sams & Co., 1978, remarks as follows on page
> 144:
>
> "The IBM PC family represents the height of folly in this regard, where the
> system -- DOS and ROM BIOS --  does not contain simple unadorned functions
> to perform character-by-character serial I/O.  

Interesting to hear a comment about the IBM PC in a book three years
before the PC appeared....
Anyway: the BIOS routines for the serial port far from fancy. But when
you accessed the hardware directly (bypassing BIOS) the Async Adapter
performed much much better than via BIOS. I never experienced the
'required' handshaking in this way.

Regards, Rob.

--
Rob Hamerling, Vianen, NL phone +31-347-322822
homepage: http://www.robh.nl/

2005\09\29@020916 by Wouter van Ooijen

face picon face
> "The IBM PC family represents the height of folly in this
> [snip mandatory hadshake story]

Strange, I recall programming such serial ports, using just direct
hardware access, and using 3 wires only. Either that was not on the type
of PC you describe (in my memory it was, there was little else available
at that moment), or your text confuses using the bios routines with
dirtectly using the hardware.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu



2005\09\29@024724 by William Chops Westfield

face picon face
On Sep 28, 2005, at 11:09 PM, Wouter van Ooijen wrote:

>> "The IBM PC family represents the height of folly in this
>> [snip mandatory hadshake story]
>
> Strange, I recall programming such serial ports, using just direct
> hardware access, and using 3 wires only. Either that was not on the
> type
> of PC you describe (in my memory it was, there was little else
> available
> at that moment), or your text confuses using the bios routines with
> dirtectly using the hardware.
>
I wrote some comm software for the original IBMPC and original serial
port (based on an intel 8250; the NS16450 was ... later.)  The PC bios
was awful.  basica had surprisingly good async support, though.  My
strategy was to to use basica as initialization and for file transfer
and etc, while using ASM for the stuff that had to be fast, like the
H19 terminal emulator and MODEM2 checksum function.  Alas, my code has
no comments about wiring; I do recall the port being unusually picky
about handshaking lines, but I DON'T recall whether that turned out
to be a firmware or hardware issue.

I do know for sure that the first time cisco shipped a router that
didn't have the "flowcontrol" signals looped back (or strapped TRUE),
we got complains of inability to talk to "PCs" very quickly, and had
to do an ECO.  Again, I don't remember whether that turned out to be
a software or hardware issue (it didn't matter.  you pretty much can't
tell a customer "your comm software is broken; get something
different.")

BillW

2005\09\29@090952 by Aza D. Oberman

flavicon
face
<I noted the following citation when contesting Wouter's assertion that
three wire RS-232 was possible in the original IBM PC>

> > As though anticipating this thread, Joe Campbell in
> > "C Programmer's Guide to  Serial Communications",
> > Howard Sams & Co., 1978, remarks as follows on
> > page 144:

<Rob Hamerling quips>
> Interesting to hear a comment about the IBM PC in a book three years
> before the PC appeared....

Yeah.  The correct date is 1987.

Great to hear from someone who's never made a typographic error on a
tedious, but optional, citation<g>

<In a more serious vein Rob continues>
> Anyway: the [original IBM PC] BIOS routines for the serial port [are]
> far from fancy. But when  you accessed the hardware directly
> (bypassing BIOS) the Async Adapter performed much much better
> than via BIOS. I never experienced the 'required' handshaking in
> this way.

I agree that the only way to get performance was (and still is) by direct
access and I envy your good fortune to not get stuck without a way to
disable all forms of hardware handshaking, Rob.

Unfortunately I wasn't as lucky.  We spent quite a bit of time before
figuring out that the NS1645 UART on the original IBM serial and
serial/parallel port adapter card was wired in such a way that one could not
use software to turn off all the handshaking requirements for reading
RS-232.  It *had* to "see" a device out there before it would receive
anything.

Compaq and most of the clones followed the written IBM specifications and
not the IBM practice. The clones permitted software to disable all
handshaking requirements.  IBM later corrected the oversight, but not before
we'd spent quite a bit of effort (all of our communications were low level
because we were wringing reliable 38.4 Kbaud communications out of that
miserable box) on the original adapters.

