NMEA <-> RS232

[Stenn]

My experience is that it's not as "open and shut" as some of you guys have described it, when using a Serial-to-USB converter.

It's extremely dependent on the Windows driver that comes with the Serial-to-USB converter, and also probably dependent on the laptop being used.

I tried two different brands of serial-to-USB converters (presumably using two different drivers) with my Compaq R3000 laptop that only has USB inputs.....and found the converter's driver would blue-screen the laptop (running XP) until I found the "sweet spot" the driver liked.....if I set the baud-rate to 38400, it would never blue-screen, but if I slowed it down to what my DSC VHF wanted, the standard 4800 baud for standard NMEA 0183, then the laptop blue-screened pretty consistently.

I was feeding from the NMEA 0183 output of my Lowrance LMS-525c (which has native NMEA2000 GPS input), to a split connection, one going to the NMEA 0183 input of a Uniden UM525, and one going to an RS232 cable to feed a PC. Disconnecting the VHF, so I was not sharing the NMEA 0183 signal, did not help....if I ran the virtual com port (the Serial-to-USB converter) at different speeds, I'd get varying degrees of successful feed to the charting program on the laptop...

....."varying degrees of successful feed" meaning it would run for a few minutes, then blue-screen, or blue-screen almost immediately, or,....if set at 38400, was totally rock-solid. But since the Uniden will only accept 4800 baud (I had to call Uniden to find that out, they omitted that fact from their manual), I was stuck not being able to feed my GPS signal to my DSC VHF and the laptop at the same time.

I finally resorted to buying a used IBM Thinkpad A31 off eBay for $180, that still uses a standard DB9 serial connection, so I can forego using my primary USB-only laptop and the Serial-to-USB converter, and hopefully solve the problem.

.....sorry to say, I have not tested that yet, as I just got the laptop recently.....but I expect it will resolve the problem.

And, speaking of NMEA2K, in Pblais' comment "...Soon we will see the new NMEA 2000 stuff. This is a network basd interface and we can chuck serial ports but with Ray Marine doing their own thing it may not really be much of a standard...." ......

....on the contrary, from what I read, the only thing that's ending up being "nonstandard" between RayMarine SeaTalk2 or SeaTalk NG and other NMEA2K implementations is the actual cable connector on their equipment, but using an adapter cable, they, and others like SimRad's SimNet equipment, can be plugged right into a standard NMEA2K backbone (network), so evidentally the "language" itself between devices is standard NMEA2K, no hardware interface box necessary.

Stenn

[Pblais]

It's extremely dependent on the Windows driver that comes with the Serial-to-USB converter,

It's more dependant on the USB to serial adapter you buy.

[Stenn]

Originally Posted by Pblais

It's more dependant on the USB to serial adapter you buy.

Yes, it does depend on which one you buy, but not quite as much as you might think. These kinds of problems can be coming from any of a number of sources within the whole "system" that's communicating....there's lots of hardware AND software variables. But one of the common weak points is the software driver.

When Windows "Blue Screens" (crashes), if you know how to read what it displays on the screen, it will tell you the cause of the crash(if you have it set to do that, instead of rebooting automatically). In this case, my laptop was crashing directly due to the software driver running the Serial-to-USB converter, and the particular baud-rate I was setting it at. How well that driver is written, how well it's tested (such as across multiple baud-rates), is a huge factor in it's performance and stability.

And that's not even taking into account that you have the "other ends" to think about too, i.e. the USB interface inside each and every laptop, and the source of your NMEA 0183 data-stream...in my case the Lowrance LMS-525c. There is just as much variation there as inside the various Serial-to-USB converters you can buy. As an example, if your laptop, and my laptop use different USB buss-converter chips inside (and their associated software drivers), there is no guarantee the Serial-to-USB converter that works great on your laptop will work the same on mine.

I have spent my whole life so far, involved in the electronics and then computer and I.T. fields, and now back in the electronics manufacturing field.

With the state of the electronics "art" these days, the way many consumer electronics items like Serial-to-USB converters or wireless network cards, etc., are built, is that one or two, at the most 3 or 4, chip-design companies like Taiwan's "VIA" corporation, the U.S.'s "Texas Instruments," etc., will design and produce the primary I.C. (integrated circuit) that is the "guts" of that item, a Serial-to-USB converter chip, or a wireless transmit/receive chip for example, and then they sell just those competing chips to 20 or 30 manufacturers, whatever, that actually produce the finished product, with their own different names on them.

If you opened up a dozen different-branded Serial-to-USB converters, you'll probably find a maximum of 2 or 3 different converter chips on those dozen boards, and you'll probably notice that all the boards that share the same chip are just about identical in design. And, if you knew how to study the code of the Windows driver, you'll find all those boards that share the same chip often share an identical Windows driver, with only a line of code that identifies one brand from another being different in that manufacturer's driver, and sometimes not even that. Sometimes the driver actually shows up in Windows identifying the original chip-manufacturer on the board, not the manufacturer whose name is on the box or case. That's why, if you bought some piece of hardware and then lost the original driver-disk that came with it, AND that manufacturer went out of business, you can still often use the generic driver from the original chip manufacturer's website, and that hardware will work fine. You just take the cover off and see who made the chip.

