Reflow oven and Solder paste?

Thanks Stephen

I did take a look for thermocouples last night and found a lot very similar to that one.

The only slight problem is that the pins on those connector don’t directly fit into the screw terminals on the MAX31855 module boards that I have.

Image

Strangely, I can get the modules cheaper on eBay than I can from AliExpress

http://www.ebay.com.au/itm/NEW-MAX31855 … SwLVZVniqU

(there are £4 versions on ebay UK)

So I’ll order another 2 or 3 of those, so I have some spares

zmemw16 wrote:http://www.aliexpress.com/item/Free-shi … eb201560_9

58p uk free shipping

i’ve max6675 on boards, offer expired though

MAX31855 on a board £4.10 ukp
bare chip ~£7

top of the heap adafruit £16-£17

stephen

Thanks Ray

I think the issue with my oven is some sort of electronics fault in the opamp.

I spent at least half a day trying to figure out why one channel was miss reading, and the other one was fine.
I resoldered various joints, and twiddled the adjustment pot, but to no avail.

So as I already have the Max31855 boards, it was just easier to do it that way.

I also can add additional thermocouples, and I’ve ordered another 3, so that I can have one thermocouple very close to the actual PCB, in the hope this will be be a better average reading of temperature in the oven.

I guess I could have gone done the analog route, and possibly managed to source another instrumentation amplifier, locally and built a complete new analog input section.

But I presumed that as the MAX31855’s are dedicated chips for this job, it would be better to use them.

Plus they also have the cold junction temperature as their output (which in this case I’m not sure I need really, because I though they were cold junction compensated, plus I already wired in a DS18B20 to act as the cold junction sensor for the existing analog inputs)

the yellow bits are held together with a nut & bolt
the wires are secured with a pinch plate and a screw

i’ve transplanted a few.

stephen

Thanks stephen

I have a similar thermocouple on my DMM, so I hoped I could just break the connector apart on the ones I’m getting from eBay

I’ll let you know how I get on..

RogerClark wrote:Thanks stephen

I have a similar thermocouple on my DMM, so I hoped I could just break the connector apart on the ones I’m getting from eBay

I’ll let you know how I get on..

Thanks Ray,

I will post again when the thermocouples arrive ( which wont be for several weeks)

I haven’t any experience with themocouples but imho a MAX6675 should be enough for temps < 300° C.
Maybe you find some ideas here: https://github.com/0xPIT/reflowOvenController

BTW..
RogerClark wrote:cazimirb wrote:A handy alternative to SMD oven is a hot air SMD rework station, like this one that I have:
http://www.aliexpress.com/item/scotle-8 … 69194.html

Well…

I finally got the oven going using MAX31855’s and an Arduino pro mini as an I2C to SPI bridge.

It took me a while to figure out how what I needed to send to the oven over I2C, but by looking at the code for the oven firmware, I noticed that if it receives 0x00000000 when it requests a TC value (during start up), it reads this a no TC attached to that port

So I have some very simple code, that listens for incoming I2C events, where the first byte is the command, and then in the onRequest handler, it either sends back 0x00000000 or it reads the Max31855 via SPI and returns the real value.

But for some reason I had to byte swap all 4 bytes of the data as it seems to be in the correct order (I’m using the Adafruit lib to read in the data from the Max31855’s but I may go back to plain old SPI)

The only other change I had to make was to the oven firmware, because the Arduino bootloader, in the pro mini, needs to complete and the sketch needs to be running when the oven scans for I2C devices. So I had to put a 2 sec delay in the oven’s init routines (and sort out the WDT which was kicking in because of the delay)

But its all done now

I’ll post a full writeup on my blog when I get time.

I’ll also post all of my code, once I’ve modified the Pro mini code to just use SPI.

RogerClark wrote:Well…
The only other change I had to make was to the oven firmware, because the Arduino bootloader, in the pro mini, needs to complete and the sketch needs to be running when the oven scans for I2C devices. So I had to put a 2 sec delay in the oven’s init routines (and sort out the WDT which was kicking in because of the delay)

Andy

The issue is that I need I2C slave functionality, and Im not sure its been written for the Wire lib for STM32 etc.

Its definitely not supported in the version in my repo, (i checked).

So I had to go back to using an AVR board for this.

I know I could have written the I2C slave code, but I just want to get the reflow oven to work.

I could also just flash the AVR board without a bootloader, as I have the hardware to do it.

But I was trying to create a fix that I could publish, for other people to use, who don’t have access to lots of fancy hardware.

So the easiest thing to do was to modify the oven firmware, as its really easy to compile on Linux.

BTW. Even the GD32 would not help, as the oven firmware scans the I2C bus within a few milliseconds of startup, so my solution would not work for any board with a bootloader unless I modified the oven firmware, to delay its scanning.

Edit.

Actually. I just found that there was a version of the I2C code written for libmaple that included I2C slave functionality

http://forums.leaflabs.com/topic.php?id=1543

I will see if its possible simply to update our HardWire lib with this functionality

In case anyone is vaguely interested

I eventually coded a I2C slave to interface with the Maxim thermocouple chips (which are SPI) using an Arduino Pro mini.

I put the oven back together using the Maxim chips for both internal thermocouples and the problem seemed to be fixed as both thermocouples read the same value at room temperature

Anyway, When I tried to use the oven, I found it was not getting hot enough to melt the solder paste, unless I set the temperature at about 30 deg higher than is correct.

