Finding Leaks in Your Vacuum Furnace – Part IV

Moving past what your leak rate should be, many times people forget that you need to leak check the furnace when it is at temperature.  All too often, people become fixated on the cold linear leak rate. You need to remember, though, that you are heat treating in a hot furnace, not a cold one. So many times a furnace can pass a cold leak check and a cold leak rate test, but when it is at temperature, something may start leaking.

Now you are saying, “Wow, I’ve leak checked the furnace with a helium mass spectrometer and I’ve verified my linear cold leak rate, but my parts are coming out discolored.”

In addition, oxides are showing up and your metallurgist is saying, “These parts have been exposed to oxygen; there’s a leak in your furnace.”

However, your operator is saying, “No, I’ve checked it.”

The problem: it might be leaking when it is hot.

This is why, periodically, you need to:

  1. Heat the furnace up to 2,000 F (1,093 C)
  2. Let the furnace soak out for an hour to degas
  3. Do a physical leak check on the furnace with a helium mass spectrometer

Trying to perleakchecksoftwareform an automated leak test with the software on the furnace (or, in other words, a hot leak rate check with software) yields a pass/fail point of reference only. It does not provide a legitimate linear leak rate or qualification.

An example of this is you put a load in the furnace, close the door and program the software to pump down for 30 minutes, shut off and check the cold leak rate (rate of rise). This is not leak checking. Leak checking is the art of walking around with a helium probe in your hand, spraying helium on the furnace and utilizing a mass spectrometer (a leak detector). That’s leak checking. Validation of the leak rate is a mechanical (either software or man-interface) manipulation of the furnace controls when the furnace is pumping, stopping, watching and calculating a leak rate of rise. That does not necessarily mean that is the true linear leak rate, but rather – as we said before – a pass/fail point of reference.

So you open the door, close it, pump down for 30 minutes and do a leak rate check (rate of rise) for 10 minutes, ending up with a leak rate of 30 microns. Does that mean the true leak rate is 30 microns? No, not at all. It simply means that – based on how you designed and performed your experiment – it is 30 for that day. So every day you do the same thing over and over and, if the furnace is operating correctly and your parts are clean, 30 is a good number. However, if one day it goes from 30 microns to 300 microns and the parts come out blue then you have a problem.

Ipsen’s software is also capable of performing a hot leak rate check. What you would do is power to a quench at the end of the cycle and program the software to check the rate of rise on a hot furnace. This allows you to get a data point called a hot leak rate point. Again, neither one is to be construed as the true linear leak rate; they are also to be considered a point of reference check.

Make sure to watch for the next post in our series, Finding Leaks in Your Vacuum Furnace – Part V, as we discuss inert gas leaks, specifically their impact on production and how to find them.


  1. Interested in the problems related to day to day operation of vacuum furnaces. Your blog is extremely helpful. Will appreciate if I can have the full series “Find Leaks in your Vacuum Furnace”.


    1. We’re glad to hear that our blog was helpful in answering some of your questions! The full “Finding Leaks” blog series can be accessed from the menu on the right-hand side (there are 11 parts in total). You can also type ‘leaks’ or ‘finding leaks’ into the Search bar (also located on the right-hand side, directly above the blog series menu) to find all past blog posts that mention leaks.

      Please let us know if there are any other topics relating to vacuum furnaces and their operation that you would like our experts to write about.


Leave a Reply

Your email address will not be published. Required fields are marked *