TVS’s, MOV’s, or RC Snubbers?

Long story made short- Purchased a hydraulic power unit for an automated manufacturing machine we built. This unit is controlled with an AB-MicroLogix 1000 (AC-Relay out) PLC. The unit did not come with Isolation relays in its control panel. The PLC’s AC Relay outputs directly power 6 hyd-valve solenoids, that fire about 20 times per min each. These solenoids state the following:



78VA



268VA inrush.



120volt.



Now the problem, we are 2 days out of warranty on the PLC (which 1/3 of which was in the power units builders shop during spec-ing and construction). We have “Burnt Out” 3 of its outputs. By that I mean that the PLC is telling its relay to break, but they are “sticking” on and continue to actuate the coils inappropriately. I can tap the side of the PLC with the butt end of a screwdriver and it will unlatch the output. We are also having a problem with failing prox-sensors and cylinder position sensors. These sensors came with the Neumatics brand cylinders that we installed on our machine. After failing, we are replacing them with Turk brand (which seems to be holding up better).



So far my investigation on our difficulties have lead me to believe that we need to install isolation relays that the PLC will actuate in order to free the relay output from the inrush (make or break) current. That is no big deal. Here is where I’m confused. After contacting numerous manufactures for advise as to which of the three inductive protection devices I should install on the coils and to what values I should choose, not one could enlighten me. My interest is to not only shunt the transient voltage that is present in “our” control panel, but to protect the sensitive electronic devices from the inrush (make or break) currents that I believe is also present. The machine's main control panel was built wrong to begin with. The builder ran control voltage wiring along with/parallel to/ and in the same pan-duit as the higher voltage conductors also present in the panel. Can someone afford a little education on TVS’s, MOV’s, and RC Snubbers, or have I missed the point entirely?

Thanks- barry!

Here`s some info. LINK

A relay won't last very long in this application, even if you add an isolation relay. The Micrologix's are cheap enough - replace them with units with Triac outputs.
MOV's won't supress the arc very well. Snubbers would work. TVS's are faster than MOV's, but I think they're DC, so you need 2.
The main problem is the frequent relay action. Even if you supress the spark that's welding your contacts, the mechanics will still wear out.
You need solid state.

Browse around the Crouzet site for some info on Snubbers, etc. Start here:

http://solid-state-relay.crouzet-usa.com/kb/ssr_kb_transient.shtml

You might want to consider solid state relays if you are firing each solenoid 20 times per minute.

Looks like I was beaten to it, I agree with keithkyll that you should be using triac outputs also. As stated, it is the frequency of operation that is your main concern here.

I can't see how AC solenoid switching would mess up the sensors, I would suspect the two problems are not related.

TVS's are nothing more than a really fast high-surge zener diode. They are also available in back-to-back configurations for AC applications, which is absolutely identical to soldering two one-way TVS's back-to-back (or front-to-front, it really doesn't matter...). But if your spikes are pretty powerful, i.e. more of a big surge than a spike, then TVS's die a quick death unless there is a bit of front-end impedence to limit the surge current.

Are you sure it's a spike/noise problem? At 20 cycles per minute and at 8 hours per day, you'll hit 1M operations in only 100 days --buh-bye relay contact (sniff...). Triac PLC outputs for sure, or SS relays.

Or stop thinking like the rest of us electrical types, and fix the problem at the dirty end instead. On a recent machine I used an OTS hydraulic valve driver (Parker TOC8) instead of cooking up my own driver scheme, partly to avoid these problems you're having and partly to get little extras like soft start & stop. Needs fancy valves too. Pretty pricey, but it saved lots of headaches and the machine runs great.

Fredderf said:

I can't see how AC solenoid switching would mess up the sensors, I would suspect the two problems are not related.

Click to expand...

Coils are 120vac at about 3/4 amp. When the relay opens, that's a serious energy spike with no place to go. AC and DC wiring in the same raceway. Surge voltages close to 400V. Probably Hall Effect sensors. All you need is the slightest knick in the insulation.

When a solid state device takes a spike, it might not fail immediately. A spike can cause a pinhole in the substrate. Over time, migration causes the device to fail.
After he fixes the relay problem, sensors will still fail for several weeks until all of the 'margined' devices have failed.

That's why nobody will repair a TV hit by lightning. After you fix it, there's good chance other parts will fail later. Impossible to warranty.

My point for this forum - After fixing something hit by a spike or surge, be aware that more failures are likely later. You can relax after 2 weeks or so. Allow for this if you're giving a warranty.

Oh, that's right. I missed the bit about the solenoid & sense cables beside one another. But the energy induced by capacitave coupling is usually hardly enough to cause a victimized sensor to die. A false reading perhaps, but death no.

BDKuhns, are your sensor readings making mistakes or are the sensors actually going dead? If it's just erroneous readings you're getting, I've seen this solved on open collector NPN sensors before with strong pullups on the PLC input side to increase the SNR (more current flowing in sense line, thus less effect from cross-coupled noise), or with simple filter caps at the PLC input to get rid of little spikes. Or soft-start those solenoids...

All commercial sensors I've seen are good for at least a human-body-model zap, which is more energy (and a higher voltage) than this scenario will generate on the sensor wires, me'thinks.

But I agree with you keithkyll, that when something does get hit with a spike that quite often it's invisibly damaged inside, only to completely die some time later. (It's always a struggle to convince people of this and to use proper ESD handling procedures, isn't it? I keep a set of pics of partly-blown semiconductor material handy for such purposes.)

Fredderf said:

Oh, that's right. I missed the bit about the solenoid & sense cables beside one another. But the energy induced by capacitave coupling is usually hardly enough to cause a victimized sensor to die. A false reading perhaps, but death no.

