Section IX QW-193/288 mock-up qualification 3

[CWEng]

We make a product that looks like a heat exchanger, but isn't. A customer is requiring us to do tube-to-tubesheet mock-ups, which have been problematic for us even though we don't make tubes or tubesheets. The product we make has tubes from .028" to .042" thick, but they are filled with an insulating powder and then rolled/compacted, so act similar a solid bar. However, due to customer requirements, we are doing the mock-up with hollow tubes. Even though our plate has a relief cut to even out the heat between the thick plate and thin tube, it is difficult in a manual weld to fuse the tube without burning through (despite the fact that ASME doesn't define "burning through"- AWS defines "burn through" as a concavity- so our customer has accepted some "melt through" that is not concave). What is the thinnest wall thickness you tube heat exchanger manufacturers manually weld without issues? Seems like typically I see references to 1/16" and greater, but I've not manufactured typical heat exchangers. Also note that our tube sticks out way past the joint, so melting the edge of the tube is not a possibility, as that seems like it would make it a lot easier. It has to be a true fillet weld.

 

[meltedEng]

Define your design and acceptable criteria and replicate the joint exactly to simulate the real welding conditions. Tell the customer to shove it otherwise. Your design, they are buying it. Don't allow them to impose extra conditions on you that aren't needed or realistic.

Also to answer your question, even with automatic welding systems the best you can do is about 0.040" wall thickness without significant effort.

 

We are welding 1/2"OD x 0.040" nickel alloy tubes using fillet welds and burn/melt-through is a constant risk. This is using orbital equipment; manual would be close to impossible.
I agree with meltedEng that 0.040" is about the practical limit for a fillet seal weld.
This is a problem that is dealt with by redesign and/or renegotiation.


"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[EdStainless]

I have watched guys hand seal weld 0.028" SS tubes, but they were the best. Anything under 0.049" I would go automated orbital with.
But back to ME's comment, it is your design so do they want to buy it or not? That is the question. If they want you to change something in the design or process then they should be paying for that.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy

 

[CWEng]

Thanks for your inputs.

We do weld some orbital/machine fillet welds, and that is a lot easier to qualify than manual welding, however we use .049" tubes for that due to other considerations. But most of our configurations are not designed for orbital welding, so end up manual welding. I would love to limit the customer's extra requirements on our design, but often by the time I see it, the sales department has already let the horse out of the barn, and committed to things like that. And purchasing of course likes to keep pushing engineering to use thinner tubes because they cost less. Of course, they don't see the issues it causes. I'd love to get thicker tubes for manual welding, but wanted to confirm we aren't the only ones that would be having issues trying to qualify things like that with manual welding. Ideally, we could at least use thicker tubes when customers require this sort of testing, and then hopefully pass a surcharge for it on to the customer.

It is a bit interesting in that since our tubes are filled with a ceramic, it kind of acts like backing, so in the fitness-for-service of the component, I don't think "burnthrough" really matters. Our bigger issue is that when welding, there is invariably small amounts of gas trapped in the ceramic, so if the weld gets too hot and the tubes are too long (to let it escape out the end of the tube), the hot gas will vent through the weld, causing a blow hole. That is a bigger issue in real production rather than mock-ups.

 

Not a good situation when sales/marketing is driving engineering design.

There does not appear to be a strong culture of continuous improvement, or sensitivity to the long-term reputation of your products.

"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[CWEng]

Sales/marketing might make silly customer commitments, but I wouldn't say they are driving the design. As for our product, it's really the same struggle every company has between building the best product possible and building one people will pay for. Mostly in this thread here, I'm just trying to make sure my push back to sales/engineering is reasonable in trying to make improvements. (i.e. if I have an issue doing the qualification, and it appears other people here would have the same issues, let's change something in either the design or what we accept for customer requirements so we can make sure we only sign up for reasonable requirements)