problem with triacs with neon transformer loads

Here's the original post:

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >Hi,

He has two triacs connected to one flasher channel. It worked for 4 months. Then one channel failed while the other continued working. Both worked when he took it to the lab. Returned to service, same problem after two weeks.

Do you think a ringing problem or dv/dt could give these symptoms? Regards,

Mike Monett

"Mike Monett"

** Absolute nonsense.

The inrush surge is caused by *core saturation* and consequent temporary LOSS of primary inductance.

The core saturates because of the DC component ( really a low frequency transient) generated by switching a sine wave.

** A rapid "fade up" using a triac with phase control is one way. ** Where it exists, it just makes the worst case inrush surge larger.

The peak surge current approaches the zero inductance situation in many cases.

So, I pk = V pk / R primary

Eg.

A 1 kVA transformer ( R primary =1 ohm) can be expected to exhibit 300 amp peak surges on a 230 volt supply.

......... Phil

Try swapping the connections to the triacs and see if the problem stays with the MOC3083 driver or BTA41 triac.

Here are the datasheets for the triacs:

http://www.ortodoxism.ro/datasheets/motorola/MOC3082.pdf

Did you know the optocoupler is a zero cross? Here's the description:

"The MOC3081, MOC3082 and MOC3083 devices consist of gallium arsenide infrared emitting diodes optically coupled to monolithic silicon detectors performing the function of Zero Voltage Crossing bilateral triac drivers."

In the datasheet, there is a statement near Figure 8:

"*For highly inductive loads (power factor < 0.5), change this value to 360 ohms."

but I can't see what it refers to. Maybe there is a way to make it switch at the voltage peak for inductive loads. This might help reduce the inrush current and reduce heating in the BTA41 triacs. Regards,

Mike Monett

Firstly as has been mentioned switch on at the AC maximum to reduce the inrush. Triacs are not good for this application since they go off at the current zero crossing point which is when the voltage is near maximum. This sudden application can cause the triac to conduct during the current "off" time so when the current tries to reverse you get conduction once again. Large snubbers can help out but its marginal at best, this is why sometimes it works and sometimes doesnt. The solution is to chuck out the triac and use two thyristors instead.

Yes:

The dv/dt problem prevents the traic from ever turning the transformer off. Every time it switches off, the dv/dt turns it right back on again. There may be some smallish change in the output voltage but the light would remain on.

The ringing can be passing a large current spike through the triac. The triac is now, in fact, damaged. When you damage a BJT or a triac, often the first sign is that the breakdown voltage is reduced. The effect is often progressive. This could cause a breakdown firing problem at high mains voltage.

--
--
kensmith@rahul.net   forging knowledge

So why does one channel fail in the field after some period of time, work in the lab, then work in the field but fail again after another delay? And the other channel apparently works fine.

That's what I was referring to when I asked if you thought a ringing problem or dv/dt could give those symptoms. Somehow it deoesn't seem possible that a ringing or dv/dt problem could heal itself, then fail again after a delay. Especially when you describe ringing as a progressive reduction in breakdown voltage.

Regards,

Mike Monett

Hi Look for SGS-Thomson application note "Control by a triac for an inductiv load" You will find the solution in using an alternistor instead of a triac for this purpose. Schematics included alternistor = TPDV 640 .... 1240 Ton Visch

snipped-for-privacy@aol.com wrote:

The OP hasn't stated whether it starts working again if he stands on his left foot or the neighbor turns on his toaster. I assume the circuit is right on the edge of the working/not working bondary.

--
--
kensmith@rahul.net   forging knowledge

Then any degredation should make it fail permanently.

Anyway, Fred thinks the post is a hoax.

Regards,

Mike Monett