Thank you very much Stephan you explained very clearly some mechanics I was always aware of but never fully understood in a technical sense.
I believe everything you said confirms what I've witnessed in my years of welding and after discussion I had come to a similar (although not nearly as well described conclusion). My fears of oxidation were mostly due to being unfamiliar with GMAW Stainless steel having only ever welded it in volume on the GTAW process which has a much smaller energy input per pass than GMAW and does not have the same surface oxidation on the bead.
However I believe your analysis, while correct, isn't perfectly describing the situation I've encountered. What I'm trying to say is I most likely didn't describe the situation in full detail, and you assumed several variables different than what actually happened.
To clarify what was said before:
- welds were made on identical material and joint detail, 304L 3/16 SS open root square groove weld
- Several welds were performed under Tri-mix at settings comfortable to the welder.
- Shielding gas was switched to 98-2 and the welder allowed to perform practice beads and adjust settings until comfortable at which point a more or less identical bead was run on several more identical test plates. The objective as given to the welder was to create similar quality, size, contour etc beads.
- Penetration, travel speed, shielding gas flow rate and machine parameters were all open variables the welder was free to manipulate in order to achieve a similar quality bead. I mention penetration as a variable and I'll come back to that later.
After completion of each test plate machine parameters were recorded, normally I recorded travel speed too, but didn't due to time constraints (I didn't have a stop watch with me). There was no expectation for a large or intrinsic deviation between travel speed for the two shielding gases and so it wasn’t considered an important variable. However I’m kicking myself now as one of your chief theories on base metal oxidation revolves around heat input based on travel speed (which makes perfect sense and I believe most welders are familiar with).
Below are the averaged machine settings between the two shielding gases:
Tri-mix:
21V
220 WFS
98-2:
27V
375 WFS
** this was an analog machine with no meters therefore settings are “knob” settings relative only to each other and not indicative of true machine settings, my feeling is that actual voltage and WFS are much higher. Generally I would physically measure voltage with a multi meter and WFS by measuring wire length over a given time. I didn’t have a chance to do this due to time constraints.**
This particular machine the welder had noted as being finicky and more “loose”, in retrospect not the best machine to conduct experiments on but the only machine at the time with stainless steel filler material.
The most immediate difference between the two machine settings is the higher voltage and WFS, travel speed however was slightly higher under 98-2. You assumed and I quote,
“Since the higher the welding speed - which should be assumed to be reached when using tri-mix (please correct me when I'm wrong)”.
Travel speed under 98-2 would actually be higher (although obviously not by a theoretical factor of 2) When trying to maintain an equal bead appearance under a “colder” less penetrative arc (due to shielding gases). When reacting to a cold weld bead the welder would naturally do one of several things: Reduce travel speed to achieve similar heat input as you theorized and I’m sure we all have done, adjust welding technique, gun angle etc to increase heat focus on welding bead, or adjust machine settings to increase heat input.
The welder I worked with adjusted machine settings because, at identical machine settings the bead surface was rough and uneven indicating low heat input (fast freezing puddle). However lowering travel speed was not a solution as that would create unwanted surface buildup for the given joint. Basically at the tri-mix settings the parameters were outside the acceptable range for creating an identical weld under 98-2. The only solution that would increase heat input so as to give good bead characteristics was to raise machine settings by a sizable factor.
Now here is a big detour and in depth explanation on how raising machine settings would increase travel speed. Anyone here with decent welding experience would know that if one raised voltage and WFS naturally travel speed would increase and can probably skip it, but I included some interesting findings
*BIG DETOUR*
Due to my previous efforts on measuring weld performance and weld cost evaluation I have done a lot of experimentation on how welding parameters affect performance. In conjunction with a statistics class I took as part of my internship I measured the prime variables of every welder at my facility under different settings to accurately estimate welding cost. This was also part of a previous experiment I had enacted to switch all shielding gas for carbon steel GMAW from 100% CO2 to C15 (85% AR 15% CO2).
As mentioned above experienced welders react to machine and material settings by compensating in several ways. In my experience the prime welding variable for GMAW is voltage, on constant voltage machines V is related to heat input, however if voltage is increased too much than there will be issues of wire burn back and puddle viscosity. Therefore WFS (filler deposition) is also increased to compensate. However WFS also controls Amperage so by increasing WFS net energy levels in the electrode are higher. However, on any individual molten droplet energy density will be roughly the same due to the increased WFS (volume) of metal exiting the nozzle in a given period of time. In order to compensate for this increased net energy and filler deposition, travel speed is increased
Due to this chain reaction of parameters for any one voltage there is a narrow range of acceptable WFS. Due to this relatively narrow range of heat inputs for a given voltage and filler deposition, there is now a narrow range of acceptable travel speeds ( to achieve good bead quality)
Now the whole point of this long detour is that if other variables are maintained and after sufficient experimentation one can accurately predict both WFS and travel speed based on voltage alone! Attached is two graphsGMAW.jpg that reflects this. After collecting data on a sample of 12 welders a direct correlation can be seen between Voltage and travel speed with the objective being identical, quality bead appearance etc. (sorry for the bad format of this particular chart I don’t have access to all my data right now). Additionally is a chart reflecting the observed maximum and minimum Travel speed (TS) for a given voltage which would follow the acceptable parameter range.
