The collected works of robespierre - Page 2

Image ?
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
The Indigo2 cases are all exactly the same pattern, only the color of the plastic is different. Besides teal and purple, there was actually a blue color and all of the R8000 Power Indigo2s I've seen have been blue.

The two types of midplane (non-Impact and Impact) are compatible with all the main board types (IP22, IP26, and IP28) since they are electrically the same. IP26 is not, however, compatible with Impact graphics (no driver was written).

The Impact power supplies have a dummy load board that needs to be hooked up when no Impact graphics boards are in the system. Aside from that they are compatible with any configuration of gfx and cpu, except that IP28 machines use specific power supplies.

The power supply shouldn't be hot when the computer is off. Sounds like a broken shutdown circuit.
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
Yes, it may be that the ones I saw began life as PowerChallenge Ms. Graphics can drop right into the *-Challenge Ms, unlike Challenge S.
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
One more detail: there are four types of Impact power supply, two of them are specified for "dual head" graphics (i.e. High+High, Max+Solid, or High+Solid). So if you run these Impact graphics combinations you need the higher rated supplies.
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
Looking at the picture Jan-Jaap posted, the copper foil doesn't appear to have enough surface area for an effective heat sink. Could it be they are really for RFI? Was there a RFI shield on the tin gutter around the two large chips?
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
I feel a demiurge coming on...
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
I remember using hdiutil(1) and asr(8) [Apple Software Restore] to create backup images of disks and restore them when I upgraded to larger disks in my powerbooks. The process wasn't entirely obvious, because asr needs to "add file checksums" to the image before you restore it. (Maybe this step was really optional but it seemed like a good idea). So the image file needed to be in a particular [uncompressed] format until checksums were added. I think it took about an hour to image a 100GB disk and restore it into a 300GB.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
what makes it hard are the ground planes that suck up all the heat. i bet it could be done by clipping off the device and using a 140 C preheater.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
I found that booting from the DVD didn't work the way I wanted, either because of limited tools available or not being able to unmount the drives. In the end I made a minimal partition on an external drive, and copied system files onto it to use it for staging backups.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
Just from visually tracing Canavan's pictures, it's clear that the 5 RJ-11 plugs on the board correspond to the four controllers and the memory unit. Each controller has a 3-pin socket, with power and ground on the outer pins and data on the center pin. the power and grounds are all the same (though they have individual ferrite protection)
each RJ-11 cord has the same pinout, with pin 2 for power and 4,5 for ground (seen from the joybus board). Whereas the controllers have 1 data pin, the RJ-11 plugs on the U64 have 3 data pins (probably send and receive plus a clock). Besides the memory unit, the joybus board clearly does buffering, 1-wire encoding/decoding, and diode protection. So what I expect is that the "controllers for U64 dev board" are different from consumer controllers by bringing the send,receive,and clock signals out directly without 1-wire encoding, i.e. they are simpler than the consumer controllers by missing that component.
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
cmonkey wrote: Thanks for the input robespierre. So based on your musings don't you think it would be possible to use retail N64 controllers on the U64 board by simply changing the cable plug for an RJ11?


I think they could be adapted by soldering RJ11 cables onto the right internal points, as long as the cable is kept relatively short. There could be other changes to the special controllers. No way to know until you try it, of course.
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
dclough wrote:
Holy mother of VAX. That looks... insane. Where's the power?


There are 8 large spring bolts on the rear side that could be used for power, BUT they are not double insulated and would be too vulnerable to short circuit. It's more likely that power is supplied via the front panel.

Quote:
Is this thing part of a larger chassis?


It was designed to slot into a frame, probably in the electronics bay of some weapons system.

Quote:
What are all those plugs on the front?


Aside from a couple of BNC connectors, they are all MIL-DTL-38999 connectors. Apart from "FLIJ1" which means "Forward-looking infrared" they are only marked with numbers.

Quote:
What sort of cards go inside this thing?

I would guess they would be XMI cards, perhaps in a smaller form factor.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
VAX 6210 != VAX 7000/860. The latter is over 100 times the VUPS rating.
I guess that's why the topic says "ultimate".

