Discussion- Multi-purpose Onboard Computers

A couple of ramblings ...

My own thinking is towards a simple RS232 line dasiy-chained from the GPS to the computer - via a PIC-based cockpit display - that way the info gets to either/both as required - and no real 'interfacing' is involved. For systems monitoring etc. a 4800 baud serial link is plenty fast enough.

Thought I'd stumbled across that crock of gold at the end of the rainbow yesterday when I came across Hibernate 1.2, a freeware program from 2003 which is intended to give Suspend-to-Disk functionality to WIN9X desktops - only it don't work. Bummer. No wonder I hadn't heard of it before now.

Some good news for a change - apparently laptop hard drives survive well in hostile environments, according to the young bucks over at the MP3 car computer forum, and they tend to treat their gear pretty harshly. They're quoting typical shock tolerance figures of 55G's for OPERATING and 350G's for NON-OPERATING. Pretty impressive. In contrast, vibration tolerance is rated at only 0.67G's OPERATING and 1.04G's NON-OPERATING. I suppose this is due to the potential risk of resonance building-up (?).

The consensus seems to be that soft-mounting the hard drive may not be a good idea, but if done, then the mounts need to be well damped to prevent oscillation (ringing).

Nice to know that there's at least one component we don't have to worry about.

Colin

 

2.5" hard drive would work ok, but for longevity you're not going to beat a solid-state drive. It's a question of cash.

As soon as you start accessing ports, WinXP (well all newer windowses) are frankly a bit of a pain. the Win9x series is easier, but notoriously unstable and insecure. My preference would be to run Linux, which will handle almost any hardware, and is very well documented. Now, I'm not saying don't use a version of Windows, I just think it will be a lot of hassle.

As I am writing the FyCom (Free Yacht Communicatio) API and Protocol (Protocol should be available for review on Tuesday) I am trying to avoid tying it to one system. The basic command-structure should be usable across serial, network and CANBus. Hopefully this will simplify everyone's workload, when getting equipment talking to each other.

For SAE's job if he's just taking Lat/Long data, there's no need for FyCom. However, for output to other equipment it should be an easy standard to implement. The intention is to make Fycom an open, flexible communications protocol.

Cheers,

Tim B.

 

Hard Drives maybe not so hard a problem..

Those guys have a lot of practical experience, and their consensus seems to be that 2.5 and especially 1.8" notebook drives do fine.. Now that 1.8" drives at 40 to 60 Gb are available for $100 they look real good. But there are starting to be 16 Gb flash drives and eventually they may be a good choice, especially for a high-speed slammer boat. But the power-on-hours on that boat will be a lot less than any cruising type boat that never really slams, and needs lots of hours.

I will collect my info from the China High Tech Show in a couple of days and do another post.

 

Communications / Protocols

Tim, can you describe this project in more detail? Did I miss an earlier post? I have lots of protocol thoughts, but plan on staying with NEMA0183 for input sensors/devices for now due to the availability of reasonably-priced instruments.

Colin, I agree that RS232 and serial at 4800 baud is enough for many onboard functions. I'd like to stick with NMEA0183 compatible stuff, though, even if that means defining new NMEA sentences. My feeling at the moment is to do NMEA to USB interfaces in and out. The expandability and the easy availability of 6 native USB ports on most current system boards, and easy expandability with USB hubs, and the available WIN-XP drivers for serial-to-USB devices seems like a good set of reasons. Opinions??

I see quite a few PIC-based "thingies", both for input stuff like engine sensors and output stuff like power controls and status displays. I'd like a series of NMEA0183 devices that all have an NMEA fan-in of 4 to 8 so you can easily set up a tree of NMEA devices onboard. But it would be really nice to have a cooperating protocol that was effective for bi-directional communications.

But PLEASE, PLEASE don't make another MAP (General Motors "Manufacturing Automation Protocol" from the 80's and 90's ) where every node is a "Peer" and nobody is in charge! I sweat way too many hours over that Monster before it finally died of its own weight)..

This thread is getting big and maybe we should spin out a protocol / communications thread some time.. ???

