• Please review our updated Terms and Rules here

The HP Apollo 700 Series I'm Trying To Get Up And Running

NeXT said:
I would like to point out that you are doing an AMAZING job at troubleshooting right now. The fact you were even able to narrow down the fault in a switching power supply goes far beyond what I at least can do.
Click to expand...

Well, I'm about 3 months from graduating an electronics engineering technology program, so that helps quite a bit, haha.

Anyway, after going through and replacing ALL the capacitors on that front board, I'm still having trouble. But it'll boot up if I get the first comparator out of the three nice and hot before powering up the machine, so I'm pretty sure one or more of the comparators has some kind of fault, and if it's not the comparator itself, it might be a cracked trace on the circuit board or something.
 
In better news, the keyboard cable showed up, and the keyboard appears to be working fine when the machine is powered up (I can navigate the boot menu and such). I still have to heat up that area of the board before I can power the machine on, though. At least now I know the specific area I have to heat up and I can try replacing that comparator first when my new parts get here. I'd try putting my one extra back in again, but honestly, it's a DIP component, and the ones on this board are all surface-mount, so I basically had to solder up 14 individual wires to the thing, which couldn't have been helping the situation any.

EDIT: Yup, I just replaced three more capacitors, this time on the main PCB, near the chip I've been heating up, and I still can't power up the machine unless I directly heat up that first comparator.

EDIT 2: I'm giving it one more shot with swapping the comparators on the board. I had already swapped two, now I've swapped the third one for one of the others, so none of the comparators are in their original locations now. It's powered up, but I don't know if it's the swapping that helped, or if it's once again just the fact that the comparators got heated through my soldering iron.

EDIT 3: Yeah, either multiple comparators are bad, or it's something to do with the board at around the position of the first comparator. Even with the comparators swapped out, I have to heat the board in that position for the computer to power up. Another interesting note, what I've done to the board so far seems to have enabled it to stay powered up for longer periods of time, although if you turn it off it still has trouble turning back on until you heat the board again.
 
Last edited:
Okay, I need some serious help now. I screwed something up. I think I may have damaged either the I/O or CPU board due to some rapid power cycling. The machine turns on currently, but the PC speaker emits a continuous tone and the machine will not proceed to load the boot ROM. Any ideas?

EDIT: Got it to stop beeping by holding down the 'service' button while powering up a little while ago. But it still didn't boot, and on the next power cycle it started the continuous beeping again.

EDIT 2: Cannot get the behavior from first holding down the 'service' button to repeat. The only thing I can think to do at this point is just leave it off for a while and try again later.
 
Last edited:
I'm getting to a point where continuing to troubleshoot this just isn't worth the effort. If it magically works tomorrow, great! If not, does anybody want to buy any of the parts from this thing?

EDIT: Okay, now I'm looking at the cost of replacement CPU boards and I/O cards. I see various websites selling them for around $100 each. Doesn't actually seem that bad to me, but it's something I'd probably wait until I've graduated and have a job to buy. But then I go looking for replacement power supplies in case I can't get this one working reliably, and these same kinds of websites are trying to sell them for $700! What the heck?
 
Last edited:
jmetal88 said:
... trying to sell them for $700! What the heck?
Click to expand...
Those sites are hoping for business customers that have legacy systems that they want working at nearly any cost.

I doubt you would have screwed up any plug-in boards if you haven't applied reverse or over-voltage. And if you have, it's not worth buying new boards.
But to be fair you should be doing PS testing with a dummy load whenever possible.

I'm leaning towards suggesting that you solder heavy wires to the back of the PS receptacle on the backplane to provide GND, +5 and +12 and hook it up to a Baby AT, or less-preferably an ATX PS. In fact thinking about it, you might be able to simply back-feed that power into the unused Molex power wire in the hard disk bay. It's all the same. I think it would at least boot without the negative voltages and as I said, +9 V is for the fans.

But if the troubleshooting is driving you nuts, better to put the project aside for a while so you don't panic and do anything you regret later.

DI-II_bus[1].jpg
 
Great, I can't get the backplane out now because one of its screws is stripped... I wanna take my vice grips to it, but I can't find them anywhere.

UPDATE: Bought some new vice grips to remove the screw. And holy crap, that had NO thread left on it. Basically, it was a metal peg with a Phillips head on it. I'm going to try the suggestion of an ATX power supply later.
 
Last edited:
Alright, I did some more testing on the power supply today. I'm not sure how critical the -12V rail is, but I was testing the pinout and what appears to be the -12V rail is only putting out -6V. Likewise, the +12V rail is also weak, putting out approximately +10.6V. This is without a load. I am hoping that this is what's causing my boot failure. We'll see once I get an ATX supply wired in. I've noticed in the ATX spec, the power supply is turned on by shorting the green wire to ground, so it looks like I can just hook that up to the existing power switch (all it does is short one pin on the power supply connector to ground) and have this operate somewhat in the way it was originally intended, still.
 