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
I don't think that's right; if you hold down command-option-S you can see all of the "dmesg" output from the boot process.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
Was this before or after the technique of molding the letters into the plastic so that they are visible even when the key gets worn down? That has to be my biggest gripe with recent keyboards, there's no excuse for the key legends disappearing.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
For a while computer companies specified the R4000 by its external clock speed, which is half of its internal speed. So 75 MHz clock on the board == 150 MHz R4000. The PROM version may be too old to recognize the PRid of the R4400, but I wouldn't expect any problems from that.

have you tried the PROM's date command?

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
ARM has never had an implementation that was performance-competitive in its own generation of CPUs, has it? I suppose that wasn't a criteria in the architecture design, given the focus on the embedded space. In some ways it is more "modern" than MIPS: no branch delay slots, POWER-like load-update and save/load-multiple, etc. But it wasn't until last week that they had a 64-bit processor!

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
I found that every time I took one of the DEC 3000 machines apart, my hands started bleeding. Watch out for that.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
thanks, I will look into it also. For some reason the VME I/O plate in Personal IRISes is especially subject to corrosion (some reaction with the adhesive on the plastic strip there)

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
Most of the home computer BASIC interpreters were written by Microsoft, including on C-64 and Applesoft BASIC on the Apple ][. The licensees made various changes to add their own features (like disks and graphics).
Bill Gates was a strong believer in people being able to write their own software, and supported the release of Visual Basic when others opposed it. As a programmer, I'm not sure that was a good thing; the experience of writing programs in BASIC is said to leave one "mentally mutilated beyond hope of regeneration".

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
Visual Basic has all of those features, but encourages terrible habits notwithstanding.

In my earlier post, I forgot the loudest cracker: people who somehow think it was difficult to hack and abuse PC-DOS never used DEBUG.COM.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
Tilera have chips with 100 cores.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
for the DECstation you need a 3W3 cable. DEC made 3W3 to 3xBNC cables and I think that's it. so it's difficult to use them on monitors without BNC inputs. all the DECstations use sync on green. You could use the DEC cable + 3 female BNC couplers + VGA to BNC cable, but that's getting a little bit long.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
the keyboard and mouse for the DECstation can be the old LK201/LK401 from VT2xx-VT4xx and VAXes, and the VAX Hawley mouse. There is a cable adapter for those, but you can just wire your own DA15 plug adapter.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
EIA423 is not exactly the same as EIA232... specifically it has a lower voltage swing, ±6 V instead of ±12 V. So it's safest to use level converters for this application, although you might be able to connect to rs232 without burning anything out.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
Coax to TP Media converter

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
even more OT...
good luck finding that Silphium...

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
converting text to paths is very easy... you just select it all and use the Type->Create Outlines command.
that way the recipient of your file doesn't need any fonts to print it. but by the other hand, he cannot edit the text anymore.
You need to switch to View->Preview to read the text now, because View->Artwork will show the paths that you made.

rendering outline fonts is not that intensive, it only took a few seconds to do each page on a 12 MHz LaserWriter back in the day. But I once printed a large graphic (4 letter size sheets vertically and 6 horizontally) and that took ages just to do the first page.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
fuses (of the miniature type) don't make any sound at all when they melt. likely you had an electrolytic capacitor rupture, or another component that burned up. the force of the explosion it made could have blown it clear of the board leaving nothing behind but some solder pads. You said something about a smell?

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
It has to do with WID planes... graphics workstations (from Sun, HP, SGI, and IBM at least) usually had the ability to use a different color depth, colormap, or double-buffering mode for each window including the root. They have hardware (either dedicated bitplanes for window ID, or a window mode display list) to change the interpretation of the framebuffer when the DAC enters its region. Something similar was possible on old home computers like the Atari 800 and the Amiga.
Some more details at http://linas.org/linux/graphics.html

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
Irix 5.3 includes both AT&T csh(1) and tcsh(1). Only tcsh(1) has line editing and arrow keys.
ksh(1) also has line editing, but with the vi command set only. So you would use ESC k instead of up arrow.
[Linux, BSD, and Darwin don't have a real csh; if you check you'll see that csh is hardlinked to tcsh on these systems, and it mostly acts like tcsh.]

