Mini Blog: IMDES Vapour Phase Purchase - Jumbo.

[wraper]

How are you interfacing the thermocouple to the microcontroller

It is PT100 RTD, opamp + few resistors.

[wraper]

RTD position was lowered by about 5mm but it still reads lower than temp on the PCB, and there is significant inertia. Right now it works like this (PID would be useless in this case):
1. Start heating
2. Once reached 80^oC, stop heating, wait 25s (after this PCB will be about 150-160^oC hot)
3. Heat 14s, likely a little bit more would be better, like 18s.
4. Stop heating, wait 25s (after this PCB will be about 180-190^oC hot)
5. Start heating
6. Once 195^oC is reached, stop heating and turn on cooling.

[wraper]

For larger oven those figures likely need to be somewhat different.

[mrpackethead]

For larger oven those figures likely need to be somewhat different.

I suspect that the the figures will vary significnatly between the boards you put in depedning on what you are soldering.  For the boards i made that had a lot of large inductors, both the ramp up and ramp down times where much longer than if i just put in a small board with a few small passives.

[mrpackethead]

Hi Wrapper,

I'm not at work, was just wondering was the RTD wired with 2 wires, or 3 wires ( for compenstation )

Just working on something useful now..  Debating if i should use a couple of Triacs with Zero Cross for the heating / cooling rather than the relays.

[wraper]

For larger oven those figures likely need to be somewhat different.

I suspect that the the figures will vary significnatly between the boards you put in depedning on what you are soldering.  For the boards i made that had a lot of large inductors, both the ramp up and ramp down times where much longer than if i just put in a small board with a few small passives.

I tried small board and 16x20 cm panel with a lot of ground plane, was pretty close. Vapor phase has big thermal capacity, after all. Adjusting heating on the go does not seem to be possible (at least with given temp sensor location). As once temperature started to rise (when you are above those 80^oC) it's already too late to stop the heating. Solder paste will already melt instead of soak which was intended. For example, if you stop heating at 90-100^oC, then board will just reflow by inertia.

[wraper]

Hi Wrapper,

I'm not at work, was just wondering was the RTD wired with 2 wires, or 3 wires ( for compenstation )

Just working on something useful now..  Debating if i should use a couple of Triacs with Zero Cross for the heating / cooling rather than the relays.

RTD has 2 wires. Unless you want to completely rip the thing apart, you should stay with relay located on the bottom of the oven. At least mine was completely not disassembly friendly. Those nasty pop rivets  , even that relay pcb is attached with them. Also to take old logic board out, I needed to cut one screw  . It was not possible to hold it from the back, and it just rotated together with a nut. So atrocious construction. Fans are 12 volt, need to be driven at high side (I used MOSFET) if you don't want to rip it apart as well.

[wraper]

2. Once reached 80^oC, stop heating, wait 25s (after this PCB will be about 150-160^oC hot)
3. Heat 14s, likely a little bit more would be better, like 18s.

After some thinking, I think it should be possible to adjust heating time here based on the temperature rise after those 25 seconds of not heating. Edit: or extend total soak time, reduce pause time to, say, 20s (time after which temperature stabilizes). Reduce heating time too. And add one more heating and pause segment. Make heating time dependent on the temp previously reached.

[mrpackethead]

 uRGH.  i've Just taken the bottom grill of the machine, and just about had kittens.

This machine is Electricaly unsafe in many regards and i will need to do a substanital amount of work to make it safe.    There are crimp rings where there is threads of live cable showing.  There are cables goign through metal openings with no glands.

THis is disssapointing. With a little more care this machine could be so much better.

DO NOT OPERATE THIS MACHINE WITHOUT A RCD/GFCI at a minimum, better still open it up and visually check.

[wraper]

I've seen the wiring on the bottom. I broke away plastic cover which was covering most of the wiring even before attaching the new controller. Dunno if it is as bad as yours. But to dig in further, I would need to basically destroy the oven because of those stupid pop rivets. Maybe your larger oven have more access space to everything.

[Spikee]

I bought one of the early imdes units, Talked with mark on the phone for a moment since I needed the machine fast
and after I told him what I do he mentioned that he might contact me in the future to re-do the controller design + firmware.
Have not heard of him after

[mrpackethead]

Marc is quite unwell at the moment, hes having chemo treatment for cancer.

