Which power device must be put in compression for out of circuit testing.
I read it in a book and I think it is a diac or triac but I can not remember and I can not find the book THANKS HYPRESS

the diac because it's bidirectional?

Which power device must be put in compression...

I've never heard of the term "compression". Please explain!

-Hal

It could be any power device, if the packaging requires it.

There is a type of packaging used for extremely high power devices called appropriately 'hockey puck' packaging. The device is built as a squat cylinder with the two faces each being electrodes/heat sink connections.

Additional electrodes are made available at the side of the devices.

I've only seen SCRs and vacuum tubes packaged this way, but anything is possible.

See figure 3 here: http://www.darrahelectric.com/trouble.htm

-Jon

I don't understand what sort of application you have in mind here. Could you explain in a little more detail?

Anyhow..... A triac is a 3-terminal device which behaves like a bi-directional thyristor. The thyristor is triggered into conduction when a pulse is applied to its gate terminal and will continue to conduct until the main current falls to zero, at which point it will not conduct again until re-triggered.

The triac works in a similar way, but allows current to flow in both directions, and it is therefore commonly used in AC power applications (where it will shut off at each zero-crossing point of the waveform until re-triggered on the next half-cycle).

The diac is a 2-terminal device which is basically a pair of parallel, back-toback diodes. The diac is frequently used in combination with a triac in AC power controllers. The construction of the triac means that it is generally more sensitive to trigger pulses of one polarity than the other. By feeding the trigger signal through a diac first, the response of the triac can be more closely matched between positive and negative half cycles, resulting in more symmetrical waveform on the output.

I think it would be the diac. (Just covered it in school <2 weeks ago )

It has a high (20-30V) 'breakover' voltage in both directions, so testing it with a meter can only verify that it is non-conducting (possibly good) or shorted (not good).

This would be any device in the "hockey puck" style packages, mostly used on SCRs and GTO thyristors.

The construction of this device relies on external mechanical force to make the electrical connections between the metal "slugs" of the package and the internal semiconductor die. This force (typically hundreds of pounds!) is normally provided by the heatsink assembly that the device is clamped into. Unless the device is suficiently compressed, it may test as "open", because the connection between the contacts and the die is lost.

I'd agree with the 'hockey pucks'. We used to have several variable speed drives in the plant that used them. One of my jobs on the annual PM was to use a calibrated gauge to make sure the bars holding them in place were tight enough that they bowed the required distance. I was always just a bit netvous when I had to troubleshoot around those things. I'm sure 'letting the smoke out' of one would not be a good experience.

Yeah, failure of one of those devices is seldom a good thing. They are EXPENSIVE items, and in the usual power converter bridge setup, you rarely have just ONE of them fail at a time.

Replacing them requires meticulous attention to the mounting arrangement, and replacement of all the hardware EXACTLY as removed, particularly all the insulating sleeves on the bolts, and the belleville washers or other tensioning devices. A torque wrench is MANDATORY here for most equipment.

Some power semiconductors pose a hazardous materials issue if they have failed in an "explosive" manner, as is common at these power levels. The ceramic insulation in some devices is beryllium oxide (BeO), which is VERY toxic if ingested or the dust inhaled. If the package has been breached or damaged, gloves and a dust mask are highly recommended as a precaution.