17-4 H900 Questions

[mchidester]

Hello All,

I work for a manufacturing company and we are currently looking at the feasibility to bringing out heat treat processes in house to reduce cost and lead time for the parts. We have an oven that use for a different purpose and I have done some testing in that oven with mixed results. Below are the particulars

Material: Either ASTM A747 TYPE CB7Cu-1 SOLUTION ANNEALED OR ASTM A564 UNS S17400 TYPE 630 SOLUTION ANNEALED
Part Usage: Compressor valve bumpers
Size: 4 different sizes ranging from .75" x 1.5" to .65" x 3"

With the current testing I have done, the I have been successful in reaching hardness criteria (40 to 47 Rc) on the smaller parts, but on the larger parts, I have only had 1 of 10 come out within spec. On the other 9, there was no change in hardness, with it staying at 32-34 Rc as it tested prior to the heat treat cycle. As far as process, I have been heating the oven to 900 degrees, then putting the parts in, waiting of the oven to re-stabilize at 900 and then starting the clock for 1 hour. After an hour, pulling the parts out and letting them cool on a metal grate table.

From everything I have read on here, I made an assumption that the larger parts were taking too long to reach 900 and therefore not staying at 900 for an hour, so on my last test, I left the parts in for 2 hours at 900, but got the same result of no hardness change. I will caveat this with the fact that the oven we currently have does not circulate air and is also old and worn out so there is not uniform heat distribution throughout. However, I am putting all of the parts at the same spot in the oven to try and control that variable as much as possible in the testing. We are planning to buy a new oven, but I am trying to validate that we are not missing something before making that purchase.

What am I missing?

Thanks in advance for the help.

 

[MagBen]

1. are you sure it was 900F, not 900C?
2. hardness no change can mean improperly solution treated. do 1900F solution again and 900F aging to check.
3. try fan cool at aging when parts are big size,or bundles.

 

[bobjustbob]

Have you attached a thermocouple to the part to ensure that the part is actually getting up to temperature? Is your control thermocouple reading correctly? Have you surveyed your furnace to ensure the working zone is getting up to temperature? What kind of oven, gas fired or electric?

 

[mchidester]

MagBen,

Yes, the temp is reading in F. I don't have the ability here to solution anneal (oven only goes to 1000F), so any parts that come out bad are just scrap (which I am ok with as we are just in the testing phase). I did try fan cooling the last batch with no change.

bobjustbob,

I have not attached a thermocouple to the part, but I will try that and see what I get. With that said, what I should be looking for in the process is the temperature of the actual part reaching 900F and THEN soak for an hour, then air cooling the parts to room temperature. Is that correct?

The oven is an electric 1000F Grieve. Like I said, it is old and beat up (the doors don't even properly close due to damage, the oven was initially used to heat and expand 40-50 pound rotors so that shafts could be stabbed in them and after years of use it is just beat up from pulling those in and out) so the temperatures inside vary each time I measure them. Again, we are planning to buy a new oven and all of the necessary equipment to do the heat treat, I am just trying to verify that we are not missing anything in the process that is going to prevent us from being successful once we have the appropriate oven.

Thanks for the help guys.

 

[EdStainless]

Take a bunch of your test parts that failed and send them out for re-anneal.
My guess is that they were as cast and had not been solution annealed.
They should age then, if the chemistry is correct.
Aging is not that sensitive,4hr at 925F will only give you about 2 points lower hardness than 1hr at 900F.
And the curves that I have show that at 900F you can age 24hrs with no drop in hardness, but at 950F hardness begins dropping slightly after about 2hrs.

I have seen castings homogenized, 4hrs at 2050F, cool to RT, anneal 30min 1900F +/-25, cool to RT, then age.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[mchidester]

Thanks everyone for all of the information up to this point. I got pulled off of working on this with a bunch of other stuff going on.

I have done more testing and I am still not getting good results. I have been using the following process.

- Place part in oven at around 500F and attach thermocouple to part
- Ramp up oven temp to 900F and start timer for 1 hour when part temp is reading 900F
- After 1 hour, remove part from oven and place on grate table to air cool (shop temp approximately 80F - 85F depending on time of day)

With this process, I am only seeing a hardness of around 32-34 Rc after heat treat.
I have tried a longer soak, anywhere from 4-8 hours with no difference in results. I have tried putting a fan on the parts while cooling with no difference in results. I also dropped a heated part in water directly upon removal from the oven just to test a rapid cooling and still only saw 32 Rc. There are 3 thermocouples involved in the process (temp controlling, safety shutoff, and attached to part) and all are reading the same temp, with the one attached to the part showing about a 100F lag during the ramp up, so I am pretty confident that the temps are reading correctly.

