rwa@cs.athabascau.ca (Ross Alexander) (12/07/90)
cmm1@cunixa.cc.columbia.edu (Christopher M Mauritz) writes: >just pointing out that Atari has pulled the FCC-is-holding-us-up stunt >so many times in the past that maybe (just maybe) the fault lies at >Atari. How many times can you tell your teacher that your dog ate your >homework without finally taking the blame for not doing it? :-) >Besides, even the stuff that did pass must've been marginal. My old >1040ST used to create snow on any TV I was watching in the same room. >Of course, matters didn't get much better when I removed the RF shielding >to make room for my RAM expansion board. :-) I can add a few things here. I have an old 1040st, and a few peripherals for it (SCSI disk, printer, modem). I also have an AT&T 6300, a rather doddering PC-clone machine; it has an internal 30 meg drive. I am particularly concious of RFI [radio frequency interference] because, as a radio amateur I am constantly running both radios and computers in close proximity. For those who are interested, the application is amateur packet radio networking, a very interesting mix of two technical hobbies :-). But I digress. In short, the 1040st howls all up and down the bands. We're talking continuous S9+20dB heterodynes and growlies from baseband all the way up to 450 MHz (to be charitable, it does fall off somewhat in the UHF bands. On 145.01 MHz it's still pretty damned loud). How this thing ever got past the FCC I can't begin to imagine. I can only assume Part 15 subpart J class B means "leaks RF like a sieve". It took fairly heroic measures to get the intereference down to the point where I could recieve anything from other stations; lots of bypassing, coax cables, chokes, shielding, and physical separation were needed. Having done all that it works fine (good serial I/O, lotsa horsepower :-). Now in contrast, the pc6300 is _silent_. A little rumble on 75 metres perhaps, but otherwise dead quiet straight out of the box. A nice steel chassis, bolted together quite thoroughly; chokes on all cables running in or out. Of course it's an 8086 abomination but at least it isn't destroying all reception in a 30 metre radius! And it's only certified to class B. Most strange. The thing that really puts the cap on this, however, is the ds2100 that I am pounding this note in via. 16MHz risc box (mips engine), 12 Megs of ram, a whole raft of externally cabled SCSI devices, ethernet connections, a mouse, a 1000 x 900 display, oceans of transistors all frantically switching, right? And it has a class A certification, which is even more relaxed than the class B of machines intended for home use. And yet, and yet, it's almost as quiet on 2 metres (~ 146 MHz) as the pc6300. What exactly is Atari doing wrong, and why are their machines so noisy? Corner-shaving and cost cutting have limits, y'know. Yours for RF sanity, -- -- Ross Alexander rwa@cs.athabascau.ca (403) 675 6311 ve6pdq
jimomura@lsuc.on.ca (Jim Omura) (12/07/90)
In article <479@aupair.cs.athabascau.ca> rwa@cs.athabascau.ca (Ross Alexander) writes: >cmm1@cunixa.cc.columbia.edu (Christopher M Mauritz) writes: >>just pointing out that Atari has pulled the FCC-is-holding-us-up stunt >>so many times in the past that maybe (just maybe) the fault lies at >>Atari. How many times can you tell your teacher that your dog ate your ... > ... It took >fairly heroic measures to get the intereference down to the point >where I could recieve anything from other stations; lots of bypassing, >coax cables, chokes, shielding, and physical separation were needed. >Having done all that it works fine (good serial I/O, lotsa horsepower >:-). > > ... and why are their machines >so noisy? Corner-shaving and cost cutting have limits, y'know. > Well, yeah. You guys have pretty much state the reality. Atari has been aiming at cheap computing and that's your answer. They leave the stuff as marginal as they can get away with for RFI because anything you do about RFI costs money. But taking it from the top, the problem starts with IO interfacing. You can build a dirt cheap computer with no IO. A grounded cage does it. But as soon as you start adding IO ports you've got troubles. The Atari ST is rich with IO ports, and not all of the best possible quality. You've got the 2 MIDI ports (the basic spec of which looks noisy I might add -- not Atari's fault), an RS-232C, semi-SCSI, Centronics, 2 joystick and cartridge port. Take that "Centronics" port. The original Centronics used every other wire as a balance which could have helped cancelled "some" RFI (doesn't help that much anyway really). But the Atari doesn't do anything with them. Actually, I'd be happier them connected for data reception like the Amiga. But the only way to silence that one is a *very* expensive fully shielded cable. The cartridge port should be given a bolt on metal "protector" when its not in use. If it was a Japanese company, maybe a hinged door would have been included. Also, I'm not sure what the keyboard looks like if you take that apart. I have a feeling you'll find it an excellent source of interesting radio signals. :-) In passing, the most impressive *looking* computer I've seen RFI-wise was the TI-99/4a. I took this apart and it looked just beautiful! All this nifty shielding all over the place. . . . It looked almost Mil Spec. I have no real idea about how effective the shielding was, but it certainly looked nice. Another matter about the basic PC boxes is that nobody has to research them anymore. A lot is known about them. You just have to look at a "good" example and copy it. Be that as it may, a lot of IBM related stuff has failed FCC over the years. Not surprisingly, from what I gather, a lot of it was IO stuff. -- Jim Omura, 2A King George's Drive, Toronto, (416) 652-3880 lsuc!jimomura Byte Information eXchange: jimomura
goldstein@delni.enet.dec.com (Fred R. Goldstein) (12/08/90)
In article <479@aupair.cs.athabascau.ca>, rwa@cs.athabascau.ca (Ross Alexander) writes... >In short, the 1040st howls all up and down the bands. We're talking >continuous S9+20dB heterodynes and growlies from baseband all the way >up to 450 MHz (to be charitable, it does fall off somewhat in the UHF >bands. On 145.01 MHz it's still pretty damned loud). How this thing >ever got past the FCC I can't begin to imagine. I can only assume >Part 15 subpart J class B means "leaks RF like a sieve". It took >fairly heroic measures to get the intereference down to the point >where I could recieve anything from other stations; lots of bypassing, >coax cables, chokes, shielding, and physical separation were needed. >Having done all that it works fine (good serial I/O, lotsa horsepower >:-). This is my first posting in this group in many moons, since I'm a discouraged Atari owner now using a clone for packet radio among other things... But I don't understand the problem. The _nicest_ thing I can say about my 1040STe is that it is _quiet_ (at rf). My no-name clone, and its el cheapo monitor, make moderately annoying noise on some of the HF bands, and a little (S1 but enough to clobber weak-signal on 2m, not packet) on 2 meters. The 1040St, on the other hand, makes very little noise. If my wife is using it in the next room, I don't know it. Note that I _never_ breached its internal shielding to, say, add memory or an 8530 or TOS1.4 or anything else. It's still factory-soldered. And there's no HD, just the build-in sloppy disk, and the Atari color monitor. If it could do HDLC (darn, no bus!) it would be a great packet computer. I'd check the internal shield integrity. --- Fred R. Goldstein k1io Digital Equipment Corp., Littleton MA goldstein@delni.enet.dec.com voice: +1 508 486 7388 Do you think anyone else on the planet would share my opinions, let alone a multi-billion dollar corporation?
dclemans@mentor.com (Dave Clemans @ APD x1292) (12/08/90)
For the record, the two ST systems I have are both relatively
quiet in a RFI sense; while I've never specifically checked, I've
never noticed any interference on nearby TV's, radio's, etc.
I heard the following from someone who visited Atari headquarters
in California... (i.e. this is all hearsay)
As far as RFI containment is concerned, they design their systems
to match the minimal requirements of the country where it's first
introduced; typically Germany. To ship product anywhere where
standards are stricter they have to "patch" the design.
However, I've heard that Germany is going to change its RFI
standards soon so that they are greater than or equal to US
requirements, so the above problem might go away soon.
The first and foremost requirement on manufacturing is keeping the
cost down. If you get lucky things will be fine. If there is
a run of "marginal" parts, boards, etc. there could be noise
problems, reliability problems, etc.
Staffing was described as being lower than you would expect in
a one product startup company. They basically can only work on
one product at once; to get different products out they have to
work on them in series. If a particular product never gets enough
priority to reach the top of the queue, it never gets finished.
