3986 driver and related fabrication

[originator]

For those interested in contributing to the processes of fabbing the 3986 or other smt parts:

I am working up a very simple application that can be OS X or PC based to graphically control an oven via a USB2serial adapter. There can be any number of phases added as needed to create the optimal profile. The profiles can be named and stored/recalled. The phases can be ramped at a rate of x degrees(F) per second. The software simply runs through each phase at the PWM rate specified by the Change Rate. When the goal temp is reached, the timer holds that temp for the value entered for that phase.

Below is a sreenshot of the beginning stages app already mentioned in the main thread. Any contributions are welcome.

The app currently will step through all phases, outputting on/off pulses to the serial port(RTS line) based on the OnTime and OffTimes specified by the Change Rate.

[markcomp77]

Idea on description of temp. profile

let's set up such profile:


using G-CODE like language:

G1 X140 F5
G1 X160 F0.1666
G1 X230 F5
G1 X229 F0.2
G1 X140 F5

---
X - temperature [C]
F - speed of change [C/sec]

=> temperature controller will work like servo driver
motor <--- heater/fan
encoder <--- thermometer(s)

[originator]

Ok, any suggested names for each phase? It looks like you have drawn 6 phases, so I will add 2 more to what is there already.

The method I have come up with so far is the RTS line from USB2ser cable adapter as control line. It toggles as some specified rate, and controls an SSR which is controlling the heating element only. I am thinking to leave the rest of my oven as is, and control the fan from its control panel.

The Rate of Temp Change buttons affect the duty cycles On Time to be ON more than off, the decrement will set the On time to be on less than the off time. This cofiguration will allow full spectrum of power, full off or full on, with sliding scale of change. The + direction change ADDs to the TIME ON, - value Change Rate subtracts from the On time, cooling off the system, so that a ramping of Temp can take place over time, hence the Change Rate factor. If set at a max rate of change, the power is set to full, and the change is fast. If set to slow change, the opposite.

I am thinking of a simpler version too, I just got a 600F oven thermometer for $4, it can sit in the oven, and you can set your profiles by watching the dial, and enterng the right values in the profile. Servo is best, but costs a little extra for the Probe, 1wire IC, 1wire to Serial IC, some level translation. With the the non feedback version, you are out $10 for an SSR, $29 for a USB2 serial cable, plus a thermomoter.

There is a chip I am exploring for the 1wire to serial

http://pdfserv.maxim-ic.com/en/ds/DS2480B.pdf

[BCwanderer]

Why not use a thermocouple and a pic chip which has a 10 bit A/D converter
Add buttons and a LCD display.
Then drive a 110 volt triac with a zero crossing optocoupler and you don't have to code phase angles.

The rate of change in the heater element is much slower than the power cycle so you don't need a fast rate of control.

The 18f252 chips have enough pins to drive everything.
They cost as little as $4 on ebay and The settings can be entered into flash memory.

Cost could be as low as $30 and there would be no need to connect to a computer?

[originator]

BC, you guys are certainly welcome to work through a hardware version along with this version, personally I like the idea of the computer's simplicity. The presets are fast and easy to use. Right now with NO hardware(only USB2serial) I am running profiles on the oven with a straight wire off the rs232 RTS to the + on the SSR. I am not trying to zero cross the ac at the moment, unless is really becomes an issue later. I have some thermocouples here that I am looking at options for getting the (super tiny) voltages out of(.000 at room temp, .002 at aound 150F). I have the ds2760 chip(1wire) ready to test, but no interface yet. Using any processor is easy, but adds hardware and extra coding. I think with the 1wire to serial 8pin dip, that may be the ultimate in cheap, low cost profiling.

I will add in a F/C choice later, plus a selection for cycle length(500ms to 4000ms). I will also add a choice for the bias for Rate Change, so each oven can be calibrated to get it close to the actual Rate of Change in Degrees/sec.

[kreutz]

If you want an open loop control (no temperature control loop) take a look at this: http://www.stencilsunlimited.com/ste...icle_page1.htm

You don't need to control the heater, use only a reflow indicator crayon and monitor its state through the oven's window. No modifications are necessary.

Even when using a temperature controller with closed loop you should place the temperature probe on top of the board you are soldering in orther to make sure you reach the reflow temperature at the board. temperature distribution inside the oven is not homogeneous.

[phomann]

Originally Posted by kreutz I Even when using a temperature controller with closed loop you should place the temperature probe on top of the board you are soldering in orther to make sure you reach the reflow temperature at the board. temperature distribution inside the oven is not homogeneous.

