Upgrade oven to PID worth it?

Y

Yugami

Joined
Oct 13, 2015
Messages
102
I've been lent on a semi-permanent basis a Blue M lab oven to use for heat treating.

The control is pretty crude, high, med, low power and adjustable cycle time. I was able to get it tuned to fluctuate between 1508F and 1481F on a cycle (I didn't time the cycle while I was out there but 10 to 15 seconds I believe) The oven lost temp pretty quickly while the elements were off.

I come from a background of precise controls and monitoring systems including using things like Datapaq's to collect both air and metal temperatures for processes. I don't have that at home :( I'm guessing based on my experience that this will even out to about 1500F metal temp. I have my first knife tempering right now. After quench (preheated canola) there appeared to be a tiny bit of decarb but the heat treat took (no idea of HRC number) A file bit a tiny bit (the decarb) then skated off. Decarb seemed really light and a quick finish grind will take care of it.

So all that aside, would a cheap Ebay PID work better (I may also coat the inside with ITC-100 or something to increase high temp stability) or am I just expecting too much.

I'm using O1 steel (I have a ton of it, almost literally) so tight control is good.

Thanks for any input.
 
A PID controller is definitely worthwhile if you have the electrical ability to change it.

The ITC-100 is an infra-red reflector and will give no benefit at all in this application.

I'm not familiar with Blue M products, but assume you are referring to one of the 1100 degC Lab furnaces, presumably an older model from your description of the control.

The description of the current models suggests that the muffle construction is very similar to the Carbolite lab furnace I had a few years ago. I fitted a PID controller to that and found I could get pretty good control at steady state and could also get minimal overshoot on full heat-up.

The biggest limitation seemed to be the fact that the elements are buried in insulating material and there is a timelag between the controller applying power and the thermocouple seeing an effect. The autotune did a pretty good job and the control was fine on an uninterrupted run, but if the door was opened and closed again, there was overshoot as the temperature tried to get back to the setpoint.

I think with a basic PID controller, it will be fine for single blades where you can put the blade in a cool furnace, bring the blade and furnace up together, open it and quench (it doesn't need to start from room-temperature cool, just far enough below the setpoint for the controller to see it as cool: switching off for 5 minutes between blades would be ample).

If you are looking to batch-treat, opening the door, taking out one blade and closing the door, you may find the first and last blades out perform differently.

To put things into perspective, the "quality" of temperature control achieved was probably as good as, or better than, that of the state-of-the-art equipment available back when many of the older blade steels we still use were developed.

At the time I had mine, reasonably-priced ramp/soak controllers were not available. I'm pretty sure I could get better results with a ramp-soak controller. If you have already mentally budgeted for the ITC-100, I'd suggest you reassign that mental budget to upgrading to ramp/soak on the controller.

Auber instruments seem to be highly rated over there. Here in the UK, I use either the Omega CN7823 or the AutomationDirect Solo SL4848VR. They are the same instrument with different badges, as far as I can tell, and I use whichever is cheapest at the time.

http://www.omega.com/pptst/CN7800.html

http://www.automationdirect.com/adc...ers/1-z-16_DIN_Size_(SL4848_Series)/SL4848-VR

Even if you go for a cheap ebay controller, I'd recommend using an SSR and programming a 2-second output cycle time.

If it's practicable, building a separate control box that you can plug pretty much anything into is a good way of dealing with things. It minimizes your interference with the lent unit and puts you about halfway to a homebuilt HT oven if you decide to go that route in the future: I find the costs split is roughly half on the control box and half on the oven itself when I build them. Obviously YMMV.

Edit to add: Mine was great on single workpieces in O1, which is pretty much the only readily-available tool steel over here.

With a $40-delivered controller/SSR package (the thermocouple supplied with it will be useless, but you give temperatures in the OP, so presumably have that sorted already), and a bit of electrical ability, you should be good to go.
 
Last edited:
Thanks, sounds like a good way to go then, I'll start doing my research. I wasn't sure on the granularity that you could achieve with the cheap units.