Nah, I think the CPU card or the I/O card or both probably did get messed up from the original power supply cutting in and out so much. The ATX supply powers everything up right away, but I'm still getting a constant tone out of the PC speaker and no booting.
 
The normal startup is a single long and loud tone. Oddly, the service manual does not describe error codes for tones, but for the eight LEDs the list is extensive.

Did you try resetting the firmware by pulling the small button cell, and is that cell in good condition, 3V? I'm assuming you have at least the I/O and CPU cards installed and that you have not recently disturbed the locations of the memory simms on the CPU card?

There is always some hope that the insufficient minus voltages are the problem and that providing those may help?

Oh, last thing, a bit of trivia, if you take the case completely apart you might find a date and signature in one corner.
 

Attachments

  • Pa270052[1].jpg
    Pa270052[1].jpg
    58.9 KB · Views: 1
Last edited:
paul said:
The normal startup is a single long and loud tone. Oddly, the service manual does not describe error codes for tones, but for the eight LEDs the list is extensive.

Did you try resetting the firmware by pulling the small button cell, and is that cell in good condition, 3V? I'm assuming you have at least the I/O and CPU cards installed and that you have not recently disturbed the locations of the memory simms on the CPU card?

There is always some hope that the insufficient minus voltages are the problem and that providing those may help?

Oh, last thing, a bit of trivia, if you take the case completely apart you might find a date and signature in one corner.
Click to expand...

Yeah, what's happening is that the startup tone never stops. It just keeps playing. I looked at the LED error codes, and it doesn't match any of them. ALL eight indicator LEDs are on, the power LED is on, and the service LED is off. I've noticed the LEDs do the same thing if I start it up without the CPU and I/O cards installed (in other words, just the power supply hooked up).

I have tried pulling the button cell battery, and it didn't help. I did wire up the negative voltages from the ATX power supply at the same time as the positive voltages. The only difference from running it with the weak/damaged original power supply was that the startup tone was louder. It still never stopped playing it and never displayed anything on screen. I've come to realize the past couple of days how much of a mess my apartment is, though, so I've put the computer 'away' (really, just against the wall in the other room) until my space is somewhat livable again. Always a chance if I just don't mess with it for a while that it'll mysteriously start working again, haha.
 
I'm honestly contemplating ordering another 100MHz oscillator for the CPU card, kind of on a whim, just to see if replacing that helps. I mean, I know if that wasn't working (assuming it's the source of the system clock) you'd have all the symptoms of a dead CPU even if nothing else on the board is damaged, and it's probably going to be the easiest thing to replace. I tried to see if I could locate the clock signal using my oscilloscope and I couldn't, but I don't know if it'd show up even if I did find it since my oscilloscope is only rated for 20MHz signals.
 
I think at this point here I would begin to consider if it's worth my time or if I should start looking for someone to give it a new home. It's gone from working to a bad PSU to an unreliable PSU and now to an unreliable PSU with a potentially and completely unknown fault elsewhere in the machine.
You can only have so much fun with a machine when it makes you pull your hair out with problems. I'm not at all educated on these machines so I have no idea what you need to do to reach the PSU. Have you been repeatedly unplugging and replugging boards or cables? Have you checked to make sure that none of these connection points have suffered a bent pin?
 
Eh, right now I'm having fun trying to retrofit an ATX power supply into the case for the original power supply. I'll decide what to do with the rest of the system later.
 
Really, I am *right* on that line where I'm considering trying to part out the machine, though. At the same time, if the CPU card is the thing that's broken and the rest of the system is fine (which I don't really know at this point, but I can hope), the worst case scenario is I buy one of the seven in stock at whatever online store I found for $95 and see if it works. If the I/O card is also fried, the same store also has it for $95, and I found another store selling it for $90. Neither of those seem unreasonably expensive to me, it's just that I'd have to hang on to a dead system for a while before I could comfortably spend that amount of money on it.

It looks like I can make the ATX power supply board fit in the case for the original power supply, interestingly enough, and while it doesn't have any screw holes that line up properly, it's just wide enough that friction from the top half of the case will hold it in place nicely. I did discover it was impossible to slide the power supply case back into the computer chassis with all those wires soldered in place, though, so I had to remove the connector from the original power supply (thank goodness for that hot air rework station I bought a while back or that would've been a nightmare to remove pin by pin) and I'll have to wire it up on a piece of perf board (which I already cut to shape so I would still have card sticking out from the sides of the power supply case to fit in the computer chassis grooves).
 