Click to expand...

Agreed. My thinking is 'insulation breakdown'.I suspect the worst. 150V wire on DC, and on the AC lines. The spike needs to be enough to blow a sensor, but not enough to toast a PLC input (no mention of failing input ports).

More likely, the sensors are reed switches rather than solid state. At 20 times/min., it won't take long to get some sticky reeds.

On a real stretch, capacitive coupling could be causing the reeds to arc and weld. The buffering on the inputs circuits may prevent this.

Thanks for the posts guys! This is helping quite a bit. Some of my problem here is to convince the other "powers to be" (who know nothing about this), as to what the problem is and how to fix it.

The output relays in the plc are actually welding closed. We have lost 3 already. The unit was obviously under engineered for the task that we are asking it to do. I thought about placing ssr's inline from the plc to the valves and placing a snubber on each coil. Problem is, I thought I read somewhere that you cannot use a snubber (RC device) in a solid-state circuit due to current leakage. If this is correct, would bi-directional TVS’s work on the valves if I were to replace the PLC that has SSR outputs? Another question is, am I correct in my thinking that TVS’s are primarily for Voltage surges and snubber are for Current surges? I am also under the understanding that a MOV is usually a 1 or 2 shot device. I need something with a longer service life. I intend to incorporate what ever we land on here into all the panels that I build.



The sensors that I am having problems with are magnetic field sensing switches. These are mounted on the out side of the cylinder for extend/retract position sensing. When these go out, they are either in the OFF or On state and no effect when I place a magnet near it. I'm replacing them with a different brand as they fail and have about 15 more to go. I'm starting to think that the brand that came with these cylinders are just cheap because I am not have the same problem with the Turks that I'm installing. If the problem was from a prior "major surge event", wouldn’t I be seeing difficulty in other areas?

Thanks!

Wow, this sounds like a major nightmare. If the valve coils are sharing the same power supply with the prox's I'd be tempeted to seperate them, maybe an additional transformer for the valves.

I would also tend to agree with using solid state relays.

Here's a link on relay contact arc suppression:

http://relays.tycoelectronics.com/app_pdfs/13c3236.pdfhttp://www.industrologic.com/mechrela.htm

BDKuhns said:

I thought about placing ssr's inline from the plc to the valves and placing a snubber on each coil.

Click to expand...

Yes. Most SSR's will have internal snubbers.

BDKuhns said:

Problem is, I thought I read somewhere that you cannot use a snubber (RC device) in a solid-state circuit due to current leakage. If this is correct,

Click to expand...

No. 2-4ma may keep unloaded inputs active. Ignore on outputs.
Leakage is enough to cause a minor shock. Be aware during 'hot' work.

BDKuhns said:

would bi-directional TVS’s work on the valves if I were to replace the PLC that has SSR outputs?

Click to expand...

Yes, but PLC will have snubber built-in on Triac outputs.

BDKuhns said:

Another question is, am I correct in my thinking that TVS’s are primarily for Voltage surges and snubber are for Current surges?

Click to expand...

In this application, they will both do the job. TVS is faster.

BDKuhns said:

I am also under the understanding that a MOV is usually a 1 or 2 shot device.

Click to expand...

Only if undersized.

BDKuhns said:

I need something with a longer service life. I intend to incorporate what ever we land on here into all the panels that I build.

Click to expand...

Design with all 24DC to avoid these nightmares.
Your AC coils are 3/4 amp. Maybe 2 amps for DC version. With DC, all you need is a diode on the coil. Then relay output would be acceptable. In future designs, Solid State is the ONLY sensible way to drive outputs at 20 times per minute.

BDKuhns said:

The sensors that I am having problems with are magnetic field sensing switches. These are mounted on the out side of the cylinder for extend/retract position sensing. When these go out, they are either in the OFF or On state and no effect when I place a magnet near it. I'm replacing them with a different brand as they fail and have about 15 more to go. I'm starting to think that the brand that came with these cylinders are just cheap because I am not have the same problem with the Turks that I'm installing. If the problem was from a prior "major surge event", wouldn’t I be seeing difficulty in other areas?

Click to expand...

Not necessarily. Probably grounded at one end, making them the path to ground for the back EMF of the coils.
Hall sensors are sensitive to spikes. Reeds can weld, or blow open if the spike is severe enough. I suspect the weakest link in this case. Difference in brands is spike protection.
Keith Kyllonen

Hey, does AB offer something like this?? (I would presume so..).

Hi current AC output with a built in snubber...

http://web2.automationdirect.com/static/specs/f208ta.pdf

Edit: I checked AB's website, it seems the best they have to offer in AC outputs would be 1762-OA8. .5 amps per point. 1 amp per common. I Hope you have small coils...

I try when ever possible to locate my MOV's at the coil. Right on the valve if I can. That eliminates noise and spikes between wires in the conduits comming back to the control panel. I also put them on all coils including the smallest relay coil. Unfortunately, that won't help the mechanical wear that many have already mentioned here from the 20 times a minute turn on. Solid State outputs is definately the way to go. Even with surge supressors built in to your output card, putting them right on the coil is still a good idea.

Locating at the coil is the best method. Very important. Thanks, Logix. Also true with DC and diodes.
elevmike, the coils are .75 amps. Micrologix has 1 amp triacs, but I think he needs to go to a L32 model to get triacs. The small PLC's are relay only. Staying with the Micrologix line, the other option is MOSFET outputs driving SSR's.
He's probably stuck with relays. Supressing the back EMF at the coil and living with short relay life is the only easy out.