PLC5/40E losing memory

[CraZAg 0]

I've been working with PLCs for 10 years and have never seen this one before. I'm hoping I can get some help. I have a customer with an Allen Bradley PLC5/40E. During normal operation of the machine, the PLC will lose coms, lock up, and have a red fault light on the processor. When you try to get online with it, the program is gone. The battery is new. We've replaced the power supply and processor. We have put a power filter on the incoming power. Still the same problem. It's very random. Sometimes will take a month, other times will happen twice in the same day. Anybody have any ideas? Thanks in advance!

[Mickey 71]

I have seen this same problem. Install a UPS along with your filter.

[Guest]

... and make SURE that the chassis is PROPERLY grounded ... the installation manual gives details ...

[CraZAg 0]

I've checked the grounding and it is correct. I'll try the UPS. Thanks for the replies.

[newpageboba 7]

I have had similar problems. We did everything you did, replaced Sola transformer, power supply & cable, processor, chassis and even ran a new ground & power source. Finally we replaced the 1771-P7 power supply with a 1771-P4S slot power supply. I don't know why it fixed it, but it did. I have also seen a data hiway connection touching the cabinet and causing the processor memory to scramble while we were doing on line edits.

[Contr_Conn 4]

Most likley this has nothing to do with power. In RSLogix 5 under Tools menu select Extract PLC5 Memory Loss Fault data file Contact Technical Support to analyze this file Log File will be erased if battery disconnected. Reloading the program will not erase it.

[paulengr 44]

If you don't find it, I have seen this happen twice before and it turned out to be a bad processor. If the fault log doesn't reveal anything (as suggested), keep in mind that a solid red fault light means just that...hardware fault. So after exhausting all avenues, suck it up, pay AB $10K, and buy a new one. If you have their troubleshooting guide, this is exactly what they tell you to do. A Sola CVT (continuously varying neutral transformer) which is what AB recommends for power conditioning/supplies is an absolutely lousy choice by the way. I've had nothing but trouble out of these except in cold weather environments where it makes for a very expensive panel heater. I've since gone to using solid state power conditioners from Tripp Lite which do the same as the SOLA except they don't fail and they are far less expensive if voltage sags are your issue. The downside is that the Tripp Lite is only available as a "brick" with power cord connectors on it...they don't offer an "industrialized" version. On the front end before even going there I install an AB "DC" surge suppressor/isolator to clean all the noise off the line. It's a nice combination of a set of zinc oxide surge suppressors (with a fault indicator...very important since these are consumable) and an Islatrol harmonic filter (best in the business). I haven't found UPS's to be all that effective because of the nearly constant cost of replacing batteries on them (you don't actually change batteries...cost of a battery is 80-90% of a new UPS!) I have never found that the ability to ride through a power loss with a PLC has any real advantages since you can't leave all the actuators and such powered up anyways. So rather than pay the maintenance expense (both bottom line and extra PM's) I've stopped using UPS's altogether if at all possible and implemented the above protection which protects the system (without batteries) against everything except a true power loss. The one exception is gosh darned SQL databases. One power blip and the database ends up corrupted. In this particular case I have a UPS but the purpose of the UPS is to signal to the PC to begin orderly shutdown. I don't want/need the battery but nobody makes a UPS with say a supercapacitor in it to get around having a battery.

[waterboy 7]

I have just had, what we assume to be, a lightning strike take out the comm channels on all devices connected to the DH+ cable. It also wiped the program and faulted the processors on each end. I suppose its concievable that a lesser voltage spike on the DH+ cable could simply wipe the program and not damage the DH+ port. Just another idea to think about.

[paulengr 44]

It's not just conceivable and doesn't take a power surge. If you send the wrong MSG instruction format (I don't recall which one) on at least one vintage of PLC 5/40E, the processor faults and dumps the program. Remember that on every program scan, the hardware fault checker does a checksum on memory and various other consistency checks. If anything is found to be amiss, it saves the data into a fault log, and proceeds to wipe out program memory and set the fault light on.

[newpageboba 7]

I had a situation once where the processor would fault and lose its memory every time we did an online change (PLC5-40B). We found it to be the unterminated end of a DH+ blue hose touching the inside of a cabinet.