He has offered me a replacement machine, many of the issues from the older machiens have been resolved.   They dont' have a new controller, and i've offered to give him the schematics / design / code for the controller i'm putting together now.

[mrpackethead]

An update for you..

Removing the perferated grid from the bottom really assists with airflow on the cool down.  Needed to raise it up so theres a good solid 50-100mm there so its not being restricted.  The current setup with it sitting just 15mm and then the grid. makes it really tough to get cooling at a sensible rate.

[mrpackethead]

so, i am going to do a bit of messing with this..  Decided easiest way to make it work will be to make a small board that i can jumper to a Development board.  I'll probably use a PIC32 Development board as i'm familar with that and i want to add ethernet connectivity to it.   A web UI will be easy to use and setup..

If this works well, then i'll jump to the next phase and create a complete custom assmebly.

Decided i'd ditch the AC relay, in favour of a 30A Solid State relay..  I'll  switch the fans with a Fet, which also means i can PWM them if i want..   Small Op Amp to deal with the sensor and buffer it to the uP's ADC.

The quality of VP soldering is awesome, its just a pity that IDMES did'tn "finish" their product. A little bit more work and it would have put it in a whole differnet leauge.   

[RolfNoot]

I was just looking for vapour phase soldering and noticed this machine. Good to see this comes within reach for most of us.

I was thinking to buy the device but right now it's priced 1500 euro ex VAT!! Regarding the crappy build and pour process control, is it worth buying this machine?

I'm willing to spend some time on the machine, rework some of the parts and build another controller.

Would this machine be a good starting point? There are also professional used machines at eBay for around 5000 usd. Some also use aspagagus cookers...
 
Comments appreciated!

[wraper]

I was thinking to buy the device but right now it's priced 1500 euro ex VAT!! Regarding the crappy build and pour process control, is it worth buying this machine?

Unless you want to modify it, no way. But if you do, you could as well just attach heater to Horeca stainless steel pan what this oven exactly is and save 90% of the price. Even overheat protection is non functional in this piece of crap, although there is a thermal fuse. The problem is how it is mounted. Had a chance to test when I forgot there was no galden in the oven, that was some spectacular experience. I had a piece of charcoal instead of PCB and burning smell in a room for more than a month. Controller had no way to detect there was so much heat in the middle (basically grill) because of how temperature sensor is mounted.

[Koen]

Some also use aspagagus cookers...

And I'm very happy with it. This was a god-sent for my FPC connectors. A year or so in and I've only added a cheap AliExpress thermo-controller to automate it. Not a single issue, it's wonderful.

[mrpackethead]

I would say avoid dealing with these guys, they just dont' get it.   Its not safe, its not well built. its a shame because he actually was onto somethign with the concept.

I love what you can do with Vapour Phase its awesome.  What i'm litterally in the process of doign is building a four tank set up so i can get some volume going on production.   My PNP line ends up stopoing because i can't get the boards through cook fast enough. :-)

I'm making some tanks from Flat folded stainless steel.

[WaldMaker]

I own a mini-condens_it since 2 years and at first I also was impressed by the solder results.

But after soldering half a dozen pcbs with 50 WS2812 LEDs and 50 0603 Caps each I think the machine does not stick to the recommended reflow temperature curve.

More than 5 LEDs on each board either didn't work at all or some color is missing. Using a USB microscope the LEDs show cracks in their clear "window". Reworking LEDs is a pain...

On each board at least 2 Caps tombstoned.

As already mentioned by others the preheat phase seems to be far too steep.  And the cooling is probably too slow. Pro machines cool with water, so there must be a reason for doing so.

So those who get working results are probably lucky that their parts survived the ride on the wrong temperature curve.

A known reason for the cracking LEDS is their moisture sensitivity and I had the LEDs out of the sealed bag for more than 24 hours. So for the next boards I bought a small pizza oven that can hold 60-70° and I will prebake the LEDS according to the manufacturers specs.

Does anyone have experience with soldering LEDS in the IMDES machine?

[wraper]

As already mentioned by others the preheat phase seems to be far too steep.

There is no preheat phase. It just heats the full power which in addition to too rapid heating and tombstoning causes excessive galden evaporation. Therefore more galden escapes each reflow and cooling becomes slower than it could be.

A known reason for the cracking LEDS is their moisture sensitivity and I had the LEDs out of the sealed bag for more than 24 hours.

Was it moisture barrier bag with desiccant and humidity indicator?

