Brazing service?

Paul Baker

My next large project will likely be an infrared toaster oven converted into a reflow oven (detailed here http://www.circuitcellar.com/library/print/0704/Lacoste%5F168/·)
The oven will reach temps of 250 deg C, the directions state that the thermocouple needs to be brazed to the wires, or using some other high temperature fusing method (makes sense you cant use solder, if the oven is made to melt solder). Does anyone know of a brazing service (or what type of local shop would be able to do it)? Or will I have to rip open a lantern battery yank the carbon rods, hook them up to my SLA brick and arc weld it myself, trying not to burn my retinas or skin in the process?

steve_b

I've heard of guys out wheeling that have used a bunch of car batteries in series/parallel that were able to arc a spark to tack things back together.·

car batteries being expensive...go to a metal shop.· All towns have them (from one degree to another).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·

Steve
http://members.rogers.com/steve.brady
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."

SteveW

Are you sure you need to braze anything? Thermocouples come pre-brazed, and I can't imagine you want to measure the temperature of anything other than the PCB, so you won't want to be brazing your thermocouple to an part of the oven.
(If you're thinking of making your own thermocouple, you won't save any money that way - pre-made ones are $10 or so, for a decent fast-response style.

Steve

Paul Baker

Ok Ive looked through the documentation for the DS2760 Thermocouple Kit, am I correct that the wires themselves form the thermocouple? If so that would be ideal, no brazing is nessesary. Additionally is the wire insulation rated to withstand temperatures for the entire range of the thermocouple?

Paul Baker

Many thermocouples you buy only have leads of a couple inches max, thats not enough to mount it near the center of the oven, hence needing to braze the leads to wire. Correction, I found a few which are long enough, but are at cheapest $15 for the thermocouple itself, Id be nice if the parallax product would be useable since it has the electronics and all three types included.

Post Edited (Paul Baker) : 2/14/2005 8:04:12 PM GMT

SteveW

Yeah - for these sort of temperatures, fibreglass insulation will do absolutely fine. It all gets more exciting when you're measuring automotive exhaust gas (hot, corrosive) or plastic in an extrusion die (ridiculous oressures, mechanical abuse). For spotting temperature in a reflow line, it's an easy life.
You might like to run a couple (or more) therocouples, to check that the whole board is getting up to temperature, especially if you're planning to reflow boards that are nearly as big as the heating element. Analog Devices have loads of useful app notes for their thermocouple amplifier chips. (I don't recognise the DS2760 - but, as long as it can accept type K thermocouples (cheapest, most suitable temperature range, most easily available), I imagine you'll be fine)

Steve

Forrest

I've always bought my thermocouples already welded from Omega Engineering. These are high temperature thermocouples designed to withstand the heat of a solder reflow oven www.omega.com/pptst/5TC.html

Paul Baker

Thanks guys, hopefully a person from parallax can specifically answer the question about thier kit, something else that puzzles me about the kit is they refer to the thermocouple being connected to the cold junction, does this just mean in reference to the other end of the wire?

Steve: the DS2760 is actually a high precision Li+ Battery monitor adapted to this application, thats likely the reason you've never heard of the product. (I'm not sure but I think they did it because the DS2760 provides a 1-wire interface instead of an analog out like most thermocouple chips do, eliminating the need of an ADC)

SteveW

Hmm. in that case, I think you're doomed.
From Maxim's site - "Temperature is measured using an on-chip sensor". That's not a thermocouple...

It can only report +-127oC, and is only rated to work from -40 to +85oC. With a real thermocouple, the sensing element is just a pair of wires brazed together at the far end, and the silicon can stay safety out of the hot area. You may yet end up having to build this out of an amp and an ADC.
http://www.analog.com/en/prod/0,,760_792_AD595,00.html are the amplifiers I use for this.

Steve

SteveW

<more rummaging>
Ah, fair enough -
http://www.sensorsmag.com/articles/0102/29/main.shtml
uses the DS2760 rather nicely. Sweet!

Steve

steve_b

Go to your local furnace repair shop.· They will have a bunch of thermocouples.· And I'm sure you will be able to buy some.

Odds are you'll find on in the length you're after.· But do some homework and know what type to ask for (temp range rating)

And you don't actually have to measure the temperature directly.· What I mean is, the temp gauge on your car doesn't directly represent the temperature that it's actually running at.· You're car runs a lot hotter than 100degF....but it's a relative thing.

SO....don't shoot me for getting this wrong....but assume heat follows the square law.· So, if you were to measure 100degF 4inches from the source, then at 16inches (4 to the power of 2) then you'd measure 50degF.· You could measure the oven, 'from a distance' (no Bette Midler impressions please), and not have to worry about melting sensors.· OF course, these numbers aren't workable, as you'll probably need to be within the box to measure the temp.· In the end you'll need to know the 'exact' temp of the box anyhow, so you'll know how much of an offset you need for you thermocouple.·

hope that made sense....

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·

Steve
http://members.rogers.com/steve.brady
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."

