B'port board problem, please help

[fredhh47]

(This question also appears in another discussion) I have a B'port Boss 6, Series II machine that I depend on, mostly for its capacity. I'm having a problem with it, I think it's getting Machine Alzheimers! I am scrapping a lot of parts recently because a single character in a line of code will get changed in memory. Like for instance, a decimal point will become a G. That can get nasty. If I fix the error, another error will crop up somewhere else, or the same error will keep appearing. I have tried all sorts of programming tricks, dummy code, macros, whatever, but the errors still appear. Initially, the "first occurance" of an error, is at a pretty predictable point in the program. After the first one appears, it's a crapshoot from then on. Just when I think I have outfoxed the machine and edited or programmed my way around all the errors, BANG, there's another one.
What I'd like to know is exactly where a program is stored when it is loaded into "memory". As in which board, what chips? Is it the LSI, or one of the others? Does anyone know with reasonable assurance? I have some spare boards I bought to try and fix the problem, but it doesn't seem to make any difference, so maybe I'm not swapping the right board. When I put in my spare LSI board, the machine won't boot at all, so I don't have much confidence in it. Of course, I don't have the $$$ to get a real service man in to fix it (and my service man is stumped, anyway), so I'm appealing to the kindness and generosity of semi-strangers. Any suggestions?
Thanks,
Fred

[NC Cams]

Old machines tend to start having interconnection problems, especially in a shop environment with vibration, coolants, moisture and god knows what sort of other contaminants are in the atmosphere.

Dirt, debris, oxidation,vibration all take their toll on low voltage connections that pass hardly no current. Especially if the OEM used tin coated IC's/sockets instead of gold plated ones (mmost don't use gold as a cost savings). Since you're dealing with 5volt TTL signals in many instances, a 1/2 volt drop due to a corroded connection is a 10% loss of signal.

First go thru all your control circuit connections and make sure they are clean and TIGHT. Then press on any socketed IC's and see if they "crackle".

If your controller had socketed IC's and they "crackled", this is probably the problem. Heating and cooling can cause chips to loosen. The socketed chips could also be corroded - even slightly, this causes flakey problems. The "crackle" is the chip resliding home and/or "breakage" of any corrosion that has occurred.

We had a dynamometer that was fairly dated and would do intermittant things just like you said - the darn thing was haunted. One day, I methodically pulled each IC out, buffed the pins with a soft eraser and replaced them. Ditto that for all the interconnections.

Thing ran perfectly thereafter.

My boss refused to believe that something that simple was all that was wrong but the machine ran and flawlessly - unti the IC's corroded again. We'd simply rebuff the chips as part of normal maintenance quarterly and that was that.

Oh, and the factory tech guy was clueless as well - his solution was that the system was worn out and needed to be replaced.

Hardly.

Learned the crackle trick from a radio station engineer who's transmitter was on a hill that overlooked a steel mill. Noxious fumes did a number on the connections.

They found a special electrical contact lubricant after time that offered some protection. I believe it was sold by Nye and was used to coat the pins. It did help.

And be sure to check ground circuits especially close - copper bonded to steel does strange things. Ground currents can cause squirrely things like let the smoke out of IC's and damage irreplaceable circuits.

Have fun...

[machintek]

Besides on a routine basis of cleaning the boards as mentioned aboveand especially the edge connectors with a soft eraser, I have seen your problemm once before. The program memory is on the LSI-11 board. You may have a bad memory chip.

George

[fredhh47]

NC CAMS- I will try that solution, but I have been fairly careful about regular cleaning and checking of the boards. It is one of my pet peeves that B'port arranged the fans so that dirt and oil vapor was pulled into the control enclosure from somewhere near the floor, where it would be most concentrated, and blown onto the boards indiscriminately. I have on several occasions suggested that the solution to other people's problems might be dirty boards, and I know from personal experience that probably 7 out of 10 times we've had a problem, cleaning and re-seating the boards cured it.
George, The LSI-11 board is what I needed to know specifically. Which are the memory chips, and what is the likelihood that I can remove and replace them? I know they're probably soldered in. Is there any order as to how the chips are "filled" when a program is loaded? I imagine they are low capacity rams, like 4K, or so. Is it possible to upgrade to larger capacity chips and get more memory capacity? I am often hampered by the small amount of available memory (somewhere around 10k as I recall). I am currently running programs on other machines that are 250k or larger (God bless Centroid, and DNC mode on my 308-SX). I wouldn't mind a solution that would let me run some of these longer programs (OK, not the 250K ones) on the Boss 6 and get some of the load off the other machines. Does the Boss 6 control do well with any DNC drip feeders? If so, which one works well, or best? I know that EZ-UTLS (the communication utility from EZ-CAM) has a DNC mode, but I've never tried it. Would it be necessary to use the "dreaded" REMOTE DATA INPUT function, since I cannot see how the machine could be run in EDIT mode?
Thanks for the input.
Fred