I have tested about 20 different parts from different vendors and our specification calls out for us to receive them in the solution annealed state, so while I have not sent any out for re-annealing, I have to assume that all 20 of the parts from different vendors that I have tested were not received in a non solution annealed state.

What am I missing in this process?

Thanks again for the help.

 

[btrueblood]

Thermocouple is new and calibrated/re-calibrated?

 

[mchidester]

Yes, all thermocouples are new and calibrated.

 

[EdStainless]

The thing with 32RC is that it could either be annealed (though that is a little high) or it could be overaged.
I suspect that the larger parts are not annealed correctly.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[mchidester]

Ed,

So it sounds to me like you are saying that my process looks correct, which would leave the only other cause of the bad results to be a material issue?

I ask that question because I am trying my best to rule out process issues on my side before starting to go down the material rabbit hole. None of the supplies for this material are in the US, which makes communication somewhat difficult.

Also, I just read back through all of my posts and I never mentioned that these are investment castings that we machine and then heat treat. I did notice in AMS 2759 it states that 17-4 casting aging time for H900 and H925 should be 1-1/2 hours, so I want to make sure that doesn't change the process at all.

 

[mchidester]

Ok, so I think I have figured out the issue, and it is a material problem.

I have 3 different parts I am working with. 1 comes from one vendor and 2 from another. All 3 prints call out the same material (ASTM A747 TYPE CB7Cu-1 SOLUTION ANNEALED OR ASTM A564 UNS S17400 TYPE 630 SOLUTION ANNEALED) and all 3 prints give an option for 2 different heat treat methods (this is from the casting supplier for the raw material, not the age hardening instructions for after machining):

Method 1:
1. Solution anneal by heating at 1900F.
2. Hold for one hour at heat
3. Water or oil quench to room temperature
4. Hardness 28-33RC max

Method 2:
1. Solution anneal at 1900F for one hour and gas/fan quench with nitrogen.
2. Age harden at 1150F for four hours and air cool.

So what I believe is happening is the smaller part that I have been successful with that comes from one vendor is being treated by method 1, allowing me to age harden at 900. The other two parts that I have not been successful with that come from the other vendor are being treated by method 2, in which they are already age hardened at 1150, so therefore I am not able to age harden at 900 and those parts must be re-annealed before they can be age hardened again, which unfortunately I do not have the ability to do.

I am going to send some of those parts out to have them re-annealed and then try age hardening again and hopefully I will be successful and that will prove my theory. Turns it looks like you were correct from the beginning Ed, it just took me a while to actually find that discrepancy on the print. The only thing I don't know now is why that method 2 was ever put on the print in the first place as the design engineers that were here when that was done are no longer here, but it doesn't really seem to make sense that you would age harden them before machining and then have to re-anneal and then age harden again after machining.

 

[EdStainless]

It looks like you hit it on the head.


= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[mchidester]

So in working with the foundry to try and figure out how they are treating the parts, this was their response. Keep in mind, they are overseas and communication can be difficult.

"The workpieces are heated in the vaccuum furnace at 1900°F, without anneal, cooling down naturally, and the hardness was achieved with 28-33RC."

What does the "without anneal" mean? This looks like the process for solution annealing (as long as they are cooling them rapidly enough). But it does not sound like they are doing H1150 afterwards like I assumed. I am waiting for a local vendor to solution anneal my test parts so I can retry H900 and see if that solves the problem, but in the meantime just trying to understand what the foundry is saying.

Thanks again for all the help.

 

[EdStainless]

They may not be holding long enough at 1900F, and they may be cooling very slowly (maybe even with a hold at 1150F?).
I think that they are telling you that there is not an true anneal or age on these.
IF that is the case then they are not complying with the spec. You should ask to see their procedure and furnace charts.
But if you really care you need to visit them, it may be easier and cost less to just re-anneal all of their products.
You need to anneal them and start over.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[ktgottfr]

I think maybe you need to be ordering 17-4 material specifically in "Condition A" in order to be sure it has been correctly annealed. Your supplier should be able to give you Material Certificates to go with that material that prove this to be the case, and the certificates also show the results of a test hardening to H900. If you can't get this from your supplier, then I would think they are not a very reliable source for your material.
A place like Mcmaster Carr sells all sorts of sizes of 17-4 bar stock in Condition A with material certs in short lengths for prototyping - you should not be needing to source this material from overseas...
Just my two cents