I imagine the Rockwell UART I cited earlier also demands some control
signals simply because their target market was in detachable external RS-232
devices that could go wonky if inopportunely unplugged and RX floated.  But
that's only speculation on my part.

Three wire (RX, TX, & Common) RS-232 is certainly convenient, but I still
question if it is a good practice to not at least strap the handshaking
lines when designing a RS-232 device. [Ironically, I believe the "RS" in
"RS-232" stands for "Recommended Standard" -- makes one wonder what the
official one  is...]

Aza D. Oberman





2005\09\29@094326 by Wouter van Ooijen

face picon face
> Unfortunately I wasn't as lucky.  We spent quite a bit of time before
> figuring out that the NS1645 UART on the original IBM serial and

Maybe you had an eralier PC than I ever used, IIRC my documentation
referred to a 8250. Is there any NS1645 datasheet on the web? Google did
not give me one.

> [Ironically, I believe
> the "RS" in
> "RS-232" stands for "Recommended Standard" -- makes one
> wonder what the
> official one  is...]

Those documents at least pretend to be a step towards a standard, the
internet depends on just 'Requests For Comment' :)

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\29@111512 by Rob Hamerling

flavicon
face

Hello Aza,

Aza D. Oberman wrote:
> . . . . . . .   We spent quite a bit of time before
> figuring out that the NS1645 UART on the original IBM serial and
> serial/parallel port adapter card was wired in such a way that one could not
> use software to turn off all the handshaking requirements for reading
> RS-232.  It *had* to "see" a device out there before it would receive
> anything.

As far as I know the IBM-PC and later the PC-XT were standard delivered
with a monochrome display/printer adapter. An 'Asynchronous
Communications Adapter' was an option.  This adapter was equiped with an
NS8250 and a 25-pin subD connector (IBM Tech Ref April 1984).

The combined asyn/printer adapter came with the PC/AT, and this adapter
was equiped with an NS16450 and a 9-pin subD connector (IBM Tech Ref for
the PC/AT, September 1985). This is probably the one you are referring
to?  I have still such an adapter in use on an old Pentium I use for
experiments. Maybe it is a later version, but I have never experienced
problems with flow control lines.


> Three wire (RX, TX, & Common) RS-232 is certainly convenient, but I still
> question if it is a good practice to not at least strap the handshaking
> lines when designing a RS-232 device.

I agree. When I build a PIC-based device with UART I either implement
CTS/RTS handshaking or use loop-backs with a current limiting resistor.

[Ironically, I believe the "RS" in
> "RS-232" stands for "Recommended Standard" -- makes one wonder what the
> official one  is...]

Isn't it an EIA or IEEE standard today?

Regards, Rob.

--
Rob Hamerling, Vianen, NL phone +31-347-322822
homepage: http://www.robh.nl/

2005\09\29@113138 by Aza D. Oberman

flavicon
face
< Wouter van Ooijen asks>
> Maybe you had an eralier PC than I ever used, IIRC my documentation
> referred to a 8250. Is there any NS1645 datasheet on the web? Google did
> not give me one.

Good question, Wouter.  I cheated when I wrote the email because I didn't
actually remember the chip number.  I went to the IBM "Blue Book" (IBM
publication #6137806) and page 29 of the August 31, 1984, "Personal Computer
AT Serial/Parallel Adapter," and transcribed the chip number -- NS16450.

Your question sent me back to that reference to confirm the chip number.
While there I went to another undated section entitled "IBM Asynchronous
Communications Adapter" (IBM publication #6361501).   On page 33 they show
the 8250 chip.

Quite frankly, I don't know which chip we had our adventures with.  I'd
assummed, incorrectly it appears, that the same UART was used on both
adapters.  Clearly, that's not so and I can't recall with authority which
chip & adapter drove us nuts (I thought we'd had trouble with both).  I do
know that we got something like a half dozen of the first PCs and the "Blue
Book" in our library directly from IBM for that project.

