Dave Dunfield wrote: This may be a stupid question but bear with me. Is the surface of the FPU chip getting hotter than the heatsink sitting on it? Peace... Sridhar

Yes, that's why I speculated about fans - perhaps it's designed for (say) three times the airflow the case is giving it, and the only real problem is the fan? It strikes me as unlikely, but no more so than any other failure mode I've come up with that matches the symptoms. /~\ The ASCII der Mouse \ / Ribbon Campaign X Against HTML mouse at rodents.montreal.qc.ca / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B

It's original from my POV in that it's the one that was in it when I got it, but the machine was out of my hands for some time between the factory and me, and I have no idea what might have happened to it then. Furthermore, fans do wear out, and it could be that the original fan, while original, is no longer up to original specs. Could be for all I know. Anyone got a fan CFM spec for the 2k? :) /~\ The ASCII der Mouse \ / Ribbon Campaign X Against HTML mouse at rodents.montreal.qc.ca / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B

A good idea, but I don't think this is the case. According to the technical manual, the power supply and fan should be the same in the main system cabinet and in the matching expansion boxes (same chassis) - I have two expansion boxes (tape and disk) and all three have exactly the same fan installed. I've not done measurements, however the fan seems to be operating correctly, good airflow, doesn't sound like it's dragging etc. The sound and airflow are not noticibly different from the other two boxes. According to the technical manual, the fan is supposed to be "variable speed". It's a 12V DC fan, however it is powered via a 7905 (-5V) regulator which has it's reference through a resistor divider which includes a thermistor. So presumably it speeds up as the cabinet gets hot - since the thermistor is in the power supply, which has somewhat of a separate airflow path (metal plate between PS and boards) it may not detect the "hot spot" forming at in the other airflow path. I'll have to do some measurements. I could try hard-wiring the fan to full power, however this thing *should* work as it is designed, and the fan supply seems to be working, as noted above, it behaves virtually identically to the other two boxes. I've done some googling and have not found any refernces to other uVAX2000's having this problem. Removing the color display buffer board should allow for better cooling of the mainboard, as the two boards are large and sandwiched together in a fairly thin air duct. But again, I think if this were a common cause of this problem, there would be references. The fan in mine (as well as the two expansion boxes) is a: PANAFLO Model FBP-08B12H made by Matsushita I didn't have to take any of my boxes apart, the fan label can be easily read through the rear air grill. I would like to know if your machine gets as hot as mine - with the main cover removed and nothing else, after a few minites, mine gets a very noticable "hot spot" on the flexible shield covering the upper (solder) of the mainboard above the CPU's (near the front of the machine, near the center but more on the drive side). Unfortunately you can't access the CPU/FPU heat sinks without removing the board (unless you have a thin temperature probe) ... on mine, both the CPU and FPU sinks get to the point where hanging on to them would probably result in skin damage (blister etc.) - it seems to me that the FPU (the one closest to the RAM array) get a bit hotter, but both are well into the "uncomfortable hot" range... Given that they both seem to run directly from the 5V rail, and that is well with spec (something like 5.01 when I measured) I can't think of any external problem that would result in that much current draw and sill have the machine appear to operate normally (can't really think of an internal problem either). Really has me stumped - there seems to be reasonable airflow through the cabinet, nothing seems to be blocked etc. It just gets so hot it dies... if the system were that "close to the edge", I'm sure I would have seen references to more problems. I did find a few references to thermal death when operated in very hot environments, but I keep my office on the coolish side. Dave -- dave06a (at) Dave Dunfield dunfield (dot) Firmware development services & tools: www.dunfield.com com Collector of vintage computing equipment: http://www.classiccmp.org/dunfield/index.html

If my response came across as critical, I apologize, this was not my intent. I welcome any and all suggestions. Agreed that it wouldn't hurt to have the fan at full power, but the point I was trying to make is that this thing gets *SO* hot that I'm convinced theres a fault, which should be found, not masked. I don't think just running the fan at top speed will be enough - it's got pretty good airflow as it is, and it is on the other side of the cabinet (fan is at the back, CPU/FPU is at the front). The only way I've been able to keep the chips cool enough to run is with a muffin fan positioned right beside them with airflow directed at the heatsinks - much higher airflow that would be present in it's natural situation. It has occured to me that with the color frame buffer removed (which I don't need in my application), I could mount larger heat sinks, and probably make it stay cool enough with the natural airflow, but I would not feel comfortable in doing so, as it does not address the actual fault. A little more information: The heat seems to be mostly dynamically generated. Holding the CPU/FPU in reset, or stopping the clock (there's a jumper on the clock line!) results in them getting warm, but not "blistering hot" which they do as soon as you start them running. Scoping around for signs of contention, odd signals etc. I found that everything looks good (nice lows, highs and full range transitions) except for two pins on the FPU - 44 and 60. 44 is not shown on the schematic, nor is it described in the techncial manual description of FPU pins. 60 is described as "TEST - this pin is not used and should be grounded" and shown as grounded on the schematic. Both pins are showing an identical waveform - which is a "square edge" signal of unknown modulation with approx. 0.5v high level. Normally this smells of contention, however both signals appear to be no-connects as far as I can tell from the board. I tried pulling TEST to ground as shown in the schematic - it pulls down with no trouble (no signficant load), and has no discernable effect. The waveform on pin 44 is not affected. I have no idea if this is normal... Unfortunately there are many pins neither shown on the schematic nor described in the technical description - there are also a fair number of pin connections shown on the schematic without names or pin numbers. There are also things like 10-pin SIP resistor modules with 10 numbered pin connections, and unnumbered 11'th and in some cases 12'th connections shown on the schematic. At this point it's probably going to have to wait until I can find another one to make comparisons with. I haven't found anything blatently out of the ordinary, and lacking better technical information I have no way to know if signals are normal or not. Regards, Dave -- dave06a (at) Dave Dunfield dunfield (dot) Firmware development services & tools: www.dunfield.com com Collector of vintage computing equipment: http://www.classiccmp.org/dunfield/index.html

No, not at all. Hence the smiley. Agreed. I assume these CPU and FPU chips are MOS devices. Is it possile that excessive leakage across the gate oxide layer in some transsitors of that chip would cause it to run hot, but still work? I'm pretty sure I've seen chips that seem to work, but get hot and then stop working, althoguh cooling them with freezer spray keeps them running. And these were not chips driving high pwoer laods -- they were things like the clock/timer microcontroller in a VCR. -tony

Hi One thing that hasn't been mentioned. His system may be unintentionally over clocking. I've seen systems where the crystal will take of at an overtone. It is even possible that the crystal is broken and running at some higher frequency. Just a thought Dwight _________________________________________________________________ Helping your favorite cause is as easy as instant messaging.?You IM, we give. http://im.live.com/Messenger/IM/Home/?source=text_hotmail_join

Good idea - unfortunately one of the first things I checked was the clocks and they all look good - decent waveforms, correct frequency etc. Dave -- dave06a (at) Dave Dunfield dunfield (dot) Firmware development services & tools: www.dunfield.com com Collector of vintage computing equipment: http://www.classiccmp.org/dunfield/index.html

Tony Duell wrote: I expect you'd be able to tell from the tears of pure sweet joy running down the user's face if a VS2000 was running three times as fast... ;^) --S.