GMAW-failure in the root

[saipemsmspetti]

We are welding an offshore pipeline with Auto-GMAW process using copper shoes for backing during root pass.
We have a problem with root pass that ocurs like concavity spot 5 mm in dia. In front and after fault weld is completely sound. What could be the reason for such kind of fault? Is it some kind of impurity or discontinuity, but which one?
Welders said that they feel it like sudden suck up. We checked magnetism and it is in the expected range lower then 20 Gauss. Welding data:
Base material: X60
WT 16 mm
Gas: 100% CO2 for the root pass, 35 l/min
Wire: Autrod 12.66 dia 1.0
(270-310A and 28-30V, v=85 cm/min )
Bevel: U with nose 1.5 mm and 60 deg 1.4 mm opening from the back.

 

[metengr]

saipemsmspetti;
It could be a low spot along the pipe root surface.

 

[saipemsmspetti]

We can check visualy both ends of the pipe before fit up and didn't discover any low spots. In the meantime we decreased voltage on the source for 1.5 volts and now we still notice some spot fromtime to time but with less depth from botom side. Now it is acceptable concavity in the root. But it is still mistery for me what could be the reason for that.

 

[stanweld]

Are you using constant current or constant voltage power supplies?

 

[saipemsmspetti]

We use the constant voltage power supplies and yesterday I decreased voltage on the power sources for 1 V and advised welders to increase the bug trevel speed (welding speed) in the area 3-5 o'clock and 9-7 o'clock (two welder operators CW - CCW) but did not solve the problem. We have 5 or more joints in a row completely sound and then these spots appear (1 or 2) but fortunately it seams on radiography shot like no deep concavity which is accepted by inspector. The spots are 5-6 mm dia with regular circle shape.

 

[stanweld]

Since the spots are irregularly intermittant, I suspect that the problem is a slight fluctuation in current due to to minor changes in electrode stickout. there may also be slight fluctuations in your incoming power to the power supply.

 

[GRoberts]

Or a slight interruption of the wire feed.

 

[saipemsmspetti]

Thanks to everybody for participation in solving my problem and helping me to search in the right direction. Just for your information problem is solved with decreasing of current in hot pass through decreasing of wire feed speed from 11 to 10.5 m/min. Process is now stabilised with no more cavities in the root pass.