However, in the final analysis I don't think it matters if handshaking is
demanded by the hardware or software, it's not a good idea to build a RS-232
serial port device for widespread use without some accommodation for
hardware handshaking.  Knowing that it's someone else's fault is not much
solace for a frustrated customer.

Aza D. Oberman

2005\09\29@131909 by Wouter van Ooijen

face picon face
> Good question, Wouter.  I cheated when I wrote the email
> because I didn't actually remember the chip number.

Our freelance librarian Rob will flog you with the correct references :)

> However, in the final analysis I don't think it matters if
> handshaking is
> demanded by the hardware or software, it's not a good idea to
> build a RS-232
> serial port device for widespread use without some accommodation for
> hardware handshaking.

That's where your opinions still differ, especially for a serial gadget
were I also provide the PC software. But I guess the dead horse won't
move at all, no matter how hard we beat it.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\09\29@153617 by Howard Winter

face
flavicon
picon face
Wouter,

On Wed, 28 Sep 2005 16:11:20 +0200, Wouter van Ooijen wrote:

> USB specialists around: what *is* important here?
> - OS
> - driver
> - usb/serial hardware
>
> anything else?

Well I'm no expert, but sitting back and watching this thread, the evidence seems to say that the driver is
the culprit.  The hardware (USB-Serial adaptor) is the same, UARTS don't decide anything, they just do what
they're told.  Win2k and XP are both NT-derived operating systems and underneath the glitz are very similar,
especially for something as mature as serial ports.  The drivers, however, can come from several places (the
chip manufacturer, the device manufacturer, third parties, Microsoft) and I think the answer is to find a
driver that works and test it on different machines/environments.  I  believe that will give you the answer.  

I think it's likely that someone, somewhere, when writing a driver made the decision not to send data out
unless there was something there to send it to, and that's wrong in the situation you are in, so using drivers
which don't do that would be the solution.  You appear to have one - send a copy to Olin, perhaps?

Cheers,


Howard Winter
St.Albans, England


2005\09\29@184358 by Dave Lag

picon face
This got me interested enough to drag out my Tech Ref copy,
 says: Aug 1981 first edition (would not have expected first edition in
'81?) refers to INS8250...

so we all predate the PC, the AppleII, the IMSA....
ho hum...
D

Rob Hamerling wrote:
{Quote hidden}

2005\09\29@195838 by Chen Xiao Fan

face
flavicon
face
Yes my university textbook in Chinese (published in 1991)
says INS8250.

I am not an old timer but I still used Apple II in my
secondary school and PC XTs/ATs in the universities.
The AST 286 PCs were the best PCs in the department.
They have 1M RAM, 40M HDD and VGA screen and runs
DOS 3.3...

In 1997, I was studying in a lab where two 286 PCs were
happily serving as email teminals (Pine for VAX OpenVMS).
Later Microsoft Exchange came and it was replaced by
Pentium 133 PCs. Still the two 286 PCs were taken over
by another department which had one program which would
not run on any PCs faster than 25MHz. They were so
happy at the time when they saw the 286...

And in year 2005, half of testing stations here is still
running DOS...

What is IMSA?

Regards,
Xiaofan

-----Original Message-----
From: Dave Lag
Sent: Friday, September 30, 2005 6:44 AM
To: Microcontroller discussion list - Public.
Subject: Re: [EE] USB to RS232 converter

This got me interested enough to drag out my Tech Ref copy,
 says: Aug 1981 first edition (would not have expected first
edition in '81?) refers to INS8250...

so we all predate the PC, the AppleII, the IMSA....
ho hum...
D

2005\09\29@212856 by David Van Horn

picon face
> What is IMSA?

IMSAI  http://www.imsai.net/

How I lusted after that thing!




2005\09\30@040358 by Alan B. Pearce

face picon face
>The hardware (USB-Serial adaptor) is the same

I am not totally sure this is the case - I suspect some of the problem may
well be revision levels of the chip itself, and changes made to the driver
to suit. Think in terms of the errata we become aware of for PICs, and then
think in terms of not being aware of similar documents for the Prolific (or
other) chips, and what errata information they may contain.

More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...