[WaldMaker]

Was it moisture barrier bag with desiccant and humidity indicator?

The bag from World-Semi had 2 silica gel bags. But I had the LEDs out of the bag for weeks @ 45 to 60% humidity and did not prebake them at that time.

Before I ruin any of my new prototype PCBs I want to try to measure the temperature curve. I started the develop an arduino mega256 / Gameduino based  controller before I bought the IMDES fryer. It contains 4 K thermocouples so I can measure the temperature curves at different heights in the tub. But I have to figure out how to seal the lid with the four thermocouple wires and how to attach the tip of the thermocouple to a test PCB.

BTW: Here are two links that could be interesting:

http://blundell.co.uk/wp-content/uploads/2017/08/IBL-Presentation-2017.pdf

On page 51 they show their patented process of moving the PCB in the vapor.
They also preheat the PCB with IR before lowering it into the vapor cloud.


WOW! look at this: http://blundell.co.uk/product/vp-mini-jumbo-prototype-vapour-phase-reflow/

They (Blundell) just put a tape over the IMDES Logo on the temperature Controlbox 
Hopefully the security flaws are fixed.

[WaldMaker]

Hi,

I finally managed to plot a reflow cycle of my IMDES fryer.

The screenshot is from my controller I am trying to develop to control the process better than the simple temperature trigger the machine comes with.

The upper curve is the rate. +/-5°C/sec.  Time measured in seconds.

The reflow graphic in the background displays the suggested reflow curve. You can see the values Time and Rate at the top.

The measured temperature tops off @ 230 as I am using Galden 230, so the value is quite reliable.

See and judge for yourself. The drop out @ 330 was probably a drop of "cold" Galden that hit my sensor.

After another 450 seconds the temperature was still 52° C.

tL measured was 42 seconds. The tL 55 was the suggested value.

[wraper]

WaldMaker, yes, preheat will be slow without serious mods like adding IR heater on the top, no way around it. As starting point you can use profile what I made with my crude controller https://www.eevblog.com/forum/reviews/mini-blog-imdes-vapour-phase-purchase/msg1104140/#msg1104140. I still use it like this and it works pretty well. You can forget about trying implementing PID as it won't work in this case. What could seriously improve control process is changing temperature sensor and it's position. Using 2 sensors at different heights might be a good idea as well. If your temperature plot is from usual sensor with standard position, you should know it's readings are barely relevant to what PCB experiences because of very bad placement. I would suggest attaching a thin thermocouple to PCB while creating reflow profile.

[WaldMaker]

wraper,
thank you for your remarks.

I did not use the built-in RTD but used one (Thermocouple Type-K Glass Braid Insulated) with a very small tip connected to a typical PCB with capton tape right in the center of the standard hole plate. So it measured something like a 0.8mm connector close to a PCB pad.

I also think the hole plate could be responsible for the very slow cool down. If you look at professional machines they use a coarse wire mesh, or even hold the PCB by its edges and some move the PCB out of the vicinity of the hot Galden pool.

As my Arduino 2560 /FT811 based controller project contains 8 MAX31856 thermocouple sensors I intend to measure the temperature curves at different heights and locations within the vessel.

The depth of the machine (150mm) is IMHO also too small and leads to vapor rising too high.

Some professional machines use a kind of gate to "close" the vapor filled chamber.

I also added a 4 relays board, a stepper controller (tic T834) and a PWM RC Servo controller plus SD Card.

So I am able to add and control an IR preheater (maybe a temperature controlled PCB rework heater?) and an additional fluid (water?) based cooler.
The servos are for an optional gate and a lid lock and the stepper is for an optional PCB elevator.

But before I go so far I think I will try to adapt your interrupted heating with a kind of self-learning algorithm one of those pizza-oven-reflow-controllers uses.

It turns off the heater after a certain time and or temperature and then measures the overshoot. As mentioned in an other post the behavior is certainly depending on the thermal properties of the PCB and its components. So I will test with and without a typical populate PCB.

[wraper]

I also think the hole plate could be responsible for the very slow cool down. If you look at professional machines they use a coarse wire mesh, or even hold the PCB by its edges and some move the PCB out of the vicinity of the hot Galden pool.

My (likely older) oven came with wire mesh plate which should have lower thermal mass. It's a piece of crap, standoff legs are bolted directly into this mesh. Because mesh is soft, this thing bends even if you look at it funny.