Alright, the power supply transplant seems to have been a success. New ATX power supply board is now in the original HP power supply housing, fully modular as is the original, and powers up the machine without issue. Still no booting, indicating the rapid power cycling caused by the bad supply while I was trying to boot the machine with it actually did cause some damage.

I did check for bent pins on the backplane as suggested, but there were none. The boot failure actually happened right after a boot success. I had turned it off and tried to turn it back on again several times -- each time the power supply briefly came to life and immediately cut back out again. The next time I got it to stay on is when it failed to boot, and I hadn't touched the cards in the meantime.
 
Well, now that the power supply is fully modded, I've put the computer back together and I'm not going to do any further troubleshooting except trying to power it up every once in a while. I had a PC motherboard once that wasn't working when I got it, but mysteriously started working after I had left it alone for some time, so I can always hope the same thing will happen here, haha.

Interesting note about the startup tone: The startup tone isn't the same every time I power it on. It changes from time to time, and sometimes sounds like two tones mixed together. I'm guessing that's the I/O card fumbling around for something to do without any CPU activity to guide it, but again that's just a guess (I really have no idea what these symptoms indicate).
 
You've done some heroic work to get this 720 back to where it is now. I've seen the behavior you're describing in the past when keeping these workstations running was my job, but I can't recall the specifics for what you're getting. Logic board replacement's the best I recall, but then, I was working for HP and we'd bag up the old board to be sent back for repair rather than do the board-level repair ourselves by the time the 700 series came out.

I've still got several old 700 series around myself, as well as my favorite old 425e. I've got a 710, the sister to your 720, and a bunch of the later 700s like the 715, 725 and 735 and a Gecko. The most recent ones have a soft power-down that's one of my favorite features of those units. I've also got a B2000 that I picked up surplus (mine were all surplus units, I picked up a pallet load of old workstations from a sale at a local university once.)

I still run HP-UX on them all, porting code isn't difficult once you know how HP-UX varies from Linux or, for me back in the day, SunOS and Solaris. I've got systems running HP-UX 9.x and 10.x, I never got into 11 though my more recent systems will run it.
 
I've got a "715t" but it has an Domain/OS keyboard port in the back. I just never get around to figuring out what's in that thing, because my 9000/350 and C8000 satisfy my HP-UX jones ordinarily.
 
Well, if anyone's willing to sell me (or even loan me) an A1094-66510 CPU board, I'd love to take you up on it, at least if you can beat the current lowest $95 price I've found so far. I got this thing booted up to the GUI twice before it failed (couldn't log in, but that's another matter) and that's just not enough to satisfy my curiosity about the machine.
 
I decided to go ahead and take the oscillator out of the CPU board. Hot air helped me out a lot there once again, as I wasn't able to suck enough solder out of the PCB holes to get the oscillator loose on its own. Again, I don't know if I should be able to see anything on my oscilloscope since it's only a 20MHz model and the oscillator is a 100MHz model, but I hooked it up to 5V and ground and it looks like it's just putting out a solid 5V, to me. I'd expect to at least see a double line or something if it was working but my oscilloscope couldn't detect the individual transitions, but then again I don't know exactly how my scope should be reacting to a signal that far out of spec.

I still kind of want to order a new oscillator and see how it does. I can't tell from my online searching exactly what configuration of oscillator this is supposed to be, though. It seems to be a Fox F5L series oscillator, but I can only find a datasheet for one line of oscillators in that series. It does seem to suggest that this is an oscillator with two outputs rather than an output and an enable/inhibit pin, though.

EDIT: The CPU card itself seems to have Pin 1 of the oscillator not connected to anything. So I don't think I should be too particular about the configuration of oscillator as long as it'll work with Pin 1 disconnected. Pin 14 is obviously VCC and Pin 7 is obviously Ground (using the 14-DIP standard), which leaves Pin 8 as the output. Unfortunately, through-hole oscillators at 100MHz are crazy expensive (the ones I was looking at on Digikey started at around $60) so it'd be cheaper to adapt a surface-mount oscillator (those can easily be had for under $10). Now the question is, do I want a sine wave oscillator or a clock oscillator? My guess would be that it's a clock oscillator.

EDIT 2: Well, eBay has through-hole oscillators in the proper size for under $10. Maybe I could go for one of those.

EDIT 3: Hmm, the oscillator installed seems to want 5V hooked straight up to the case. I guess that's why what little info I can find on it calls it a -5V oscillator. The CPU card PCB just has the same 5V that everything else uses hooked up to pin 14, the housing pin, and GND hooked up to pin 7. So in other words, it's a -5V oscillator hooked up exactly like a normal 5V oscillator. I could probably get away with ordering a normal 5V oscillator in that case, I'd think.
 
Last edited:
You must log in or register to reply here.
Back
Top