 

Hi, Bagger,

Nice to hear from you again. I apologize if I created the impression that I thought a computer won't work outside an air-conditioned environment in the tropics. I was focusing on heat because I had just witnessed a fried HDD. The point of my experience is that however you do it, provision for cooling is critical. Blockage of cooling ports was what cooked my drive. I found a file in BIOS that stated my drive failed at a temperature of 113 C. Any reasonable cooling system, whether static (fins) or dynamic (fan), so long as it is functioning, should have no trouble maintaining enclosure temperatures far below that. You're right about water; even the warmest tropical harbor water will rarely exceed 35 C. Air conditioning condensers operate with some efficiency in ambient air temps up to 45 C. The key is to recognize that cooling by some means is critical, and equally important, that the cooling system needs to be checked carefully to verify that it is working. My HDD cooked because I was not careful enough in seeking out tiny passages that had become blocked with dust. Operator error, without question.

 

Don't worry, Charm, I really don't know much about computers, but I completely agree, that cooling is very important. But in this context there won't even be dust, nor (I hope) will people be editing movies or using photoshop and the like (?). You don't need some dual 3.16GHz processor set-up. And for what it's worth, the power drain on that type of processor sucks big time, imo.


When that is said though, you have a very good point about the HDD. And makes me think linux (or OS X, if you prefer – I do), because it uses the harddisk less. But further, it wouldn't be that hard connecting the hdd-case to the outer-case, or "glue on" (with that heat sink paste) a heat sink.

The computer I used when I was out in the heat was an alumium computer - was yours plastics (the one with the fried hdd)?

And in this context, I would mount the hdd externaly, so it wouldn't get any heat from the CPU, RAM etc., and thereby it would have a bigger surface to cool down from.

Also, I'm thinking, that one could simply look at harddisks (still 2.5" preferably) that are used in laptops. They use less power, thus producing less heat, and there is less mass, which translates to more surface per x mass (if I got that right). Further, how about speeds of the HDD.

Take a look at these harddrives (In essense, it's just the case that makes these "rugged", but notice that they're closed off):

http://www.lacie.com/products/product.htm?pid=10722

https://www.olixir.com/products.php


And this one claims to be able to operate (yes, operate) at temperatures ranging from -30 degrees celcius (-22F) to a whopping 85 degrees celcius (185F)!

http://news.softpedia.com/news/Rugged-Hard-Disk-Drives-From-Fujitsu-49386.shtml

Mind you, when they say that it'll work at those degrees, it'll propably work at even higher temps if not used to its utmost capabilities.

And this one – for server use for the U.S. Army and to the U.S. Marine Corps:

http://www.spacewar.com/news/miltech-05zzzzl.html

This place will build one for you to spec:

http://www.mt-optech.com/rugged_hard_drive.html


P.S. One would certainly want a HDD with a "drop-sensoring" thingamajig to protect the data (belts and suspenders etc).

 

Bagger,

My dead drive was a 3.5" EIDE internal in a desktop tower. Metal frame, but the enclosure panels are plastic (Hmm, good insulator, keeps heat inside, so the computer components won't get cold ).

My external HDD is in an aluminum housing. The manual mentions that it's normal for the housing surface to be hot, as it's radiating heat away.

You're right on the newer notebook HDDs. Minimal power draw and heat generation are important design criteria now. Selection of suitable components is getting larger (Yay!). Terry should have the latest info now; he's near the source of most of the new components.

 

Very impressive HDDs. I love the article's opening line: "Listening to music at 100 mph might kill you, ... but at least your player's hard drive will survive."

 

Hard Driving Over Water: F1 Racing

The Fujitsu drive is available now at $200 and seems like a good choice for a fast planing boat that gets some good slams and vibration. The most vibration I've ever felt was when I ran a little 3-point hydro with a 20 HP Merc in about 1960. A bigger displacement boat will never get those high-frequency movements, I think.

AND in another bizarre development, I've just been asked to shoot photos this coming weekend of the Shenzhen Grand Prix F1 Powerboat races, for the Shenzhen Daily. Poor me, I get to go anywhere I want, talk to the drivers and (more importantly to me) the mechanics, and possibly fondle the boats. I'm beginning to really like China.