In fact with a large enough sample size (the number of welders at my facility) I was able to establish a formula using statistical methods to calculate travel speed for a voltage with about 90% confidence, of course due to the parameter range there is a always going to be variation in weld parameters while still maintaining acceptable bead appearance. If requested I can dig up and attach this information.
Anyways, although travel speed is increase to compensate for increased heat input the ratio is not exactly 1:1 that is to say welders don’t increase travel speed as much as they increase voltage.
So what happens at higher voltages?
Through observation it has been my conclusion that by increasing Voltage, and hence by proxy WFS at a 1:1 ratio, but by not increasing travel speed at the same ratio;filler deposition and heat input is slightly increased. On square grove welds and fillet welds the heat and WFS means more penetration for groove welds and larger leg size for fillet welds.
This also corresponds to a higher heat input “ceiling” for acceptable welds at higher voltages with colder gases. Below is the observed heat input range by switching to a “colder” shielding gas and allowing welders to change all other parameters to ensure acceptable (identical) welds.
*Averaged settings*
22V: CO2
354 WFS
145-165 A.
Heat input: 10,000-13,600 kJ/in
29V C15
648 WFS
210-240 A.
Heat input: 12,000-16,000 kJ/in
Due to the colder, less penetrative gas heat input (voltage) is raised to the necessary range to insure similar bead fluidity(good bead appearance), which leads to the cascade of events as previously mentioned. From this information we can ascertain that when switching to a “colder” arc situation, heat input will generally be higher even if travel speed is increased!
it should be noted that that when switching between 100% CO2 to C15 this opens the ability to use spray transfer, a higher energy mode of transfer, thus it’s impossible to raise CO2 above 24v (erratic globular transfer) and as such this skews data slightly as naturally C15 will have a higher energy input. Tri-mix and 98-2 can both achieve spray transfer at similar Voltages.
To further correlate my findings that a colder shielding gas will increase heat input and, by proxy, penetration; are pictures of two welds taken from the above trials, one of C02 and one of c15 titled (C02.jpg and C15.jpg respectively). Those are nick break tests from actual welds performed during the formation of the information above, it’s easy to see the increased penetration (deposition rate) of C15 if surface appearance is maintained.
What this all means is that the one metric welders use to judge non code welds (bead appearance); can improve welding characteristics (travel speed, penetration, distortion) If they switch to a colder shielding gas and try to maintain bead appearance via machine settings!
*END DETOUR*
Now how this all relates to the stainless steel situation at hand?
Well as previously mentioned and generally accepted 98-2 is a “colder” less penetrative shielding gas than tri-mix. If the fixed variable is bead appearance (the metric welders at my company use) than a colder shielding gas will necessitate higher heat input and by proxy, higher penetration, energy density, filler deposition etc. Then by the mechanics you explained so well, increased latent heat in the weld bead will lead to higher oxidation which is also coupled to the increased travel speed under the colder gas.
Basically we came to the same conclusion with two slightly different versions of what transpires under a changed shielding gas.
As you mentioned about heat input into the base material due to higher travel speed and therefore subsequent discoloration. I saw no noticeable change, I suspect a macro cuts and acid etching would more easily highlight HAZ size, which I am assuming would be parallel to surface oxidation in base material and which would correspond to your good explanation of heat input as a function of travel speed.
I think most welders are aware of this phenomena which means that sometimes although a weld can be achieved at lower speeds travel speeds with lower heat settings, over heating of the base metal can occur, which is important on a sensitive metal such as Stainless steel.
Finally you make mention of the complexities of gas flow etc and it’s relation to shielding.
Quote, “the high helium containing "tri-mix", I suppose that the flow rate has to be increased due to the appropriate lower density of helium which is the "base gas" of even this composition”
However under the tests the welder actually increased flow rate under 98-2 as opposed to tri-mix. This was his reaction in trying to counter the increased oxidation. I believe flow rate was around 22 CFH for tri-mix and 28-30CFH for 98-2 but I’m going off of memory of an event a month ago. The only basic guess I have is that some of the welds were performed vertical down in which the lighter than air helium might rise somewhat to cover the trailing weld bead. Similar to its use for back purging in the high areas of a piping loops. However the rest of the welds were performed in the flat position where I would think the opposite would be true of argon which would dissipate slower and downwards as opposed to the rising helium. Again that is just a guess and I’m not sure how much that would actually influence shielding characteristics.
Very amazing posts Stephan! It made me think quite a bit about fundamentals of welding physics I’ve always taken for granted. I posted my rather large detour as I thought other forum users might find the experimentation I’ve been doing on shielding gases informative.