Quench Problem

ddavelarsen

ddavelarsen

Joined
Dec 7, 2000
Messages
5,179
Well, I recently switched to actual quench oil - Houghton G - for my O1 blades. I did a test to see just how good the stuff is, and I have a major problem.

Here's the setup. 3/16" O1 (cutoffs of my normal steel) soaked at 1455 for about 8 minutes. The program comes to 700 for 15 min then full ramp to 1455. I quenched in 140 deg G. I didn't time it but it might have taken 1.5 seconds to get from the oven into the quench.

When I went to break the test pieces before tempering ... they bent to about 45 degrees before breaking. They were fully cooled to room temp.

Can someone suggest where I've gone wrong here? I've used Brownell's Tough Quench with my one brick forge and gotten excellent hardening on small blades. I've quenched all my 'full sized' blades in common veggie oil at 140 deg previously - the only difference being the quench was much closer to the oven so time from oven to quench was likely shorter, though I've never timed this.

I don't know what's gone wrong here but this has to be resolved before I can make any more knives. I appreciate any ideas for process change and further testing. I can get some pics if needed but don't have any at this point. (I was so freaked out when my cutoffs didn't really break that I just shut everything down and went to do yard work.)
 
Just my opinion. Your soak should be at least at 1475° for 15 to 20 minutes. Your preheat should be around 1250° soaking for maybe 20 minutes. Your oil was a tad to hot, try around 125° to 130°. get your oil closer to reduce travel time.
 
Dave,
I ran into the same problem in working with the damascus order for leatherman, they had a third party HT the blades and they were coming out in the mid 40's Rc. I asked them how long they were soaking and they said about 5 min, I told them to soak for half an hour and bang Rc64 every blade.
Bump your temp closer to 1500 and soak and you shouldn't have any more trouble.
Thanks,
Del
 
Also the thermocouple readings would be not right. I don't know how to calibrate but I am always afraid of it. For this reason I plan to buy another TC and a pid just to test the actual working one...
 
Thanks guys, I'll change my program and give it another run. I think you're right about temp and soak time. Emre, I hope it's not the thermocouple though it might be. I have a digital controller that really hasn't been cycled all that many times. I may check out some temp cones (whatever they're called, research in order ;)) to check I'm getting the temps I think I am.
 
Hi Dave , I have never done my own heat treat but have been doing a lot of research over the last 6 months on O-1 . What the others have said is what I have read every where . 1200 degree initial soak , time based on steel size per inch . Then 10-20 mins at 1475-1500 . I have also read that going over 1500 could be critical . Also remember Phillips recent test and Kevins evaluation . It seems long soak times does not hurt O-1 .
 
Dave, I wouldn't count on temp cones honing in too close on your control accuracy. They are "interpretive" in that they react not only with temp but with time-at-temp. One has to learn to read how much they slump. I experimented with them at one point, using 5 cones, bracketing my setpoint with two cone temps on either side of the target centerpoint. While I found that I could tell whether it was performing "about the same" on multiple tests, I didn't feel the response was consistent enough to say I could "read" the results accurately. I have since abandoned the test.
Personally, I prefer having a second digital thermometer and thermocouple. I drilled a couple holes in the furnaces where I can insert an additional probe. I bought a long thermocouple so it can "see" various places in the furnace. If you do this, you'll be surprised how very inconsistent the interior of a funace is, end up placing your probe very close to your blades, and making sure neither probe nor blades are in the radiant path of the coils.
Just my 2cents after a lot of experimenting.
 
Last edited:
One other comment: if you have programmable ramping on your controller, and are ramping to final temp at full rate, add an additional program step that slows down the rate the last 100 degrees. I ramp at "full" to within 100*, then switch to 100*/hour so there is time to equilibrate and ease into the final temp slowly, giving an increased probability of homogeneity in the furnace and, thus, accuracy.
 
galadduin said:
Also the thermocouple readings would be not right. I don't know how to calibrate but I am always afraid of it. For this reason I plan to buy another TC and a pid just to test the actual working one...
Click to expand...

I calibrated mine by measuring the temp of ice water and boiling water, use the offset up or down to adjust what the readout says to match the actual known temp of the water.

Do multiple readings after adjusting the offset because it changes both readings (ice and boiling).

Also make sure you get the exact temperature that water boils at your elevation because if you use 212 and you are not at sea level your readings will be off.
 
69_knives, I have done the same for checking thermocouples, but never to adjust them. It can open a can of worms. You are using two points very close together (32 and 212*F) that are outside the intended use range for us and hoping it extrapolates to linearity many hundreds of degrees higher.
As an analogy, consider a long pole you are holding in two hands a couple feet apart. Move the end of the pole with either hand a few inches, and at the other end the change is magnified dramatically. Thus, any errors in the assumptions about the temps of the ice/water or boiling water or errors in setting the linearity magnify themselves drastically up at 1500-2000F.
 
