Dirty connector update-more info

I would inspect all visible solder joints, check the CR voltage and around the 555 and change all the electrolytics. Record changes and take representative voltages around anything analogue, if problem recurs then attack something else. I assume all the TTL is soldered direct to pcb , not socketed, if socketted then likely suspect there. Any use of vias ?

All solder joints inspected and good. Do the dipped type of electrolytic caps go bad? I thought they were immune to the problems that the liquid filled ones have. No sockets. The board is double sided with all device pins soldered on both sides of the board. So I guess those are all vias? Thanks, Eric

--
This email is free from viruses and malware because avast! Antivirus protection is active. 
http://www.avast.com

OK, not an opto anywhere, so my idea is toast! So, is it clear that powering off doesn't fix it, but reseating the board DOES, at least temporarily? If so, that pretty much indicates a connection problem. Most of this stuff was wave soldered, so it is not real likely there are bad solder joints on the chip-board connection. What type of connectors are on the board-backplane connection? If Elco Vari-lock or similar contact, they need to be squeezed. My old Allen-Bradley 7320 control was all Elco Vari-Lock contacts without the plastic housings on the board contacts. This allows them to get bent apart when the board is wiggled out. Elco vari-Lock contacts are hermaphroditic, and sort of a narrow fork. The close faces of the tines are angled, and wipe against the other contact, assuming the tines are parallel.

So, I have no idea whether your control uses this type contact, but I definitely ran into this on that old A-B. Card edge contacts can get dirty or sprung after many insert/remove cycles.

Jon

Do the dipped type of

Ugh, those dipped tantalum electrolytics? They can be awful when old, the worst case is gear that has been run for 10 years, then parked for a couple years and then powered back on. But, they don't become mechanically sensitive, so that doesn't seem to be your problem. When they go bad, they can catch fire when connected across the main power supply.

Jon

It seems like there is a heat problem and some sort of removal and replace problem too. When the machine is run with the cabinet closed it alarms out faster than when it is run with the cabinet open. I have been so busy running the machine and working around this problem that I may be wrong about removing and replacing. Today I'm a little less busy and I'm going to try a couple experiments. Eric

--
This email is free from viruses and malware because avast! Antivirus protection is active. 
http://www.avast.com

You might be able to isolate the fault with freeze-mist or maybe a jet of compressed air, one chip at a time. Of course, if you could get enough technical info to know which input on the board is the oil or air sensor (maybe you can glean this info by tracing wires on the machine that does have these sensors) then you can just replace all chips in the signal path through the board.

Ahh, for the days of Allen-Bradley when you got a 3" thick circuit diagrams manual and a 1" thick theory manual with each CNC control.

Jon

I can get a wiring diagram which I'll download tonight. Maybe that will help. The control was made by Fadal as far as I know. Some Fadal machines could also be had with Fanuc or Siemens controls too. Eric

--
This email is free from viruses and malware because avast! Antivirus protection is active. 
http://www.avast.com

All that TTL. We're gonna party like it's 1979!

After seeing this parts list, I'm gonna re-suggest checking the power supply to this board, probably +5 V DC. If it's at all low or high, you might look into why that is. (TTL parts draw relatively a lot of current from the +5 V supply. Some of yours draw less - the 74LSxx and

74ALSxx part numbers - but in general, TTL boards require stout power supplies.)

Here are some comments on what these chips are *probably* doing. This may or may not help your problem, though...

This probably works with CR to measure some kind of analog signal, probably produced by a sensor with a variable resistance.

These guys might be in charge of talking to the rest of the controls (the motherboard, other cards, etc).

These guys are probably in charge of receiving commands from the main CPU.

This should have an output of 1.20 to 1.25 V at 25 C (77 F). If it's way off from this, you could have trouble. It probably lives pretty close to U8.

The "square film" caps are probably OK. The "dipped electrolytic" are probably tantalums - they look like a little epoxy blob, often blue or gold (but can be anything) on the end of two wires. I don't know much about if these go bad or not; I will defer to others on this one.

Sometimes this is done because they forgot to design that cap into the original board. Sometimes it is done to solve a problem that didn't appear until later. You might make sure that U21 has a cap "tacked on" to it somewhere - it can be on top of the chip, or on the bottom side of the circuit board under where the chip is.

Matt Roberds

Greetings Matt, Thanks for telling me what does what on the board. You are correct about CR being close to U8, it's right next to it. Someone else suggested testing the output of CR. I was going to look at the datasheet online to see what it is supposed to be putting out but your post has that info, which will save me some time. There is indeed a cap bridging U21, that's why I put it in the list. All the components listed are what actually populates the board. I used my 10x magnifier to get all the part numbers that are on the list. I really hope the faulty part is CR. As you know it is a voltage reference (which I didn't know 'till I looked it up) and a replacement is available cheap. And it's a two lead device. Can I safely test the electrolytic caps in-circuit with my DVOM? It has a cap test function. It only tells you the value in farads, not the ESR. The meter uses a 9 volt battery but I have no idea if 9 volts is applied to the cap under test. Thanks again, Eric

--
This email is free from viruses and malware because avast! Antivirus protection is active. 
http://www.avast.com

It's an easy test. :) You have to do it with the board powered on - it needs input current to work. Basically it's a diode, and when you put current through it, it will very reliably drop 1.23 V across itself.

One or both comparators in U8 probably use this steady 1.23 V to compare against some external signal.

I found the full datasheet for your part number at

- this part is by Maxim. The part number suggests that it was originally made by Intersil and their data sheet is at .

Ah. I thought maybe the list came from a service manual.

I don't see any stock on it at Mouser or Digi-Key. Going through Mouser's component picker, the TI (ex National) LM4041 may be a replacement.

It may apply up to 9 V to the circuit, which the TTL parts won't like. It probably won't give an accurate result in-circuit, no matter what voltage it applies. If it costs $500 to get it wrong, I would take the caps out of the circuit.

If you unsolder one lead of the cap and test it that way, the test results should be reliable, and nothing should get damaged. If these are the caps I am thinking of, though, it's hard to unsolder and remove just one lead from the board - you may have to unsolder both leads and remove it entirely.

Matt Roberds

I'm going to agree on both points that the board is a "classic" and that it may be time to check the power supply. It may even be that the cool off period the board gets when you mess with connectors is really doing the magic.

If it's safe, I'd run my fingers across the chips to see if any are hot just after the machine acts weird. The 7400 stuff will warm up, but should not be hot.