That's why I recommend and use a fan forced toaster oven. It provids a better temperature distribution throughout the oven.

Cheers,

Peter.

[originator]

http://www.studiotoyz.com/TempProfile.zip


This is a link to the early stages concept app.

It does not do much yet, I haven't set the Temp and Rate to do anything yet till I get through experimenting with how to implement a few things. You can change all the sliders and Rate value, store the presets and recall them in the Preset Menu to see the idea of how it will work. To do a simple test, put an LED or meter on RTS pin(7 I think). Enter some low values for time in each phase(like 1 or 2 value). Before pressing Start, locate your ports, and choose one. Then press start. You should see The phase names changes as it goes though each one, the counter below the name shows how many on/off cycles it performed on the pin. I am driving an SSR direct with this pin to the + side input, but keep in mind this values goes negative too. (If you press Star prior to selecting port, quit and re-open app.)

Very soon I'll have the Rate set up with some options to bias it to attempt to achieve close numbers of change.

Closed loop will be up within a week at most, soon if time permits.

The goal for my purposes is to be able to crank out boards with the least amount of thought and tweaking after the initial setup. A stencilling process will be next on the agenda similar to Stencils com idea, only I'll be lasering mylar.

I was told that there is a vapor phase process that is excellent, and can be done in a pasta cooker with a liquid called Galden. Supposedly it does not allow overheating of the parts, no profiling to deal with. It may be something to look at as well.

The experiments with the thermocouple using microprocessor is going well. The temp change is very quickly updated, so tracking the real oven temp will be easy. The instruction set for the processor is not to complex to get temp in F and C every sec, so I may be able to convert it to the computer using the serial to 1wire chip. If this works out, the hardware involved will be: the serial port(or usb2serial), a thermocouple, a DS2760, and a DS2480B interface IC, an SSR. Very minimal parts required.

[originator]

If anyone requests it, I'll post the code for the SX28(Parallax) that I used to interface the 1wire sensor. The chip is $4 dollars, requires a 4mhz resonator. It also requires an interface to program it that fits on the RS232 cable. It is called the Blitz, I think they are around $19. The software to DL to the chip is at their site called SXkey. The processor will simultaneously drive an LCD from Parallax(backlit is $29 or so), so the SX can be used as standalone or with the PC for temp display, it simply reads the DS2760(in C), and outputs the values to the LCD and serial port once a second. Temp range is 0 to 1000C. It requires a Ktype thermocoupler.

Now that I have the temp into the computer, the rest will fall into place ove the next few days, a few tweaks here and there. Hopefully I'll have it where it can be calibrated for each user without the need to rewrite code all the time.

The first working version should be up in a few days. I have set it up to run in either timed or servo per phase by selecting the WaitTemp button. If it is checked, the timer will not start until the target temp has been reached per phase(at the rate of change selected). If not checked, the phase just runs for however long the time is set for. During the servo mode when waiting to hit the target temp, the progress bar will show the barber pole until the goal is reached, then it shows the progress.

[smarbaga]

i did a job and proposed an oven controller in a place that re-balanced and re-hardened jet engine fan blades and shafts.
these things were finally re-heated and cooled.
if i remember correctly some of the temp. times were like:
from 200 - 1100 degrees in 3 hours
stay at 1100 for 8 hours
cooling took 4 hours ( down to 200 degrees)
different rise, cook and cool times for different materials.
so this unit could also be used for metal treating as well.
heating elememt wire is not expensive.
altho the toaster would have to be replaced with bricks, or the oven would be toast.
i think powder coat machines do a similar process,

[pminmo]

So far my findings make me question the ability of a toaster oven to heat at the rate necessary. Admitadly I bought a really cheap small one, but at $9 I figured what the heck. I'll probably buy a second and move the two heating elements to the original. http://www.allegromicro.com/techpub2/an/an26009.pdf

[originator]

Smarbaga, you are right there are a lot of applications for temp control, for me paint, glue, pcb's, and other stuff will popup to make autmating temp control easier. With this system the device is 1wire, which means you can add any number of extra devices, and get that info iinto the computer just as easily for either averaging or controlling multiple areas.

Phil thanks for that link, lots of good info in there. I think you need a good oven for sure. Did you try putting in an analog temp gadget to see what it will do? I bought one at a drug store for $4, it was dead on 500F with the ovens temp readout. Convection will certainly smooth out the temps, adding a fan shouldn't be a big deal if the oven will hit the temps.