SLC Memory Drop Problems

N

Thread Starter

Niall Watson

Dear List

Has anyone experienced problems with SLC (5/05) "dropping" entire memory...
It appears to happen if the power supply is cycled rapidly.
The processor faults and becomes "default".
I am reluctant to use EPROM as the system contains a lot of dynamic data.

Any suggestions or similar experiences???

Thanks in advance

Niall
sysero: systems integration solutions

Niall Watson

Phone: 01283 743022
http:\\www.sysero.co.uk
a subsidiary of BI Industries plc
 
Yes, I have had similiar experiences with even 5/03 and 5/04 PLC's. What I did to resolve the problem was add some power conditioning to the AC feeds and in one international location actually added a small UPS per PLC. It seems the PS/PLC combination doesn't take well to voltage spikes or brown-outs.

John Kost
[email protected]
 
There have been a few postings lately about the SLC-500 controllers "losing" their memory.

Let's clarify this; the SLC controllers are purposely loading default programs when the running application program has been declared corrupt by a hardware or firmware watchdog.

The SLC's are built to be particularly careful of their onboard memory and chassis I/O, and they check the validity of their program and I/O every
program scan. The goal of these watchdogs is to prevent a machine from running when it is out of control; i.e. the program is incorrect or some
of the inputs or outputs cannot be controlled because a module is bad.

Niall Watson's problem might be related to one of the circumstances that can result in a default program; when the 24 VDC onboard supply (either
through the backplane connectors or the screw terminals above the AC line terminals) gets shorted, the SLC goes to default because it cannot control it's I/O if it doesn't have solid 24 VDC on the backplane. If you have a
particularly large load (a lot of cards or some large field devices using the onboard power), cycling power quickly might result in a startup surge which the power supply could detect as a short circuit.

The other posting, in which the SLC was actually insulated from the control panel backplate might have a problem with conducted noise. The 1747-SN
scanner for RIO networks has an unusual grounding arrangement for the shield wire from the "blue hose" cable and the installation instructions
should be followed to the letter.

I have encountered only a handful of SLC memory corruption instances in which a fat green wire to a solid ground didn't solve the problem. There
are two good technotes on the A-B site (docs 7568 and 9656) that address noise and installation issues.

Good luck,

Ken Roach
A-B Seattle
 
Sir:

I have not had these types of problems, but, my a-b distributor's PLC guy always tells me that when intermittent and strange things happen like this, it's usually an undersized rack power supply (This is true in any microprocessor based controller). I always use the biggest they offer--no matter what the rack size, etc. Perhaps you should look at the i/o card count and current draw and compare it to your P.S. capacity.

j paley
 
N
Niall,

We have experienced similar problems to what you mention at times although thankfully not that often. We find the same result as you. Its always following some type of supply disruption or disturbance.


The SLC's have a trick where you can short out two tracks on the board for a certain length of time and this restores the processor to factory
defaults for comm ports etc it also clears any program. I wonder if the ability to do this in the design some how leads to susceptibility
to "clear" itself. On our real important SLC's we provide a more secure supply with CVT's, UPS's etc.

Our real PLC's :) PLC5's just keep going, and going ,and going!!!

Neil Firmin
Yallourn Energy
 
Hello All,

Earlier today Ken Roach posted the following:
>...snip
>when the 24 VDC onboard supply (either
>through the backplane connectors or the screw terminals above the AC line
>terminals) gets shorted, the SLC goes to default because it cannot control
>it's I/O if it doesn't have solid 24 VDC on the backplane.
>...snip

Well I found this quite interesting and had to try it. Guess what he wasn't kidding. If you short the 24Vdc terminals on the front of the powersupply the processor is faulted and the program is wiped. The processor is returned to its default state.

WARNING
For all of you out there wiring your 1761-NET-AIC devices to these terminals to interface to a DH485 port on a SLC-500/01/02/03 processor beware. If you short the terminals the program in the processor is lost.
WARNING

Bradley G. Hite
Intertech Incorporated
mailto:[email protected]
http://www.myplc.comTeaching Practical Skills for a Technological World
 
M

Michael Griffin

I haven't seen this with an AB SLC, but I have seen a similar problem with another brand (Siemens) when the power didn't go off and on
cleanly. The problem seemed to be an erratic power recovery would cause it to lose its memory. This was a difficult problem to understand because only one (S5-115U 945) out of more than a hundred Siemens PLCs (mainly smaller CPUs) was doing this.
We put an on-delay timing relay in the AC supply to the CPU power supply. If AC power was lost momentarily, the relay would stay off until
power had recovered for several seconds (at least a 5 second delay). We installed this a couple of years ago, and have not had a similar problem
since. This was much easier and safer than a UPS.