dgcBob_BobR_Retelle@cup.portal.com (12/08/90)
Ross Alexander asks: >What exactly is Atari doing wrong, and why are their machines >so noisy? Corner-shaving and cost cutting have limits, y'know. Speaking of which, Atari has apparently killed the STacey... couldn't get it past the FCC... Perhaps somebody else ought to ask John Townsend to confirm or deny this.. If I asked, it'd probably be taken as a bash or insult... BobR
ekrimen@ecst.csuchico.edu (Ed Krimen) (12/08/90)
jimomura@lsuc.on.ca (Jim Omura) writes: - They leave the stuff as marginal as they can get away with - for RFI because anything you do about RFI costs money. - - But taking it from the top, the problem starts with IO interfacing. - You can build a dirt cheap computer with no IO. A grounded cage - does it. But as soon as you start adding IO ports you've got - troubles. The Atari ST is rich with IO ports, and not all - of the best possible quality. You've got the 2 MIDI ports (the - basic spec of which looks noisy I might add -- not Atari's fault), - an RS-232C, semi-SCSI, Centronics, 2 joystick and cartridge port. - Take that "Centronics" port. The original Centronics used every - other wire as a balance which could have helped cancelled "some" - RFI (doesn't help that much anyway really). And I bet the TT with all its ports causes lots more problems. -- Ed Krimen ............................................... ||| Video Production Major, California State University, Chico ||| INTERNET: ekrimen@ecst.csuchico.edu FREENET: al661 / | \ SysOp, Fuji BBS: 916-894-1261 FIDONET: 1:119/4.0
cmm1@cunixa.cc.columbia.edu (Christopher M Mauritz) (12/09/90)
In article <36668@cup.portal.com> Bob_BobR_Retelle@cup.portal.com writes: >Ross Alexander asks: > >>What exactly is Atari doing wrong, and why are their machines >>so noisy? Corner-shaving and cost cutting have limits, y'know. > >Speaking of which, Atari has apparently killed the STacey... couldn't get >it past the FCC... Last I heard, the STacey only had FCC class A certification. This meaning that it isn't intended for home use, only "industrial" use. > >Perhaps somebody else ought to ask John Townsend to confirm or deny this.. >If I asked, it'd probably be taken as a bash or insult... I doubt he'll say anything about it at all... > >BobR Cheers, Chris ------------------------------+--------------------------- Chris Mauritz |D{r det finns en |l, finns cmm1@cunixa.cc.columbia.edu |det en plan! (c)All rights reserved. | Send flames to /dev/null | ------------------------------+---------------------------
trh@atari.UUCP (T R Hall) (12/09/90)
Bob_BobR_Retelle@cup.portal.com writes: >Ross Alexander asks: >Speaking of which, Atari has apparently killed the STacey... couldn't get >it past the FCC... > >Perhaps somebody else ought to ask John Townsend to confirm or deny this.. >If I asked, it'd probably be taken as a bash or insult... > >BobR My, aren't _we_ testy? Last time I (an Atari employee) spoke to someone in sales (in the person of Gary Trameil), we were selling each STacy we brought into the country. Which tends to indicate: (A) STacy has passed FCC (B) We haven't dropped it Of course, now I'm terribly insulted, and will now proceed to go away and pout and not talk to you nasty users any more. B^). TRH
stephen@oahu.cs.ucla.edu (Steve Whitney) (12/11/90)
In article <1990Dec7.115131.26377@lsuc.on.ca> jimomura@lsuc.on.ca (Jim Omura) writes: ... >They leave the stuff as marginal as they can get away with >for RFI because anything you do about RFI costs money. >But taking it from the top, the problem starts with IO interfacing. >You can build a dirt cheap computer with no IO. A grounded cage >does it. But as soon as you start adding IO ports you've got >troubles. ... Not to mention that the FCC and VDE require you to attach something to every port which can conceiveably be used simultaneously. For FCC testing, these peripherals have to be turned on! So you get their emissions and emissions from the rat's nest of cables added to your own product's. For VDE testing, the peripherals don't have to be turned on, but their cables still act like antennae. Atari computers have so many ports, I'd hate to see the test set up. =Especially= the TT's testing. That thing has more ports than you can shake a stick at! --Steve >Jim Omura, 2A King George's Drive, Toronto, (416) 652-3880 >lsuc!jimomura >Byte Information eXchange: jimomura -- Steve Whitney "It's never _really_ the last minute" (())_-_(()) UCLA Comp. Sci. Grad. Student | (* *) | Internet: stephen@cs.ucla.edu UCLA Bruin--> { \_@_/ } GEnie: S.WHITNEY (To Cal: We still have the axe!) `-----'
kdarling@hobbes.ncsu.edu (Kevin Darling) (12/16/90)
>Not to mention that the FCC and VDE require you to attach something to every >port which can conceiveably be used simultaneously. For FCC testing, these >peripherals have to be turned on! So you get their emissions... Plus the horror stories of things like this: you hook up a monitor or printer or terminal (which supposedly has passed FCC), and *it* is putting out so much RFI that your own unit "fails". I can sympathize with Atari... we're going thru an FCC approval right now. Something always holds things up. For instance, first the lab wanted a ground strap on the keyboard connector. Okay, so they ship the unit back, the strap is installed, the unit goes back to the lab, but now it was damaged in shipping, etc etc. Then the lab is busy with other stuff. It goes on and on. What should take a week, takes literally months instead. Sometimes I'm surprised that *any* equipment actually makes it to market. - kevin <kdarling@catt.ncsu.edu>