Initially I thought the issue was that the oven has hotspots as its not got anything inside it to circulate the air I.e no internal fan.

So, in case this was the issue, I thought I’d get out one of those hot air paint stripping tools, and blow air in through the front by opening the drawer a bit, and blowing it in.

Anyway, Looking at the thermocouple readouts on the display; even if I blow air in, the thermocouple on the right seems to read higher than the one on the left,by around 20 deg, perhaps more. This was even if I move the hot air tool around from one side to the other.

So, I’m now not sure if one of both of the thermocouples is faulty, which may or may not have been the original problem.

It did also give me an idea for a better reflow oven…. Use a hot air paint stripper !
As far as I can tell, they are capable of delivering hot air well in excess of the 225 deg C that is required for lead based solder and actually capable of delivering air in excess of 300 deg (some places even quote 400 deg C or more, much more..)

I can get a “variable” heat one for $50 locally, but in reality, you would need to strip the thing apart, and control the motor and the heating elements separately, as the motor speed will be too high, as its likely to blow the components off the board.

Perhaps in conjunction with a toaster oven as a enclosure, this could be a better solution to reflowing boards .

… So, I’m now not sure if one of both of the thermocouples is faulty, which may or may not have been the original problem.

Is there any way to swap the thermocouples round and see if the problem follows the thermocouple, or stays at the same place, or are they bonded in place? If you can, then if the hotspot remains in the same physical location, then the thermocouples are fine, and the physical layout is suspect. If the hotspot remains on the same thermocouple, then the thermocouple is suspect.

Hi Andy

I must admit Im getting a bit lost about whats going on.

I looked inside the oven, and all it has is 2 of those cheap thermocouples from eBay with the woven white covering on the wire.

They stick down into the middle of the oven, about half way between the 2 IR heater rod things.

Initially, I thought it was a heat spot issue, but blowing hot air inside (no heaters on), still have me the same sort of error.

I can swap over the connections, as I suppose there could be a fault on one of the Maxim chip interface boards, which would not be impossible, but its strange that swapping from analog input circuits on the main PCB to extrnal interfaces on the two thermocouple digital interface modules / boards which use the dedicated Maxim chip to do the conversion, still seems to end up with the same sort of problem as I had on the analog electronics

I have 3 more thermocouples on order from China (via eBay), so I think I will wait until they arrive, so I can do some comparisons

It does sound like the thermocouples are the issue. You could probably compensate for this in software, but it would be better to use two known, good thermocouples. Furthermore if Googling for the issue doesn’t turn up lots of results, you have to conclude, that this is a local fault, rather than a design issue.

Do we have any spec for these cheap thermocouples?

maybe for each thermocouple, connect to same circuit in turn, use melting ice and boiling water.
if different output, its the thermocouple else if pretty close try with the other circuit if still pretty close wait for
thermocouples and repeat, if still consistent try a third circuit else scratch head?

out of ballpark thought, try squeezing bead between insulation and a higher temperature, ‘dry’ joint?

stephen

Try grounding the “–” side of the thermocouple.
Or, as some documents suggest, a ≥ 1 MΩ resistor.
Floating TC’s can behave very erratically.

jcw wrote:Try grounding the “–” side of the thermocouple.
Or, as some documents suggest, a ≥ 1 MΩ resistor.
Floating TC’s can behave very erratically.

i know the max6675 requires grounding of tc- for open circuit thermocouple detection to work.
is the MAX31855 similar? seems not from the pdf

seems mainly 5v/3v3 problems(?)
MAX31855 k type site:http://forum.arduino.cc need a tongue in cheek smiley :-J

stephen

Guys

Thanks for all your responses.

The interface modules I am using are the ones with MAX31855 chips on them.

When i finally took a look right inside the oven, the thermocouples looked like this type

http://www.ebay.com.au/itm/1PCS-K-Type- … 1216872256

except without the insulation at the end of the cable.

But I noticed that the white coloured woven sheath around the cable has turned light brown with heat damage.

It looks like that woven heat coating is not designed for the temperatures that the thermocouples are exposed to.

Actually, i have a Digital Multimeter which has a similar thermocouple input, and i had taped it to the metal tray of the drawer where you put the PCBs and after my test runs while debugging the thermocouple stuff, I found the woven outer on my probes cable was also rather burnt.

I think they have used the wrong sort of probe for this, in order to save money.

Initially when I was going to use a toaster oven as reflow oven, I bought a thermocouple temperature probe, like this one
http://www.ebay.com.au/itm/Temperature- … 1556418430

but unfortunately it has now stopped working, but I will take it apart to see if its a cable fault.

I see that there are probes like this

http://www.ebay.com.au/itm/K-Type-5cm-P … 1243583738

which I think look like they are more appropriate for the job.

I didnt really intend for this to be a development project. It was supposed to be an easier way to reflow boards, than using my hot air reflow gun. But its ended up taking several weekends already.

But its hard to know at the moment, whether a IR heating based oven is ever gong to give good results, or whether I should have gone done a home made route, and buy another 2000W hot air paint stripper gun, and use it to heat a toaster oven.

I really think this could actually end up being a very good convection reflow oven.
And it would probably have been quicker to build than the time Ive spent trying to get this commerically made oven to work.