I suspect that any battery backed RAM is subject to similar problems under the right (or wrong) circumstances. Some of the newer PLCs have
eliminated the need for batteries. So far I haven't seen any problems with these ones, although I must admit that problems with battery back-ups are pretty rare.



**********************
Michael Griffin
London, Ont. Canada
**********************
 
D

David Lawton Mars

Niall
The only time I've come across this problem is when the backup battery is either disconnected or flat. I'd check the battery is ok, maybe even replace it as a precaution considering they're inexpensive. Apart from that I don't know of any other problems with the 5/05 unless another member of the automation group can shed any further light on this.
Regards
David
 
My observation rather I will say it is a horrible
problem getting RAM along with EEPROM losing their programms.I am facing this with 70 slcs running from past 4-5 years & have done all & all
whatever is necessary.

I do not understand why there is NOT a provision of protecting application software from external
power problems as said by somebody from AB.
Since the reason being given as the processor is loaded with default programe in case of application programme is changed.

Why does application prograame should change because of power problems? Should not be there
enough protection built in into SLC?
 
S

Steve Myres, PE

I've never had this occur with a SLC, although I built a small machine a couple of years back with a Micro 1000, and it seemed to fault periodically, although not with a loss of program. Unfortunately, the fault diags in the Micro are not all that great, but when we finally did find the problem, it turned out to be an occasional short on the 24VDC line (used throughout the machine for input power).
 
J

Joe Jansen/ENGR/HQ/KEMET/US

This is not entirely true.

If you have an EEPROM installed and *properly* configured, it will load the EEPROM program, and will carry on without a problem.

There are obviously filters on the power supply, etc. The situation that Ken described is caused by the *user* drawing too much current from the
24VDC power supply. If this is a problem, buy an external power supply.

Also obviously, at some point, you have to have an action if the power drops to a point that memory corruption -could- occur. You can not expect the power supply to be able to filter absolutely every possible power fluctuation perfectly, and still have an economic unit. At some point, you have to recognize that corruption may occur. If that point is reached, I
am much happier having the SLC dump the program, or load it from EEPROM, than have it try to execute who-knows-what out of potentially corrupted memory.

If you are having a lot of difficulty with this, I would suggest that you:

Check your grounding. This is the single most common source of problems in this area.

Check your loads on your power supply. Buy an external supply. In my practice, I never use the 24VDC supply off the plc supply for anything
beyond very simple testing. And NEVER in production.

If you are having your EEPROMs dump out with the PLC RAM, you're not configuring them properly. Make sure that you go into program mode and
tell it to transfer the code to the EEPROM. Make sure it is flagged to load from EEPROM either on memory error, or on power cycle if you are real
paranoid.

I have not had any problems with this that needed anything beyond grounding properly. I don't use UPS's on my power or any exotic filtering, just
mainline power thru a transformer to get 120VAC. I put EEPROMs into systems that are critical, or have a lot of power cycling. I set it up to
load on fault, so if it is loading, I don't even have to know about it.

HTH

--Joe Jansen
 
Hi

Application programs in EEPROMs get clobbered when the design of the hardware does not block the spikes or surges or the electrical "noise" in
the power line. This happens quite often with controllers that are fitted in hospital environments, textile plants etc, where high harmonic content due to high energy, high frequency systems exist in the vicinity and are
also fed power from the same source as that feeding to the SLC/PLC whatever. PLCs are normally more robust, I have found.

I have soved problems with a cent worth of silicon (Tranisl diodes or similar) in several cases (including a carrier Chiller plant which
required the EEPROM to be changed every 6-8 weeks or so, like clock work!)

Check out the plant power and the other connected loads (VFDs?) Your answers and solutions lie there.

Suresh
ICON Technologies pvt ltd, Chennai, India
[email protected]
 
J
everybody thal sells small PLC's gives you a little 24vdc power as a "user power source", but they're only kidding. ALWAYS use a separate 24vdc power supply for i/o. this will eliminate hours of heartache and suffering. A field wiring short will crash your PLC. Who wants that? It's only about a hundred bucks for a good, din rail mounted, easy to install power source. everybody sells 'em now. Use one!
 
I had some SLCs at another plant that were constantly dropping memory everytime the power was cycled. I checked with AB and they said to check the grounding, as a poor grounding system could cause this problem. The tech there insisted that the grounding was OK, but when we moved the machinery to another location it was discovered that all the grounds were conduit grounds. We reinstalled the machines exactly as they were except that we ran ground wires instead of relying on a conduit ground. Two years later, we have never dropped the program. No other items on the system were changed during the move.
 
Top