That's because the original chip manufacturer usually produces a pretty much finished engineering design, including software drivers, and sells the rights to use the design and driver to anybody that wants to manufacture that product. The majority of manufacturers usually don't change a thing, although, in rare cases, the "retail" big-names actually have the personel and financial resources to go in and actually customize the circuit board and/or driver, and tweek it in little ways that might improve stability, or add features that are optional in the original engineering design from the chip manufacturer.

Beyond the original hardware design of the chip, there follows update upon update of the software driver, because the hardware, no matter how good it's silicon design is, is totally dependent on how good it's software driver is before it will function well in Windows, Linux, MAC, etc. (which is why Microsoft decided to get involved, and established their WHQL (Windows Hardware Quality Labs) certification process, to try and establish some degree of standardization and certainty that a given piece of hardware will actually work.

But it's not just about the converter or driver you're using either....

I also have to look at the BIOS/Firmware versions of the computing devices I'm using....i.e. whether my laptop has a BIOS update, and also whether my Lowrance LMS-525c has a Firmware update that might have corrected errors in it's NMEA 0183 output stream.

So, in the end, you might just have been lucky that the one Serial-to-USB converter you bought worked fine the first time, in your system taken as a whole (from source NMEA device to laptop). There can often be a number of interfering and interacting variables that could prevent that.

But it's definitely useful for us all to share which devices worked for us, so as to minimize "reinventing the wheel." That's what's supposed to differentiate us from the rest of the food-chain !

Stenn

[hellosailor]

Stenn-
" In this case, my laptop was crashing directly due to the software driver running the Serial-to-USB converter, and..."

These days, most folks are unaware that they are running Windows NT. Windows2000 is NT5.0, XP is NT5.1, Vista is NT6.0, but they are all NT, not the old Windows 3.x which Windows 9x really is.

And for WindowsNT, in any version, the hardware requirements are always more stringent. For Windows2000 there were only two USB-to-serial adapters that were actually submitted for certification and passed (IBM and Keyspan). Vista now has some two dozen brands and models certified, including big names like Radio Shack and Belkin who used to have many problems.

Certified, means they were submitted to a test program and awarded the right to use one of the Windows logo programs ("Made for..." etc.). I'll bet even money that your adapter which doesn't work, like MOST of the adapters on the market, is from a company which either didn't want to pay for certification--which is pricey--or a company which could not pass the certification tests.

Bottom line? Garbage In, Garbage Out.
No certification? Garbage! There's no excuse for blue screens from driver crashes, the certification programs are there to tell you "this gizmo has been extensively tested and found to play nicely with others in normal use" and it pretty much accomplishes that goal. All the certification lists (for the currently approved and older OSes) are on the Microsoft web sites, look for "HCL" Hardware Qualification List for your OS.

It is very similar to the UL or CSA approval on electrical appliances, or a USCG approval on marine equipment.

[Pblais]

If you refer to the above link for the pfranc USB / Serial device you'll find that that link takes you to a lot of discussion about this issue. I still stand by the idea that it's mostly hardware drivers. Chipsets and BIOS may be a factor but in those situations you are probably dead. I know the pfranc works with both USB 1.0 and 2.0 chipsets. They also have drivers for Vista and Linux as well as many others.

You will not get good support underWindows less than Win98 SE2.

Stenn. It's not the drivers because the drivers are chip based and a lot of the chips just don't work well enough for decent RS232 operation.

[Stenn]

Originally Posted by Pblais

If you refer to the above link for the pfranc USB / Serial device you'll find that that link takes you to a lot of discussion about this issue. I still stand by the idea that it's mostly hardware drivers. Chipsets and BIOS may be a factor but in those situations you are probably dead. I know the pfranc works with both USB 1.0 and 2.0 chipsets. They also have drivers for Vista and Linux as well as many others.

You will not get good support underWindows less than Win98 SE2.

Stenn. It's not the drivers because the drivers are chip based and a lot of the chips just don't work well enough for decent RS232 operation.

I'm confused by your answer Pblais.....in your first paragraph, you say "I still stand by the idea that it's mosly hardware drivers...."...but then to me you say "....It's not the drivers because the drivers are chip based....."

...so what "hardware drivers" are you talking about?

Also, that's not really accurate about "....you are probably dead" if the issue is a BIOS or firmware update that's needed. (I take it you mean the thing won't work at all, if a BIOS or firmware update is needed?) If you read through the ReadMe's of IBM Thinkpad or other BIOS's, you'll find many instances of BIOS tweaks found necessary only when using some specific hardware peripheral or another, and not just in go/no-go situations where it doesn't work at all, but often just to correct minor operating issues like what happens to the peripheral if the PC goes into Power Saving mode, etc....