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
sounds kind of like a sonicwall, but hopefully better support and performance?

I use a mikrotik RB532 with RouterOS. doing everything from a command line gives me a warm fuzzy feeling, and I don't need fancy virus scanning etc.
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
please don't use code as gross as that article though. ((65535 * (long) v)/256) isn't even correct and is perverse compared to a bit shifting operation.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
Have you had a SPARC ELC? Deeper case but much easier to find. The Sun Voyager is also rare.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
jan-jaap wrote:
So, in my case, either I'm going to go with a decent 'consumer' (integrated wireless blabla) router, or I'm going to slap together a real, dedicated router. See, there's one thing I didn't mention so far because a dedicated router simply never was something I considered: I still have an FDDI network segment. Right now I use my Linux server to route between FDDI and ethernet.

Energy isn't cheap here, so low power is crucial. A dedicated router would have to be some Atom platform, with a PCI slot (for the FDDI card). With 3x gig ethernet (WAN, LAN, DMZ).


If it weren't for your >100Mbit requirement, I could recommend the Soekris net5501. It has a 32-bit PCI slot that could host an FDDI attachment (although my SK-NET card won't fit in the standard case). I know it's looking long in the tooth now, but most of the new Atom embedded boards do not have classic PCI and won't be able to host FDDI. The Soekris firmware is special in that it transforms PC BIOS calls to serial console operations, so almost any x86 operating system can be installed headless. The power consumption is 6W.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
At least look at the Zephyrtronics stuff—they have pretty reasonable prices on accessories like flux dispensers. Their non-contact rework solution is almost ideal, with preheating from underneath and air pencil from above, and the board cradles are amazingly convenient.
I also have a Hakko 939ESD and while digital temperature control is convenient, it isn't really a big deal for hobbyist rework and assembly. The 936ESD's control is good enough, although its iron doesn't respond quite as quickly. I do appreciate the ability to control tip temperature for different solder types (there is more than just Sn63 and "lead-free"!) and use different settings for soldering and desoldering. The 936 is one of the most affordable all-duty products on the used market now, much more practical than Metcal I think (needs different tips for different temperatures and the tips are fairly costly).

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
You may be able to find old products from Pace like the MBT series soldering/desoldering gun/hot air stations, very good I have heard, but parts and accessories are less commonly available.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
There is a Texas Instruments SuperIO chip in the O2 that controls the serial ports. Its mask revision is probably printed on the chip package, so it would be simple to see if there is a difference between various moosehead board revisions.
I found one site selling the 030-1327-002 as "CRIME1.4" which implies that the board has a different revision of the CRIME asic. This may not be true, as I have some docs which state that CRIME 1.5 was required for R10K processors, and all publicly released O2s support R10K. In any event, there is a revision register within the CRIME address space, and likely a revision register in the SuperIO chip as well, so a small program could be written to extract that information.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
A real Hakko 936 isn't much over $40. They sold millions of them because they always work. Temperature controlled soldering works in one of two ways: either the Metcal method that uses a magnetic material with a Curie point at the required tip temperature, fed by radio waves that are only absorbed if the tip cools below that temperature; or the method used by every other vendor, where the heater has an integrated thermocouple. The first has its advantages: heat recovery is almost instant, doesn't need calibrating, and there is no sensor or microcontroller to fail. The second is more practical because of the user's control over the temperature setting (although many stations have lockout keys/codes to prevent this being changed once it's set). But that flexibility comes with the need for more consumables (the heater/sensor elements) and a requirement to calibrate the sensor if absolute temperature accuracy is called for.

For CMOS devices you need to use a system that guarantees that there is no static charge on the tip by keeping it grounded, and shutting off if the ground has failed. Both cheap soldering guns/irons (that use 120V AC directly!) and ad-hoc air pencils built from same are not safe in that application.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP:
The Hakko FX-888 you mentioned is made in Malaysia. My 939 was made in Japan.
The Weller WES(D)51 is made in Mexico.

_________________
:PI: :O2: :Indigo2IMP: :Indigo2IMP: