Thread: OT: How to choose proper varistor to protect bridge rectifier

Westom- Motor is not needed to stop load in any way. Did not explain myself very well. When we get a unit back with stuck relay the rectifier is also alway blown. Dead short accross AC side of diode which blows circuit breaker. Also in case of just stuck relay user can put control in opposite direction which again causes short and blows fuse. With that said still might make sense to have seperate power relay from reversing relay for both saftey and help isloate from voltage spikes.

The motor is connected to 50:1 worm gear reduction drive this is then attached to boat lift winch. The lift can have full load and you can remove the motor from the gear box and nothing will move. It is fail safe in that regard. We have tried rapid FWD to REV switching under load and have never had a unit fail. With that said when you do remove power from motor it does take maybe a second or two for it to come to a stop. So the idea of a voltage spike being sent back through the rectifier is very possible. Maybe some diodes are more susceptible to this type of damage.

Tofurther help in understanding things, would a 30 amp rectifier be able to survive a dead short on a circuit that is protected with a 20 amp circuit breaker? For example lets say a contact on the "up" relay sticks and then the operator pushs the "down" button. This will cause a short in our control box. Would it be expected in addition to blowing the circuit breaker it would also destroy the 30 amp rectifier.

Kilroy- I have not purchased scope and do not plan to purchase one. Why would I buy somthing I do not know how to use? Would you be willing teach me with out getting huffy? Doubt it. I like your idea of looking for sparks on the contacts. However the relay contacts are not visable through their housings. I will see if we can remove the housings or get a set that has visible contacts

Originally Posted by ORAGRAG

would a 30 amp rectifier be able to survive a dead short on a circuit that is protected with a 20 amp circuit breaker? For example lets say a contact on the "up" relay sticks and then the operator pushs the "down" button. This will cause a short in our control box. Would it be expected in addition to blowing the circuit breaker it would also destroy the 30 amp rectifier.

That's going to depend on the specific characteristics of your rectifier and your circuit breaker. It's entirely possible the answer would be "no, won't survive" or "yes, will survive". With the specific part numbers for the rectifier and fuse (so we can go get the manufacturer's detailed specs), we might be able to give you a specific answer. But frankly, one of your engineers could do that homework, too.

Let's go back to basics for a moment. Does your motor control have any form of motor switching transient protection? If not, that's probably your problem with field failures, rooted in a company failure to have someone with the right skills review the design.

Originally Posted by ORAGRAG

Dead short accross AC side of diode which blows circuit breaker.

First, a blown diode bride is acting as an emergency backup system. Similar to backup O-rings on the shuttle. Those too almost failed multiple times. But management (business school graduates) said a disaster did not happen. Therefore a disaster would never happen. Your reasoning implies similar reasoning.

As demonstrated by the original post. You wanted to "fix the backup o'rings" (diode bridge) rather than first identify and then solve the actual problem.

Suggested was how to expedite a solution obtained by an outside consultant. But to expedite that meant creating a reproducible problem. Or understand some basic concepts here.

To say more means also knowing numbers for those relay contacts. Your posts suggest it is not an isolated failure. It also should be treated as an "almost burned through secondary o'ring".

Above is an engineering concept that must be grasped to properly solve this and other problems.

Second, to know something means explaining what can happen AND test that it does. Without both theory and experimental confirmation, then only speculation exists. Does not matter what a motor reversing test with a boat demonstrated. You must also explain why that would not cause damage - ie a fused relay contact. And even with other variables such as the lower voltage apparently found in those customer venues. Is the motor and relay contacts generating only 950 volts? Then your tests did not see a failure that exists on all equipment. Concept taught over 50 years ago by Deming. And necessary for any engineering analysis.

A fused contact can be from also anything. A motor reversal was simply one of maybe 15 different examples I could have posted to demonstrate the larger point. You must explain why every action does not fuse relay contacts both by using numbers and by worst case testing. Concepts taught by William Edward Deming absolutely apply here. For example, it does not matter if all parts are in spec. How those parts remain in spec can expose a major defect. Which is why an oscilloscope is a cheap and simple tool.

Does a circuit breaker trip on 20 amps? Of course not. If a circuit breaker is doing protection, then these numbers must be known before you discuss it. 30 amps could flow through a 20 amp breaker for minutes before it thinks about tripping. 100 amps for a shorter time. Hundreds of amps could blow that diode bridge. And the breaker not trip. Circuit breaker may not be the protective device you have assumed if relevant numbers are unknown.

Oscilloscopes were used by every student in my high school science generations ago. I was using one in sixth grade. An o'scope is a dumb simple tool. Can be rented from a technical supply house. But this is the point. If you do not even understand basics tools, then your company desperately needs someone who does understand these simple concepts and tools.

I walked into an operation that sounds similar. They were lucky. I found reams of defects. Some that almost killed people. But they just knew their product was good. Because their testing never found any of those problems. They did not understand the above and critically important concepts.

kilroy's comments - although sharp - are the mindset necessary to only maintain sales of a product line (at minimum). Or to not kill someone (worst case). The only engineer that should be doing your job should have basic knowledge so that every failure is explained before fixing it. Only that means other existing problems are discovered and will not harm anyone.

BTW, chances are you cannot do anything about this. If the boss does not understand these basic concepts, well, you (or customers) may become frustrated. That was also taught by Deming.

We do have a rather heavy-handed person here......

Questions......

1) Is there EVER a blown fuse that does NOT also have a shorted rectifier?

2) What is the rating of each relay, what type is it..... number and mfgr if possible... how is it hooked up in circuit?

3) What is the motor rating?

4) What is the DC resistance of the motor?

5) What TYPE of motor is it.... I would assume (since it is chinese) that it is a permanent magnet DC motor.

6) WHICH relay is stuck? It sounded as if it is the AC side relay, before the bridge.

7) Is tehre a fuse? What is it? Where in the circuit is it?

8) Is there a power switch? is it before the rectifier?


The motor has a back EMF. If "plug reversed", the back EMF and the line voltage add up, and the max current is approximately double the line voltage divided by the circuit resistance. If that is larger than teh DC rating of the reversing relay, you have trouble.

I would recommend the interlock I suggested, because of the large currents from reversing under load. More current equals more energy in any arc or spike.

Some points:

A) diodes differ. Most are good significantly above their rating, especially when a 1000V type is used, as that is a typical highest rating for many series, including the one you mentioned. But a few will be right at the rating, which of course is still perfectly acceptable. Those obviously will fail more often.

B) conditions differ..... the engineering sample usually is better than most, courtesy of Mr Murphy. The lab power is better than site power (nearer to spec), site power is usually lower or higher in voltage than you would like.

C) Yes, no amount of quibbling will make the strong potential for spikes go away from the described system.....

D) Examination of failed parts will usually offer no particular info, unless you find seaweed, etc.... or the like. Once the diode has failed, it usually looks much the same regardless of reasons.... I have sent a number of parts back, and never once was that the defining piece of information.

E) As THOROUGH a description of what was going on when it failed is important...... was the boat within the weight limit? Had there been any storms? Is power usually low or high in voltage? Are there any other high current devices in the area, including up the road a mile or so? (manufacturing, welding, etc). Was any welding done recently at the site?

F) Unless the problem is REALLY bad, you normally will not see it in testing unless you do a LOT of testing at limit conditions, and with regular field conditions of temp, humidity, dirt, water, etc.

G) You normally need to test a number of samples That can take a long time, and cost money..... you need to determine if repairs are more or less costly than the effort of proving a fix. (sounds like a bean counter, and it should.... some things are not WORTH fixing, unless they impact reputation or bottom line, you will always have some failures).

H) Complaints notwithstanding, some simple cheap fixes can be applied, and should be, regardless of whether they can be proven, because the proof will cost a lot more than the fix, and the fix fixes known potential issues.

I) Fuses and semiconductors..... There are fuses made for that.... they may not be what you need. But they are fast enough to "save" the rectifier, etc, at the cost of a blown fuse and annoyance, not to mention that teh fuses are not cheap.

J) Circuit breakers will NOT save the rectifiers, in most cases. The "let-through energy" of teh protector must be less than the "failure energy" of the device to be reasonably certain the protector will protect. Breakers let through a lot of energy, even though they open pretty fast.

I have a much better understanding of what is involved here. Thanks everyone for the help.

Nick

There have been a number of posters want to turn a boat lift designer into an electronic expert. I am not sure this is practical, what is needed are a few principles that will ensure that the diodes have a long life and that there are no designed in short comings (or should I say not designed in safety features).
1. All brushed motors need to have suppressor devices fitted else they will radiate noise which according to most wiring codes is not permitted. Look inside any cheapo hair dryer/vacuum cleaner/washing machine/corded drill. . . .
2. If you use electronic devices they have to be protected from voltage spikes coming down the mains or from the switching transients caused by your own control equipment (because the DC motor is already suppressed!).
3. All kit should have a safety margin built into it, so components are not rated at their absolute maximum.
4. Fuses (or circuit breakers) should blow before too much damage is caused to the equipment (hopefully none).
So looking at these points wrt this specific problem:-
Suppressors MUST be fitted to the motor, It sounds that this size motor will require quite an expensive device. My AQUA-VAC (good'ol USA made vacuum cleaner ~ 30 years old) has a metal cased filter with the case connected to the motor chassis and to the earth lead. The L&N leads go through it out to the mains and the switch. This then keeps all the noise within the confines of the motor.
Now to protect the diodes a very small high voltage capacitor ACROSS each diode, diverts the incoming voltage spike into the more robust circuitry. Additional capacitors from L&N to earth would be a good belt and braces approach.
The diodes are adequately rated for this task as theoretically a 400V device should be enough. On current overload, the diodes will heat up but this takes time (2 secs?) and the fuses should blow first.
Frank

Originally Posted by chuckey

...On current overload, the diodes will heat up but this takes time (2 secs?) and the fuses should blow first.
Frank

Excellent post Frank.

Only issue I would take is above sentence. In all my years of working with power semiconductors, 99.999% of the time a CB provided no protection at all. Semiconductors seem to pop most often in usec, rather than seconds due to constant slight overload. We provide semiconductor fuses when we try to protect the semis and they work maybe 50% of the time at a cost for a 30 amp fast fuse of like $ 150 each or more.

Depends on the semiconductor......

Mosfet/IGBT usually 10 uSec "square load line"..... i.e. over 10 uSec at full current + full voltage and its dead. Only the most expensive fuses will potentially work, The fuse cost can be more than the repairs

Rectifier/SCR..... Survive 10X to 100X overloads, depending on time..... basically a heating issue. it is possible often to find a fuse to protect them pretty well.

What do you guys think of this idea. Bridge rectifiers are cheap so why not use two of them wired in opposite polarity to the motor. Then place the FWD & REV relays on the AC side of each diode. Then the bridge rectifier would never be disconnected from motor under power and diode would only have AC supplied to it when lift is in operation.

Originally Posted by ORAGRAG

What do you guys think of this idea. Bridge rectifiers are cheap so why not use two of them wired in opposite polarity to the motor. Then place the FWD & REV relays on the AC side of each diode. Then the bridge rectifier would never be disconnected from motor under power and diode would only have AC supplied to it when lift is in operation.

I think you will find that each will short out the other........

Simpler to avoid switching under load, with an interlock. Relay contacts will CARRY a lot more than they can safely make or break.