Unfortunately I know almost nothing of detail about Formula 1 boats. They do not run in the US, I guess. Mostly the Persian Gulf, Europe and Asia, maybe Oz?

These are all Merc V6's 400 HP out of 2500 CC, 130 MPH and Zero to 100 MPH in 4 seconds.. Couple (NOT stolen: I am now blessed by F1 to use their images!) photos here:
AtTheDock:

OnTheWater:


And I get to be in the herd, here:


I'm doing some fast reads, but if anyone knows a bit about F1, or if you have any ideas of what would be a cool 'angle' (photographic or editorial) on this kind of racing and these boats, please drop me a note at: terry@terryking.us

I'd like to appear like I'm not too stupid about this kind of boat.

... I'll ask what kind of hard drives they're running !

 

Broadly agree.

The reason I suggested a simple RS232 link is that if set to 4800 baud, then it becomes NMEA-compatible.
One obvious limitation of NMEA is that it's a one-way protocol, and as such it's not possible for the PC to upload waypoint data to a GPS or send out any receiver-specific commands. However, by using a bi-directional RS232 link this can be achieved, albeit with the manual pressing of a button or two.

An RS232 link has limitations, the most significant being that it is a one-to-one connection, and therefore not useable as a bi-directional *bus*. However, very few devices on a boat need to be written too: GPS and automatic steering mechanisms are the only ones which readily come to mind (perhaps this is why Autohelm developed their SeaTalk protocol ?).
The simplest way around this limitation is to construct a PIC-based GPS Repeater (such as Jon Ficks design), which could very easily be modified to function as an 'NMEA-hub', taking in data from numerous peripherals, filtering out essential data and forwarding that data to the computer's serial port as a single composite string - which is broadly in keeping with your outlined idea.

I really don't see why it has to be more complex than that. NMEA is a proven, well-documented and common industry-standard protocol, as too is RS232, with the data protocol from the latter (inverted data, start-data-stop format and standard baud rates) still being widely recognised, whereas admittedly the RS232 voltage level standards are not so widely respected. On the last point - I tend to employ 6N139 opto-isolators wherever serial devices might be at risk, and I'm surprised not to see other designers taking such a simple and inexpensive precaution.

And the playing of MP3's and DVD's etc ... ? Well - this is where it appears that my philosophy differs from others on this thread. I firmly believe that a navigation/ instrument monitoring system should be a discrete stand-alone unit, with as much built-in redundancy as is practical, and with some form of dedicated battery-backup - with entertainment systems being kept completely separate. WIN9X-based embedded controllers are remarkably stable: removing Internet Explorer and all the rest of the web-bloat is the answer to achieving stability.
I accept that this approach will not be attractive to those who seek a more sophisticated approach to electronic system design - but then - is sophistication *ever* desirable ?

With regard to your last post - and focussing on the engine control of power-boat engines for example - I can't ever see such control being handed over to a general-purpose computer, it's just too much of a demanding application. Monitoring perhaps, but any control functions would certainly warrant their own dedicated controller complete with hi-speed interrupts.

'best,
Colin

 

Data Link Options

Colin, we're thinking along the same lines here, for sure..

I don't see any compelling reason to go to a more complex system like multidrop RS485 or Modbus, or a true network like CAN / NMEA2000 for the kind of DIY and small boatbuilder applications I'm considering here. I understand why big boatbuilders and Marine Electronics players want and need NMEA2000, but that's a different scenario.

Also, I want to figure out an approach here where many of us with existing boats, refits, and custom builders can put together a working system which includes the many NMEA0183 Instruments out there from CruzPro, Clipper, etc., and including those already onboard and feeding existing cockpit displays.

I don't see any important limitation to the NMEA0183 concept that any segment of a network has only one talker and multiple listeners. NMEA specs that a Talker should be able to drive 4 Listeners, but that's very easily exceeded, especially with sensitive optically-isolated receivers like a 6N139 with a few other components.

