Modicon 984 Frequent CPU Failure


Thread Starter



I am presently connected with a Diesel Power Plant here in the Philippines. I would like to get insights/thoughts as to how we are
going to minimize/eliminate the PLC CPU failure that we are encountering. We had an average of 7 CPU failures per year. These PLC's are Modicon 984 which are responsible for Diesel Engine monitoring of critical parameters like temp, pressure, etc. The plant is a floating power barge. These CPU's are mounted inside a panel in the engine rooms which has an ave. ambient temp. of about 40 C.

I'd suggest a grounding problem - possibly thru your data-comm ports. Most PLC like to claim thier comm port are "isolated", but if you take an
ohm-meter you'll learn that they really isolate the power entering the PLC, so the comm port shares the "logic" ground with the PLC - and any other comm ports. So if you have 2 comm ports, your PLC may be acting as a fuse between the 2 remote grounds used by the 2 attached devices.

This is greatly increased in power plants, generator systems, and if your I/O is not the isolated type.


1) make sure all I/O is 100% isolated (over 1 meg ohm) from the CPU logic power
2) add RS-232 isolators to any Modbus comm ports.

40 DegC shouldn't be a problem - I spent 12 years in Singapore and worked on several 984's which always ran in 34-38 DegC.


Lynn August Linse, Senior Product Application Engineer
15353 Barranca Parkway, Lantronix Inc, Irvine CA 92618
[email protected]
Tel: (949)450-7272 Fax: (949)453-7132

R A Peterson

You will need to define the failures better. Are you talking about memory losses, CPU replacements, CPU leaves run mode but can be sent back to run, etc?

40C is pretty warm. I'd be nervous that the spot temp exceeds the rated temp of the CPU. Could also be nasty power coming in.

984s are pretty reliable once you get past the infant mortality stage so I suspect you have a problem there that is killing them.
Experience with over 75 984 type PLC's in our plant yielded failures almost always caused by poor grounding practices. On your barge you
might be experiencing this. Contact Modicon for technical advice on how to test this possibility.

Our Quantum's have failed only due to being zapped by dirty power and when placed on a power conditioning UPS's they virtually eliminated failures. Good luck.

The plant is a floating power barge. These CPU's are mounted inside a panel in the engine rooms
which has an ave. ambient temp. of about 40 C.
Hi Lito:

Well, the first thing I would look at is finding some way to lower the ambient temperature, probably by using a cabinet air-conditioner. I
know the 984's are environment rated to 60 C (per the Jan 98 Modicon Specifiers guide), but heat is ALWAYS your enemy in electronics. Get that temperature down.

The other thing I would look at is AC line imbalance and/or grounding problems. 7 PLC's a year is HUGE. I would check for serious ground noise, ground current, and even if there is a
potential difference between your ground and your neutral line (assuming a grounded system). Another thing that wouldn't hurt to check is for proper isolation if you have different power zones. In other words, if you get all your AC from the same source (transformer), then don't worry about this. Otherwise, check for
differences in the voltage on the various high sides' of the line, the neutral sides, *AND* the grounds. Not at all uncommon to have different voltage levels in this situation. This can cause all sorts of neat things to happen to electronic circuitry

Ron Gage - Saginaw, MI
([email protected])

Phillip O'Meley

I may have a solution. What is the power supply to the PLC. i.e. AC or DC, what is its source, generator, battery etc., ?
If you can answer the above we can go forward from there.

[email protected]
I have several customers using the 984 processors for years without fail, this is not a usual situation. What are the Failures you are having? My first guess would be a power quality issue.

Darold Woodward

The prolonged heat and vibration is probably the problem. A popular way to perform highly-accelerated-lifecycle-testing is to put equipment on a shake table mounted inside a temperature chamber and subject the equipment to vibration and heat cycles. It sounds like you've built a pretty good HALT chamber.

My understanding is that there are some extended temperature rated compact 984 products used in the RTU-PAK products.

It may be possible that you can swap the CPU out for a Remote I/O communication module that is simpler and more rugged. You can then mount the CPU in a more controlled environment.

You could also put a cooler or heat exchanger of some type on the panel to keep the internal temperature down. Hoffman makes cooling products that you can add to your panels.

I'd like to hear how this gets resolved. If you get the opportunity to let me know or post to the list it would be great.

Darold Woodward PE
SEL Inc.
[email protected]
Please ensure that :
Whether you are supplying noise free power supply. Unforthunately DG Engine power is very noisy. So please ensure that you
* filter at first stage at AC inlet using a good EMI/RFI filter
* Use A good regulated DC power output supplier to the system with excellent input and output isolation.
* use once again a good RFI Filter at the DC side. Use MOV to clamp surges in parallel and NTC s in series to absorb hiccups.
* There has to be a good electrolytic capacitors at the DC LOAD followed by smaller ones sprinkled throughout your board.

Please ensure that that loads which you are switching have good isolation with your system. ie. preferably optocoupled. etc.
* Check out if there is overloading on CPU outputs that leads to partial failure of CPU drivers leading to full death ultimately.

* check all the instrumentation parameters fed have unloaded and isolated interface with high impedence.

Let me know in detail about your problem so that I can discuss in detail.

[email protected]
Ashok Gupta
Research Engineer
II Floor
#39, Main Pusa Road
N. Delhi 110 005