Capacitive Coupling ???...

[SG-1]

Only one timer out of the 8 in this order malfunction in this manner. There are two True Time Delay timers in this order with the .1 to 10 sec range. Neither will work in that circuit without help, the resistor.

I am betting a resistor will be wired across it's voltage input " coil " this time.

One contact in the other timer stopped working without cause. It was picking up an industral control relay.

The engineer had some concern with the realiability of the relays before testing began.

 

[SG-1]

100315-0807 EST

Smart $:

I am not clear on what was his bench test.

Without doing any calculations I believe I could use a GE RR series relay, a several thousand mfd capacitor at 30 V DC, some electronics, and be able to work to a limit of 15 minutes.

GE RR relays are a mechanical snap blade bi-stable relay.

.

Set timer to .1 seconds.

I connected the timer to a varable power supply preset to 120VAC.

I connected a bell to one set of contacts.

I connected my Fluke 27 in series with the timers coil to monitor the current.

I turned on a breaker to energize.

I rapidly decreased the voltage by spinning a knob.

The relay remained latched.

Repeated process with the second timer with the same results.

I believe experiments should be repeatable. They should all fail in the same way :grin:

Feel free to ask questions. I will replay as I can.

GE is out of the question. Old Westinghouse people here. That is like putting the cross in Dracula's face.

 

[gar]

100315-1629 EST

SG-1:

Your test of quickly turning down the voltage with a Variac may not represent what Magnacraft expects the input to the relay to do. Namely abruptly turn off.

On the other hand I can see a basic design flaw in their design from what you have described. This is a 120 V relay and I expect it should not start timing until the input drops to maybe 60 V, 1/2 nominal rating. This would be a design criteria I might use to avoid false triggering of the timer with droops in voltage. But, also I should expect that it might start timing at about the 1/2 point to avoid the noise problem you are encountering where it does not drop out at all. At half voltage I can design the device to have enough energy in the storage capacitor to be able to change the relay to its normally off state at a time delay as long as 15 minutes.

If you have a bad relay you might take it apart and analyze the circuit and see if you can determine what design defects might exist.

Many years ago we used a P&B time delay relay to supply an initial trigger pulse at the start of a machine cycle. The machine produced a lot of impact shock that transferred to the electrical cabinet on the rear of the machine. These relays consistently failed after about 3 months. The reason was P&B used a small electrolytic capacitor mounted with two leads from one end. These wires fatigue failed and the capacitor broke off. The solution was to RTV the capacitor in place. I might point out I never design with this type of capacitor. I always use ones with axial leads.

Do any of the relays fail to time out if you use a switch to control the coil and there is no residual output voltage from the switch in the switch's off state?

If the relays operate properly with a discrete on-off signal to terminals A and B and your system does not produce a slowly dropping voltage to A and B, then I do not think the slowly varying test is useful for qualification.

Does AB still make a pneumatic off delay relay?

.

 

[Smart $]

...
The engineer had some concern with the realiability of the relays before testing began.

His concern seems justified

So are you and/or the engineer now considering a replacement?

As ELA asked, in essence, does the relay need to be true off? It needs to be if the relay has to time out even in the event of power failure. As such, it limits the viable replacement options.

 

[SG-1]

Gar, the bench test replicates the failure seen in the circuit. A fair test for the timer ?... The circuit tested it first. :grin:

I agree the manufacturer never expected this condition.

Smart $, the engineer will talk with the customers rep. in the morning. We cannot change anything without their approval.

The 15min timer is the spin down timer for the motor so no re-start can be attempted until it times out. It is a TDPU Time Delay Pick Up. If the AC fails then all 3 breakers will be open by redundent means.

The True Drop Out is is more critical in that is is part of the incomplete sequence. It function is to open the START Breaker if the RUN Breaker does not close.

The AC is usually supplied by the sola transformer that powers the exciter field. This time the AC is a remote customer source. UPS I hope.

 

[Smart $]

...Smart $, the engineer will talk with the customers rep. in the morning. We cannot change anything without their approval.

I felt that one coming

The 15min timer is the spin down timer for the motor so no re-start can be attempted until it times out. It is a TDPU Time Delay Pick Up. If the AC fails then all 3 breakers will be open by redundent means.

The True Drop Out is is more critical in that is is part of the incomplete sequence. It function is to open the START Breaker if the RUN Breaker does not close.