Multiple Instruments, Sensors, Engine monitors etc that are Talkers can be combined (multiplexed) into one NMEA0183 signal going to the computer. At 4800 baud, such a path could saturate at about 10 talkers. BUT there is no reason why the Combiner needs to talk at only 4800 baud, and in fact some combiners use a USB output to the computer at high speed. I like that approach.

So that ends up looking like a 'tree' of NEMA Talkers combining together and going IN to the computer. And devices can be added or removed from that tree without affecting any other devices.

Now, Outputs: As you said, there are very few NMEA devices that need to be talked TO: mainly GPS for upload of waypoints and routes etc, and an Autopilot. I would dedicate one USB/Serial port to this NAV stuff: GPS and probably Depth IN, and Autopilot out. Keep this separated from all the other monitoring and control 'Dashboard' stuff.

Now, if someone builds the hardware and software to do a more complex boat with "Onboard Systems Outputs" as shown on my diagram (far below, now).. these are all Listeners. They may as well be NMEA compatible and listen to one or 2 NMEA outputs. These do not need to be High Speed. Like you said, if there is some high-speed control system it needs it's own dedicated microcomputer.

Hey, Where Is Jon Fick?? He and I worked together at IBM in Vermont. I tried to email a couple of addresses I had and got no response. Anybody hear from him? Jon, please Google This

I looked at Jon's PIC stuff a while back. I will be writing some NMEA stuff in PIC soon (I have a prototype of a part of a combiner done..). I am sure there is plenty of $8 compute power in the newer PIC chips for anything we'd want to do.

And we need to see how Tim's work may plug into all this...

Oh.. Sorry about the F1 over-enthusiasm, it can go elsewhere. But there IS a thing about Real Fast Boats, even if I'll never own one...

 

Hi Terry - I spoke with Jon back in July/August time, using the address jon@NOSPAMVermontFicks.org (remove the NOSPAM, obviously), and I think there's also a contact link on his webpage.

I still can't come to term with PIC's - how Microchip have cobled together a complete controller, the like of which would have taken half-a-dozen fully populated boards a few years back, is beyond me. And all for a handful of millamps when running, and just a couple of nanoamps when asleep. Truly amazing.

And as for the latest DsPICs - I took a quick look at those, got m'self a headache, and decided to leave 'em alone. That be technology for the youngsters ....

Colin

 

Hi again - don't worry about the lack of response to your F1-questions. If the other guys are anywhere like me – even to the slightest degree - there's a much defined line to where their knowledge about that type of boat stops. And begin, for that matter.

Anyway, for other reasons, I happened to stumble upon an interesting site pertaining to the threads topic:



He makes an interesting computer, with some nity details:

Here:
http://www.archer-marine.com/

And here's the products:
http://archer-marine.com/album_0.htm

 

Found Him.. thanks!

Jon is still designing IC's at IBM, and sailing his boat on Lake Champlain in the Summer. I'll be talking with him about the NMEA stuff.

The PIC Microcomputers are amazing for sure. This youngster started in Broadcast in 1956 in a radio station that was 100% vacuum tubes. Now a million transistors is a couple bucks.

Fortunately, the DsPICs also have a simpler cousin, the 24F series which is very much like the earlier PICs and much of the older code will work. But they're FAST.. like 20 million instructions per second.. And like you said they have SO many built-in peripherals and connections. At the risk of being too tekkie, here's the block diagram of these animals:

For onboard purposes, the "A-D Analog to Digital Converters" can measure voltages and currents from temperature and pressure sensors, tank level senders, etc. and the "GPIO - General Purpose I-O" pins can read switches and alarms, and turn things on and off. The Timers can, uh, time stuff and create pulses to control motors etc, the UARTs (serial ports) can talk to NMEA. etc. These newer PICs cost $8 to $12 instead of $4 or so. Probably not an issue..

Sorry for the digit-head part here, but some of us know nothing about piston-ring clearances, apparent wind, or F1 racers either. But we can use what others figure out, or learn it if we need to or want to.

 

F1 Powerboat Racer stuff is now HERE:

http://www.boatdesign.net/forums/showthread.php?t=19836