thinking of building a forging furnace

C

ChuckTilbury

Joined
Oct 26, 2010
Messages
27
I am thinking of building a forging furnace. I have a background in electronics and software design, so I am thinking of also making a computer controlled thermostat for it.

I was wondering if anyone has any interest in a programmable furnace controller. You would be able to use propane or natural gas to run the furnace and it would have 3 settings: high, low, and medium. You would be able to program a sequence of heats over 24 hours by 10 minute increments. It would also have the usual controls, too.

I am thinking of selling kits and/or completed furnaces. Does that sound like a good idea? I am interested in opinions.

Thanks everyone!
Chuck Tilbury
 
Chuck,

I think most makers end up making their own forges, there are a lot of plans on the net. But with that said, high quality forges at low and/or reasonable prices are few and far between. I suggest you research the market, if you can make something better-at better prices, then go for it.

Dave
 
David Stifle said:
Chuck,

I think most makers end up making their own forges, there are a lot of plans on the net. But with that said, high quality forges at low and/or reasonable prices are few and far between. I suggest you research the market, if you can make something better-at better prices, then go for it.

Dave
Click to expand...

Thanks, Dave.

All of that makes sense, since most of the commercial forges I have seen are 2 or 3 times the cost that I think I can build one for. On the other hand, you can get a pretty good PID for about $150 or so. I am not the first one to notice that. :)

What I am talking about is using that old junk PC in your closet for a really smart PID. There would be a hardware box that connects to the PC over a serial port and provides the PC with access to relays and the temp sensor. It would be very cheap, extremely accurate, and easy to use with a "home made" forge. Mostly software. It could also provide nice things like graphs and temperature logging and being able to handle any arbitrary heating and cooling cycle over any length of time.

I would probably license the software and the schematics as "open source" and provide kits and fully assembled units. I think the interface hardware (without the probe) would cost around $50 as an assembled unit.

A PC is a little bit fragile for a shop environment, but it would not need to be physically next to the device it is controlling. Also, you could control several devices from the same PC at the same time.

I have not really seen anything like this. Have you?

Thanks,
-=chuck=-
 
No, I haven't seen anything like you've descibed. If you hang around here long enough, I would hope some of the really experienced guys will chime in. If they don't, it's probably because they just missed your post in the mix on this forum.

From my perspective, I think your idea sounds promising, and if you do go forward, best of luck.

Dave
 
Chuck
Well if we are going to make the perfect forge controller using electronics. Why not just use a Stamp (Microcontroller) and program it for three different heat settings. And use a three position switch to have the Stamp trigger the solenoid to fire or go to idle the forge.
It should be easy to do and you could forgo having a computer in the shop.
I am pretty sure we could get the stamp to read a thermocouple easily. Next the temp is reached and wham solenoid closes and forge is in idle mode.
Stamps are only about $50 and pretty easily programmed with Pbasic. I used to play with them alot in my last job.
Upside is one could program the three temperatures to anything we want and every one could be different if needed, or hell do 5 temperatures if one wanted to. And the electronics package would be tiny, maybe the size of two packs of smokes.
Electronics is not my forte in the engineering world but I do dabble in it from time to time so let me know if you think this is in the not feasible category.
Chris
PS If we used a rotary switch we could even get 10 temperatures.
Only hardware we would need is a stamp, thermocouple, relay, and solenoid.
 
Using a stamp-like controller is exactly what I have in mind for the part that goes between the PC and the forge. You can really get them for around $20, if you know where to look. Check out http://olimex.com/dev/avr-p20.html, for example. The MCU (Micro Controller Unit) by its self can be had for about $2. I am thinking of a very simple MCU that has at least one A/D channel, at least one serial port, and at least 6 GPIO lines. All of the code that goes in that will be intended to communicate with the PC. The cost of the external dongle (with the MCU in it) would mostly be in relays and the case. It would have 5 or 6 "electronic" components at most.

I thought of using a rotary switch, but I really want to be able to do precisely timed and temp controlled sequences of temperatures that could take hours. For example, annealing and hardening aluminum parts in a toaster oven. I know that I would get involved in doing something else and skroo it up about every time without something like this.

I have also toyed with the idea of using an old electric range that I have gathering dust as an annealing furnace. You could easily hang 4 or 5 devices at the same time on 1 PC and all of them would be "smart". All you would need would be a serial port and cable, a thermocouple, one of these external dongles, and a relay for each one. The PC would not even need a keyboard. Just a row of 6 buttons on each side of the monitor. :) Almost all of the real effort would go into the software for the PC.
 
Chuck
OK I see where you are going, you want to make a very sophisticated forge or heat treat oven. One that the user can tweak using the computer to do whatever it is they want to do. IE Anneal stainless which takes a very long slow ramp down in temp etc.. and or run several devices at one time etc.
I like it, I think its very doable.
Only thing I wonder now that I think about is whether the Microcontroller will accurately measure the Thermocouple reliably vs a PID.
Guess you would just have to do a side by side comparison to ensure its up to the task.
PS Cool thing about this potential product is you could use it to run a heat treat oven or a forge.
Chris
 
blindhogg said:
Chuck
OK I see where you are going, you want to make a very sophisticated forge or heat treat oven. One that the user can tweak using the computer to do whatever it is they want to do.
Click to expand...

Yup! :) That pretty well nails it.

blindhogg said:
Only thing I wonder now that I think about is whether the Microcontroller will accurately measure the Thermocouple reliably vs a PID.
Click to expand...

I was wondering about that, too. I looked up a table for the V produced by a type K thermocouple. (http://www.omega.com/temperature/Z/pdf/z218-220.pdf) It did not look difficult. It looks pretty linear. The A/D converter on a AVR261 (http://atmel.com/dyn/resources/prod_documents/doc2588.pdf) is 10 bits wide, giving it a range of 1024 steps. And the usable range of the application is from 100F to 2500F, which is 2400 steps. That makes the resolution about 2.3F. The rest of the deviation, if it were around 10% per step would not add much to the error. It seems like +/-5F would be close enough... :) A problem would be that someone would have to calibrate the sender if they built it. Otherwise, I could calibrate it.
 
You must log in or register to reply here.
Back
Top