This is what the unit looks like (not mine but better pictures than I can take in my garage right now. http://www.used-line.com/PicturesAuto/643/10388643.jpg

The analog temperature sensor isn't working but I compared mV output of the thermocouple to direct readings by thermometer in the chamber at lower temperatures and the Thermocouple is still working fine.

At this time one at a time should be fine. Down the road I'll look into a different unit or build one.

Did you do any thermal cycling with yours? The amount of time I left the door open when I pulled my blade out dropped the temp to ~1200F or so. Which seems close to the ramp temperature suggested in this data. http://www.speedymetals.com/information/Material9.html. So I'm thinking that thermal cycling should be doable for grain refinement.

I was hoping the ITC would give a little bit more temp stability but you make a good point that its mode of operation is probably not whats needed. I'll look for some foil type insulation I can put on the top and exposed side, both of which were too hot to touch.

Thanks for taking the time for such a detailed answer, I appreciate it.
 
Somebody more knowledgeable can come along, but couldn't you just stick some additional metal pieces into the oven for thermal mass? That's what I did with my small tempering toaster oven and it seems to help.
 
You mentioned you had previous experience in automation controls. If you can set the parameters of proportion of error , integral and the anticipation of error time you should be able to regulate temperature real well. I'm not certain how your oven works , primary element of control would be a thermostat or thermocoupler and final element would be a contractor? An SCR or IGBT would be ideal for maintaining steady heat or modulating heat Rather then a contractor or multiple contractors picking. Or much simpler ... Multiple elements on different contacts. The closer you get to a high alarm the contractors gradually un-pick to lower heat as it gets closer to set point.
 
The photo shows the elements visible from inside the chamber, so you will not have the lag that I thought you would: Mine had the elements separated from the chamber by a layer of insulation:



The fluctuation you mentioned may have been on the same time period as the output cycle of the controller: rising as the elements are powered and falling as they are off, to give a saw-tooth temperature profile. I've seen this myself. Reducing the cycle time reduces the size of the saw-teeth and about 2 seconds seems to be as good as it gets in my testing. A cycle time this short really needs an SSR. I saw no improvement when I tried going to 1 second cycle time on my home-built ovens, but I am on 50 Hz mains and 2 seconds seems to be the minimum recommended for 50 Hz.

I'm guessing you are forging if you are heat-cycling. I've only used O1 as Ground Flat Stock, so have not needed to heat cycle.

I've always had access to some fairly useful monitoring equipment through work, but I picked up an ABB SM3000 paperless recorder on ebay about 3 years ago for 100 GBP (about 155 USD) and it really helped me to get a feel for what is going on with an HT oven.

If you really want to even out temperature fluctuations during tempering, bury the workpiece in a tray of dry sand and increase the time to allow for the increased heating time. You can put a thermocouple in with it if you want to check the temperature. If you have a pyrometer already, great. If not, I'd recommend a TM902C off ebay: between 5 and 6 bucks delivered, it comes with a 3' long glass-fiber-insulated bead thermocouple that is good to about 400 degC (about 750 degF). It only reads in degC, but it's not difficult to convert if you think in degF. It has a miniature socket (the flat-pin type) for any type K thermocouple and will actually read to 1368 degC (2494 degF) if you use a suitable thermocouple. I've had a few on the calibrator and accuracy is as good as any of the big-name units I've used at over 20 times the price.
 
Thanks for the suggestions. The idea of more thermal mass is a good quick fix.

Yeah I'm forging.

Thanks for the equipment suggestions, I'll look into those and pick up the TM902C to have kicking around. I'd like to play around and make something to monitor metal temp. I just started an extra long soak timer to be safe during my HT.

I sharpened the knife and it was cutting leather and paper though wasn't shaving (but I'm not great at sharpening knifes, something else to work on) It cut out all the 8oz leather I'm using to make a sheath for it so thats a good sign.

grumpy_grinder I like the way you think, I'll have to look closer at the different cheap PIDs being offered, I'm not 100% sure of their capabilities.
 
You must log in or register to reply here.
Back
Top