Last edited:
fitzo said:
One other comment: if you have programmable ramping on your controller, and are ramping to final temp at full rate, add an additional program step that slows down the rate the last 100 degrees. I ramp at "full" to within 100*, then switch to 100*/hour so there is time to equilibrate and ease into the final temp slowly, giving an increased probability of homogeneity in the furnace and, thus, accuracy.
Click to expand...

That's another excellent idea. I will do exactly that, thanks man.

I think before I do the next round I'll lay out my program here and let you guys vet it one more time. I thought had it nailed pretty well using my previous quenchant (based on cutting stuff, not on rockwell tests) but this has me questioning even those years of experience.
 
Dave, I just re-read what I wrote, and had 100*F on the brain too long.:o
I went and checked my records. I bring it at max rate to within 40 degrees of final temp and then ramp at 100F/hr to get to the hold point. Given there's a bit of overheat at full ramp, it only takes about 15 minutes to stabilize at the correct final.
Hope that makes sense.
 
fitzo said:
69_knives, I have done the same for checking thermocouples, but never to adjust them. It can open a can of worms. You are using two points very close together (32 and 212*F) that are outside the intended use range for us and hoping it extrapolates to linearity many hundreds of degrees higher.
As an analogy, consider a long pole you are holding in two hands a couple feet apart. Move the end of the pole with either hand a few inches, and at the other end the change is magnified dramatically. Thus, any errors in the assumptions about the temps of the ice/water or boiling water or errors in setting the linearity magnify themselves drastically up at 1500-2000F.
Click to expand...

What assumptions have to be made about the boiling or freezing points of water? I don't understand.
 
Boiling point of water changes with pressure .So BP at sea level is not the same as in Denver !
But more to the point is that if an instrument has a working range of 1000-2000 F you don't calibrate at 100 F - the instrument is not designed for 100 F.
A point that is not well appreciated is the fact that O-1 is not a simple steel. It has enough alloying elements that soak time is important .
 
Ice coming out of the freezer is not at 32F , but colder (freezers are frequently set at approx 20F). That means, when you mix it with water, the inital temp is not 32, either. The "bath" has to equilibrate to a steady state. Anyone not knowing this who just mixes some ice in water and doesn't let it even out will have something other than 32.
Likewise, boiling water in a pan, pumping in more heat than it actually needs for a simple rolling boil can result in slightly higher than 212F before it suddenly boils over. It's what's called super-heating.
I made my comments to warn people that it's not as simple and straightforward as it may sound, and there are caveats. Missing one or the other of those calibration points even slightly could cause a big error far upfield in the measurment range.

When someone calibrates a device, it is usually verified for a linear range for use within the extremes tested, not "extrapolated' from a couple points at one end of the total range of measurement. If it isn't linear, then one has to generate data at sufficient points to create a "curve" on a graph so you can then "interpolate" the results.

There's a chap checks in here occasionally named "Howie". He's a professional metrologist, a specialist in measuring and calibrating devices. Perhaps he may see this and give a better explanation than I am able to without the big words.
For all I know, he might tell me I'm full of crap :D, but I don't think so.
Hope this helped rather than confused the issue. I'm not very good at explaining in plain language stuff i know at a technical level, and I apologize for that.
 
Last edited:
Well, thanks for that I guess.

I did say to take multiple readings, which takes time, time in which the ice water temperature would equalize. I also said to

Also make sure you get the exact temperature that water boils at your elevation because if you use 212 and you are not at sea level your readings will be off.
Click to expand...

so no assumptions had to be made.

Thanks for clarifying your post.

Forgot to mention that the De-facto calibration uses an ice bath so if one can't get an ice bath right, maybe needlepoint would be a better hobby. :)

I know personally, I felt better knowing that my PID wasn't way off by doing a couple of simple tests, That's all I got to say about that subject.
 
Last edited:
69_knives said:
...... so if one can't get an ice bath right, maybe needlepoint would be a better hobby. :)
Click to expand...

I couldn't agree with you more. :) At the very least, they shouldn't be changing the calibration on their meter.
 
Okay, here's how I plan to lay out the program. You folks can vet this to see if it needs any tweaks; I probably won't make another run until the weekend. I have a Set Pro control, so this will be familiar to those of you who have one. Otherwise I think it'll be pretty obvious what's going on anyway. Please let me know what you think.

Segment 1: FULL to 700* Hold 15 min (this is for coating with PBC which I won't bother with for this test)
Segment 2: FULL to 1250* Hold 20 min
Segment 3: FULL to 1425* Hold 10 min
Segment 4: 100* per hr to 1475* Hold 45 min (soak 30 min, 15 more min to quench a few blades)

Thanks for your help!
 
You must log in or register to reply here.
Back
Top