Paul Baker

ick:

Features

• 1-wire interface

• Temperature range: 0C to +127C, Resolution: 0.125°C

• 16 bytes of general purpose SRAM

• Active Current: normal operation 60uA sleep typical 1uA

• Type K, Chrome/Alumel thermocouple min/max temperature = 0C to+1000C (+32Fto +1832F)

• Type J, Iron/Constantan thermocouple min/max temperature = 0C to +1010C (+32F to +1850F)

• Type T, Copper/Constantan thermocouple min/max temperature = 0C to +400C (+32F to +752F)

so it seems that while the themocouples are capable of measuring to the nessesary 250 C, the interface can only measure to 127, sigh.

achilles03

You don't need a thermocouple kit. All thermocouple measuring systems do is measure very small voltages induced by the thermocouple, which is a function of temperature (a thermocouple is a glorified battery). You can set up a low reference (maybe V@0C) and a high reference (say, V@300C), and use those as Vref- and Vref+ respectively on a sensitive ADC or other reference chip. After that it's just a matter of a curve fit, or linear scaling.

Dave

SteveW

You do need to know the temperature of the cold end, though, unless you're prepared to just declare that it's 'room temperature' and accept the inevitable errors, especially if heat from the reflow area might affect it. That's what the DS2760 gives you, in this application - a nice cold-junction temperature measurement, and a nice ADC stage, in the same cheap, small package.
With the realitively tight temperaturecontrol you want for reliable soldering (the gap between reflow temperature and component death temperature isn't that wide, and lead-free processes will only make it worse), I'd definitely not ignore the cold junction temperature.

Steve

The Dead Bug

Steve wrote:

-snip- SO....don't shoot me for getting this wrong....but assume heat follows the square law. So, if you were to measure 100degF 4inches from the source, then at 16inches (4 to the power of 2) then you'd measure 50degF. You could measure the oven, 'from a distance'... -snip-

Doesn't the inverse square law tell us that if you increase the distance from the source, as in your example, by a factor of 4, the heat would be reduced by a factor of 4 squared, or 1/16 rather than half? Also, temperature is an absolute thing, relative to -273C. I think the numbers you gave were based on the assumption that 50 deg F is half as hot as 100 deg F ? It's not.

Don't worry, I'm not shooting. I think your point is correct. [noparse];)[/noparse]

Bruce

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Name: Bruce Clemens

Work:· Clemensb@otc.edu
Good Stuff on my Bolg: http://theDeadBug.journalspace.com

achilles03

Steve,
If you're trying to get it within 2 degrees or so, yes, you probably need the cold junction temperature. Otherwise, you're gonna still be close. An easy fix is to put 2 thermocouples in series REVERSED, and put one in an ice bath. That way, you're input voltage is always in terms of 0C.

I'm not quite sure what you mean by "assuming room temperature" for your cold juction reference, though, since it's not acting as a real voltage reference.

Dead Bug (nice picture by the way),
Heat doesn't follow the inverse square law. For a very conductive material, temperature will decrease slowly from the source. For a insulator, the temperature drops rapidly from the heat source. The equations for conduction are linear, although geometry can make it complex. Radiative heat transfer does follow the inverse square law, more or less. And convection kind of escapes definition with those regards... convection is more of a black art, especially at high Reynolds numbers (aka vigorous flow).

Dave

steve_b

Hi Bruce,

ya, I was trying to use the square law as it works with light....figured it'd be similar to heat (radiant energy) but translating what the mind is going through has never been that easy.

achilles....square law certainly applies to light.· So, if you had an EXTREMELY conductive medium (wrt heat transfer) then wouldn't apply?· It's the point that materials don't acquire/release heat near instantly the way light dissappears.

I just had vegetarian lasagna....so I may be rambling from lack of red meat!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
·

Steve
http://members.rogers.com/steve.brady
"Inside each and every one of us is our one, true authentic swing. Something we was born with. Something that's ours and ours alone. Something that can't be learned... something that's got to be remembered."

achilles03

steve_b said...

Hi Bruce,

ya, I was trying to use the square law as it works with light....figured it'd be similar to heat (radiant energy) but translating what the mind is going through has never been that easy.

achilles....square law certainly applies to light.· So, if you had an EXTREMELY conductive medium (wrt heat transfer) then wouldn't apply?· It's the point that materials don't acquire/release heat near instantly the way light dissappears.

I just had vegetarian lasagna....so I may be rambling from lack of red meat!

It does apply to radiation, assuming it's radiating in all directions (i.e. not·a laser).· As for conduction, no.· Heat transfer·in materials (conduction) usually travels linearly (point A to point .· With light, it's spreading out spherically.· More specifically, light is·going from "point A" to·an increasing·number of points (which increases as the square of the distance travelled).··This usually isn't the case with conduction.· With some geometries, you could get a temperature distribution that did obey the inverse square law, but that is the exception, not the rule.

Let's say something has a thermal conductivity of 20 W/m/K (close to·stainless steel).· That means if you have a·1 meter cube of material, and one side is 0K and the opposite side is 100K, then 2 kilowatts of heat will travel through the block.· The temp in the middle will be 50K, the temp .25m from the cold side will be 25K, and so on...



Hope that helps,
Dave