[fredhh47]

George, one other thing, my service man, in a phone consultation (which he does for free because I have kept him busy for about the last 15 years), said the reason that the machine doesn't boot with the alternate LSI board installed might be due to a voltage setting being out of range. I forget exactly where he said I would find the pot to adjust, and he's unavailable for comment because he is on a well-earned vacation. Any ideas what he was talking about? Other than the board being outright BAD or a voltage setting, are there any other reasons why the machine won't boot with it installed? Thanks,
Fred

[machintek]

Why would it not boot with another LSI? The logic power supplies are massive.
Do you get a 70007 on the display or just the ERR LED?
On the large white logic power supply is a small board. It has 3 multi turn "pots" on it. Top to bottom, it is the 5V and then the -12 and then the +12. I am certain of the %v, the 12s may be reversed.
At the top of this power supply and near the logic baords is a terminal strip with 6 very fine wires in it. 3 are blue and 3 are lavender (?). I remember that one is labeled 76. Go to the two alongside of each other of each color. This is where I measure the 5VDC and set the pot to adjust. This one is critical. If you see that it varies a lot with the new vs the old LSI, you may have a LSI problem.
The next two wires out (one in each direction) is the 12 VDC and the two outside ones are -12VDC.

Google Digital Equipment Corp LSI-11 and you will find user groups dedicated to data about this mother board. This is a KD-11L module PER DEC as it has the 5 main ICs. The Boss 4 and such had a KD11-F module.
There are a lot of jumpers on these boards to configure them correctly.

George

[NC Cams]

Old memory plug in IC's are REAL hard to find anymore. If you do/can find them, they are HORRIBLY expensive as you have to buy them from liquidators who KNOW how valuable they are. Been trying to find some to plug into my Fanuc 5T (HM4513's and 6508's if I recall).

What used to be available for a buck or two each are now $15 to $40 /chip. Don't think you can simply plug in more memory unless there are plug in provisons (IE: empty sockets on board already)

Since the chips are probably hard wired or socketed for row/column accesss, unless you do some major surgery you may be stuck with what you have.

[Jonp]

Fredhh47, the boss 6 control is quite different than the boss 3-5, Boss 6 is more like the boss 7 as far as the electronic architecture, except it uses steppers and has no tool changing logic. The LSI 11/2 in todays world is not much more than a cpu, really only eqivelent to a 8080 or 8086, the RRS board has the boss program in eprom, the DEC bootstrap in prom and a handfull of 4104 rams to hold your program in memory. I've replaced these before but slim chance this is your problem, probably a bad conection as the others have pointed out. LSI 11/2 were quite common in there day and had many uses, we have one we were told came out of a early dollar bill change machine like in a laundromat, it works fine in a bridgeport too!!!! Good luck, Jon P.

[machintek]

Hood and I have discussed this. I have never seen a BOSS 6 with the half size mother board this side of the pond. All I have seen are the full size with a NTP, ZDI, ZCK and ERS boards going with it.
Stateside the BTC BOSS 7 had the half size mother board, RRS, CPL, etc.

George

[Jonp]

machintek, You have seen a lot more Boss 6's than me for sure so I'm sure you are right, I can remember seeing the configuration I mentioned once but who knows where it came from. I'm really puzzled now so I'll have to do a little research, I've studied the early DEC equipment pretty hard guess I need to study the earlier Boss's, Good day Jon P.

[fredhh47]

To all who offered help: I think I got it fixed. I swapped the 5 main chips from my LSI-11 board into the board that would not boot up, and everything seems to be fine now. I think the bad memory was on my board, and something was wrong with one or more of the 5 main chips on the other one. Thanks for the tips and help. I will save the adjustment info for a rainy day, just in case.

[fredhh47]

Yes, I was getting the 70007 code output. The power would come on, the display would light, but I would not be able to move the axes, load a program, etc. The machine is up and running now, but I will take some time to check voltages in the near future.
Here's a question: Has anyone else had their machine lock up if you hit the hold button at the moment that the quill gets to the clearance point above the part? I have had this occur many times and only recently figured out how to get things going again. This remedy also works when I do whatever it is that I sometimes do that results in the display showing T14 and the machine taking off on a rapid trajectory towards the X and Y limits. (I still don't know what causes this to occur, exactly, but it has to do with a MDI error, I think)
Solution for both problems (if you remember to hit the HOLD button instead of the E-stop). If the machine is not moving, gently tap the E-stop button, just enough to make the light come on. Then press the Override button, home the quill, check your zero settings, just in case, and you are back in business. I used to go thru the whole process of Clear/Reset/Reload/re-key the TLO's, etc, until I tried this. Much easier. We have so many power failures out here in my neck of the woods that I ALWAYS write down my zero settings and TLO's.