Sailor.....I guess you skipped over my paragraph "Beyond the original hardware design....", where I talked about the Microsoft WHQL program and certified drivers? Although that is obviously the preferance...buying a converter that has a certified WHQL driver, unless you're standing in a "Brick and Mortar" computer store looking at the box, and plainly see it printed "Certified WHQL Driver," the average consumer has no way at all of knowing what driver they're getting. In my case, my favorite PC hardware source is NewEgg, on-line, but there, and other on-line sources, are at the mercy of what documentation the manufacturer gives them as to whether their driver is WHQL-certified or not, which is usually not mentioned.

.. I'll repeat that my converters worked perfectly well if I set the baud-rate to 38400 and fed GPS via NMEA to my laptop's charting program. It was only when I wanted to slow it down, and especially if slowed all the way down to 4800, that XP on my laptop didn't like talking to the 2 DIFFERENT BRANDs of converters I tried. I haven't opened them up to look, but maybe both of these use the same crappy chip, which would explain it.

Which leads me back to my previous closing statement about not reinventing the wheel by Trial and Error, by sharing information and experience about which devices have those certified drivers we know to work. I was not a member here when I bought those 2 converters, or I might have tried the one Pblaise is recommending.

Stenn

[Stenn]

Pblaise....I should have mentioned....I tried looking at that link you posted a while back, and it was a total confusing mess on their site.....I read through the whole page, and didn't find a thing about Serial-to-USB converters, just custom Garmin plugs.....

But then I clicked on their "Home" link, and found mention of their converter finally.....about a third of the way down the page....something about "New laptop, No Serial Port?"

Garmin GPS eTrex compatible cables, USB serial converter.

And, incidentally Pblaise, to underscore the "not so open and shut as you guys suggest..." here's a post right off of pFranc's website about their converter.....

".....Problem: Our converter would transfer lots of data then up and lockup!
Fix: Upgrading Windows from 2000 to XP on a Dell laptop needed BIOS & video driver upgrade too."

....so, again, it's often a systems-analysis that's needed to get to the bottom of a malfunction involving PC's and other processing devices....even to the point of updating a video driver!

Stenn

[Stenn]

Pblais, I ordered one ($30+$10 shipping), which will be here in 3 days....so I'll report back to you guys as soon as I can get down to the boat to test it....but that might not be for a week or two....it's REALLY cold here for the next week, and I brought my Lowrance unit home for safekeeping!

It will be interesting to see if, everything else being equal, i.e. same hardware setup, same Serial-to-USB converter,etc., this pFranc unit allows the laptop to work perfectly. But I must admit that since I last tried my converters feeding my Compaq R3000, I have updated the firmware of the Lowrance unit ...it was about 3 firmware versions out of date, right out of the box...but I didn't know that at the time I was testing! (for all I know, it might ALREADY be fixed ! ...so ideal test would be to duplicate failures with original converter first, then switch to this pFranc unit.

Stenn

[Pblais]

One tip with the pfranc and most any other USB / Serial adapters is do NOT start the GPS before you boot the PC Windows auto hardware detection often sees the GPS as a mouse. The NMEA signal starts moving the cursor around. Wait to power on the GPS and connect it to the serial port until after it boots. The good news is when it happens you will know right away. It takes a reboot to fix it however.

In general I think your Presario R3000 under WinXP should work well. My wife's has a slightly newer Presario and it works just fine under WinXP. As an O/S WinXP performs better with USB devices than the older versions of Windows.

I'll repeat that my converters worked perfectly well if I set the baud-rate to 38400 and fed GPS via NMEA to my laptop's charting program ...

Not all chips operate the same. Most modern chips will auto negotiate the transmission rate all on their own when you set the rate higher and it really is too high. It does not mean they actually operate at that speed it's just the speed they start with. The primary settings for NMEA would be 8 data bits, one stop bit and no parity. Those can not be negotiated. If as you say the device was working perfectly at 38400 then perhaps it was. It was not uncommon for the highest setting to mean auto negotiate.

I'll be anxious to here how the connector works.

[hellosailor]

Stenn-
"Sailor.....I guess you skipped over my paragraph "Beyond the original hardware design....", where I talked about the Microsoft WHQL program and certified drivers? "
Not intentionally but I did miss that.

"unless you're standing in a "Brick and Mortar" computer store looking at the box, and plainly see it printed "Certified WHQL Driver," the average consumer has no way at all of knowing what driver they're getting." That's probably why the folks who do pass, print the logo on the box. But the consumer does have another way to check, when they are ordering online they can easily pop over to the HCL online. For some products, the list is quite extensive. For others (like USB convertors) the list can be quite short.

Most odd that your converters would work at high speed and not low speed, that sounds more like the automatic negotiation process and the port settings themselves was somehow failing. But then again, I suppose a truly faithful serial port emulation WOULD fail that way, since set-up problems are all too often normal with real serial ports.