The AC is usually supplied by the sola transformer that powers the exciter field. This time the AC is a remote customer source. UPS I hope.

This means little to me without the grand picture. Sure I can speculate... but what good is that??? Even now that I know a liittle more, it still does not answer the question if timing out of the relay in question is necessary on power failure? If yes, the relay hass to be a true-off type. If not, it can be of the "non-true" type, with slightly different wiring of course. Would need a lot more info if there was a UPS in the system... but speculation says it could be the the "non-true" type.

 

[gar]

100315-1854 EST

SG-1:

Your circuit actual circuit did not test the relay in the manner that I believe Magnacraft would expect. In other words my guess is that Magnacraft expects the input to A and B to abruptly fall from about 120 V to 0 V, but your circuit only drops to 10 V or so, about 14 V peak. and never gets to 0 because of some leakage.

If their trigger threshold to start timing is not up where I would put it, 60 V, but is maybe 10 V, then I would expect the failure. Quite often the design criteria for the threshold between logic 0 and 1 is put at about mid supply voltage for maximum reliability.

.

 

[SG-1]

100315-1854 EST

SG-1:

Your circuit actual circuit did not test the relay in the manner that I believe Magnacraft would expect. In other words my guess is that Magnacraft expects the input to A and B to abruptly fall from about 120 V to 0 V, but your circuit only drops to 10 V or so, about 14 V peak. and never gets to 0 because of some leakage.

If their trigger threshold to start timing is not up where I would put it, 60 V, but is maybe 10 V, then I would expect the failure. Quite often the design criteria for the threshold between logic 0 and 1 is put at about mid supply voltage for maximum reliability.

.

When the voltage was lowered slowly, the timer would drop out around 50 volts. Pick-up was about 60 volts. As you expected. When the voltage dropped abruptly or very slowly the timer operated like a champ. I needed to know if it could be made to malfunction.

I would estimate I lowered the voltage from 120 to 20v or 30v in about one second on the bench to cause the malfunction. My bench has KC-241 type meters. ( The meters might be older than I am. ) One would not expect the rate of rise or fall to adversly effect the device to the point of it locking up.

 

[gar]

100315-2032 EST

SG-1:

Your normal circuit opens and closes abruptly. The time delay relay works correctly with this type of input. Your problem was a residual voltage, probably capacitive or resistive leakage. Resistive loading of the open circuit wiring apparently reduces this residual voltage to an acceptable level.

Thus, it appears that by correcting the residual voltage problem that erratic operation of the time delay relay is eliminated. Thus, you do not have a relay problem if operated like I believe Magnacraft expects it to be operated.

I think there is some weird characteristic to the relay operation at low partial voltage, and it would be interesting to know why. But, this would require analyzing the internal circuit if possible.

.

 

[SG-1]

I felt that one coming


This means little to me without the grand picture. Sure I can speculate... but what good is that??? Even now that I know a liittle more, it still does not answer the question if timing out of the relay in question is necessary on power failure? If yes, the relay hass to be a true-off type. If not, it can be of the "non-true" type, with slightly different wiring of course. Would need a lot more info if there was a UPS in the system... but speculation says it could be the the "non-true" type.

I felt it too !! Should be interesting later today, after the dust settles.

IMHO the timer does not need to be a true time delay, because if the AC control power is lost everything will trip offline. I am not an engineer or a MV motor control technician, well I am this week.

 

[SG-1]

100315-2032 EST

SG-1:

Your normal circuit opens and closes abruptly. The time delay relay works correctly with this type of input. Your problem was a residual voltage, probably capacitive or resistive leakage. Resistive loading of the open circuit wiring apparently reduces this residual voltage to an acceptable level.

Thus, it appears that by correcting the residual voltage problem that erratic operation of the time delay relay is eliminated. Thus, you do not have a relay problem if operated like I believe Magnacraft expects it to be operated.

I think there is some weird characteristic to the relay operation at low partial voltage, and it would be interesting to know why. But, this would require analyzing the internal circuit if possible.

.

The relay with the bad contact is on it's way back to the vendor. So cracking that puppy open is no longer in the realm of possibilities, for now.

On the bench I actually lowered the voltage to zero. If it did not change on the way down it was latched until re-energized & abruptly de-energized.

I checked the circuit controlling it. I used my Fluke 27 & measured between the coil signal wire and every other wire across the top of that unit. Every resistance reading was off the scale. Checked over 60 wires. Then checked to ground, then to the the ungrounded conductor of the supply circuit. All were beyond the meters resistance scale. I visually traced the wiring to every terminal block all the way to both MOC switch points. I suppected cross circuiting, found nothing.

 

[gar]

100416-1846 EST

SG-1:

Since you see no leakage resistance between your time delay relay control wire and any other wires, then it is reasonable to assume that there is no resistive leakage component.

An experiment you can run on the bench, and then a similar one on your circuit.

Bench --- Need a 0.01 mfd 400 V paper or Mylar capacitor, a 0.1 mfd 400 V, and about 25,000 ohm 1 W resistor. I used a CD capacitor box for the 0.01 and a 100 V axial lead for the 0.1 because they were convenient. Two 10 k and a 5.6K 1/2 W in series for the resistor.

Connect the two capacitors in series and across a 120 V 60 Hz source. The voltage read across the 0.1 mfd was 12.6 V. Replace the 0.1 mfd with the 25 k resistor. The voltage across the resistor was 12 V.

Actual circuit --- In your actual circuit connect the 0.1 mfd in place of the relay at the relay location. Open the control line to the relay at the source, and measure the voltage across the 0.1 mfd capacitor. Assuming the leakage is capacitive coupled, then this voltage can be used to estimate the equivalent leakage capacitance assuming all the other wires in the bundle are at 120 V.

.

 

[SG-1]

Update

Update

Customer realized today that the 120 volt AC should be supplied by the sola constant voltage transformer, not a remote source. Wiring changes in progress to return circuit like it started out.

He will accept either a pull down resistor or use a timer with a 1 to 10 sec range that seems to be immune to this issue. The timer has been operated about 10 times with no failure. The 1-10 sec timer is perferred if it does not malfunction. They will test for this during start up.

I have to try it again after the sola is supplying the 120VAC.

 

[SG-1]

100416-1846 EST

SG-1:

Since you see no leakage resistance between your time delay relay control wire and any other wires, then it is reasonable to assume that there is no resistive leakage component.

An experiment you can run on the bench, and then a similar one on your circuit.

Bench --- Need a 0.01 mfd 400 V paper or Mylar capacitor, a 0.1 mfd 400 V, and about 25,000 ohm 1 W resistor. I used a CD capacitor box for the 0.01 and a 100 V axial lead for the 0.1 because they were convenient. Two 10 k and a 5.6K 1/2 W in series for the resistor.

Connect the two capacitors in series and across a 120 V 60 Hz source. The voltage read across the 0.1 mfd was 12.6 V. Replace the 0.1 mfd with the 25 k resistor. The voltage across the resistor was 12 V.

Actual circuit --- In your actual circuit connect the 0.1 mfd in place of the relay at the relay location. Open the control line to the relay at the source, and measure the voltage across the 0.1 mfd capacitor. Assuming the leakage is capacitive coupled, then this voltage can be used to estimate the equivalent leakage capacitance assuming all the other wires in the bundle are at 120 V.

.

I do not have the required capacitors to try the experiment. I do not think Radio Shack will have any with a 400V rating. I wll look through my stuff at home to see.

 

[gar]

100316-2257 EST

SG-1:

You could use a 200 V cap for the 0.01 mfd that connects to the 120 V supply. Your test does not need to be long and the likelihood of failure with a metalized Mylar would be very small under these conditions. The 0.1 can be a 100 V. If there is some failure the voltage on the 0.1 could get a lot larger.

With the 10 to 1 ratio of capacitance you get a output voltage about 1/10 the input.

Using an 0.1 cap for a test shunt in your actual circuit you will estimate the leakage capacitance approximately by (Voltage across the 0.1 cap) / (Source voltage) = (Leakage capacitance) / 0.1 . This fails as an estimate when the voltage across the shunt capacitor gets to higher voltages relative to the source voltage. At the 10% point it is sufficiently good for the purpose at hand. You can write the more accurate equation for any voltage ratios if you want.

If testing your actual circuit produces voltage across the 0.1 cap much above say 15 V, then make the 0.1 an appropriately larger capacitor. Preferably use only paper or Mylar capacitors. Mylar can be read to mean a number of other plastics like polystyrene, and polypropylene.

.