SparkFun Forums 

Where electronics enthusiasts find answers.

Questions about the BatchPCB service

Moderator: robacarp

By sarmitage
#91630
The other problem is you really have to heat soak for a long time with the T962 if you're doing QFN or BGA... I've not got a reliable recipe for that yet.
Yes, it has no bottom heat, so you need to slowly raise the heat to make the whole board heat up properly. I find that 5 degrees C per 10 sec step works well. You may also find that fiber washers under the board will help prevent the cold drawer surface from drawing away the heat.

It's just a simple ATmega controller
That's interesting. Mine has a LPC2144 ARM7. Either way, a rewrite would be feasibile.
By 3One
#93621
djd328 wrote: This is about right. The issue is the internal construction uses masking tape to hold the fiberglass insulation in place, and even scarier, to insulate the connection to and as grommets for the elements(!) They had what looked like floor tile pieces wedging things in here and there.

I gutted the whole thing, bought some sheet brass, new fiberglass and high temp silicone sheet at Tokyu Hands (I'm sure such things can be had elsewhere, it's just easier in Japan) and redid all the internals, building heat shields and some ducting for the fan with the brass. It's very simple construction, it's just an evening's work if you're so inclined.

I also wanted off-then-shelf... but the stock T962 wasn't safe to use. The other problem is you really have to heat soak for a long time with the T962 if you're doing QFN or BGA... I've not got a reliable recipe for that yet. Small parts or QFP, works great (did before, except the burning tape/floor tile/plastic fumes) It's just a simple ATmega controller, I'm sure it would be "only" a few days work to rewrite it's horrible firmware... but at least it mostly works in spite of it's self.
About the mods you did on the T-962... Would you be so kind and share us in more detail on how you did it? So as for those who already has a T-962... Pictures will be very much appreciated...
By DC REdD
#93720
I am also having problems with my T962. It burns the PCB directly under the heating elements while not melting the solder paste elsewhere. Detail of any modifications would be gratefully received!

DC REdD
By blakec1
#103695
I have never been happy with the skillet / hotplate. As an ME I just can't reconcile how heat will travel through a fiberglass board with multiple heat contuctors (via's and traces) and not have hot/cold spots. Many times I will put exposed copper on both sides of the board for heatsinking - and I can just feel that hotplate roasting the Power IC's due to the better thermals of the heatsink with numerous vias.

I have used hotplates on aluminum backed boards (for LED's) with some success.

I would not be comfortable using a hotplate on complex boards especially with small SMT connectors and electrolytics. Electrolytic caps can and are damaged by temperature spikes and are giving people fits as the ramifications of temp spikes during reflow may not show up for 6 months to a year. Basically they can get their lifespan cut significantly. Think of all the Mitsubishi TV's or Escort Radar detectors that have died due to dead Electrolytics. CM's will tell you that RoHS has significantly increased the problem due to the higher reflow temps.

All that being said I have tried it and it "works" but again - my comfort level is low.

I do have the TechFX controller and using the B&D oven I put together a working unit in a few days. I think I have an old rev of the controller so I may get around to buying another soon (the controller is cheap - the other parts add another couple of hundred including oven). It gives me some problems not kicking off a cycle occasionally (have to reboot the controller) - but all in all seems stable as hell. I need to rig up a vented cool-down at some point as opening the door spits out a nasty fume plume. It does a nice job but I make sure and keep boards in the middle of the unit so as not to run into issues around the edges of the oven.

It was non-trivial to convert an oven - and it makes the house lights flicker as I don't think the oven is synced to AC Zero cross (not sure about that...). Perhaps Corvette123 can answer that?

Again - though - it is a great base platform to start with and I have been very happy with the solution over the two former ovens I had and my two hot plates. I am debating dropping a couple of grand on a used desktop reflow with conveyor or building another B&D IR or two and running them in tandem.
By MichaelN
#103744
blakec1 wrote:I have never been happy with the skillet / hotplate. As an ME I just can't reconcile how heat will travel through a fiberglass board with multiple heat contuctors (via's and traces) and not have hot/cold spots.
I understand your reservations, but I've found it to work really well. What I do is carefully observe the process & apply more heat to the top using a heat gun in the areas that need it. This minimizes the overheating in the areas that heat up fastest. "Proper" reflow ovens actually have the problem you are worried about - in order to reflow the entire board, certain components will be overheated.
blakec1 wrote:Basically they can get their lifespan cut significantly. Think of all the Mitsubishi TV's or Escort Radar detectors that have died due to dead Electrolytics. CM's will tell you that RoHS has significantly increased the problem due to the higher reflow temps.
Yes, the higher reflow temperatures are a problem, but IMO the problem is the electrolytics themselves. I try to avoid using them in my designs whenever possible, since I hate the idea of components that have an inbuilt finite life. With ceramic caps available in larger values at lower prices, it is often possible to completely eliminate electrolytics from a design.
By Roko
#104253
blakec1 wrote:I have never been happy with the skillet / hotplate. As an ME I just can't reconcile how heat will travel through a fiberglass board with multiple heat contuctors (via's and traces) and not have hot/cold spots. Many times I will put exposed copper on both sides of the board for heatsinking - and I can just feel that hotplate roasting the Power IC's due to the better thermals of the heatsink with numerous vias.

I have used hotplates on aluminum backed boards (for LED's) with some success.

I would not be comfortable using a hotplate on complex boards especially with small SMT connectors and electrolytics. Electrolytic caps can and are damaged by temperature spikes and are giving people fits as the ramifications of temp spikes during reflow may not show up for 6 months to a year. Basically they can get their lifespan cut significantly. Think of all the Mitsubishi TV's or Escort Radar detectors that have died due to dead Electrolytics. CM's will tell you that RoHS has significantly increased the problem due to the higher reflow temps.

All that being said I have tried it and it "works" but again - my comfort level is low.
In the engineering world, "Perfect" is often the enemy of "Good Enough"

Hot-plating works really well, and despite the non-perfect heating of a PCB, the difference is generally negligible for hobby and prototype production. The likelihood of frying a component or significantly reducing it's life-span is pretty low and that sort of thing would be indicative of a greater problem in the way you're going about the reflow (Too much heat, not enough patience is the enemy of both cooking food and making electronics ;) ). Would I have used hot-plating in mass production? No, the temperature controlled ovens we used in MP are useful because better control means less fall-out when you're doing runs of 10,000 PCBs at a time. But there's a huge difference between MP and small runs.

We used hot plating extensively in my last job for prototyping, including very small, complicated circuits with SMT electrolytic, various 0.3-0.5mm pitch SMT connectors and ICs (some with ground slugs, some without on the same board), and never had serious problems because of re-flowing. On occasion we'd rework components if the connections didn't take, but those were more indicative of a non-uniform paste application than a problem with the reflow method. (Again, in MP there are machines that do that precisely for you, but in the prototype/hobby world it's generally done by hand with stencils).

Not to knock the toaster-oven reflow, I'm thinking about setting one up for my hobby stuff because I do double sided PCBs almost exclusively, and am tired of always putting the leadless packages on one side, and doing the other side by hand.. That's one area where hot-plating is less convenient than an oven.

(As reference, my last job was double hatting as a design engineer and manager of production engineering in an electronics company)
By jekutin
#139170
:x If someone stll happens to come here in searching for information about reflow ovens and especially T-962, here is my first impression.

I bought a Puhuit T-962. When it came, there were transportation damages. I mailed to the seller, but they did not comment.

There were two spare bulbs with the oven. In both bulbs the heads are not parallel to the tube. In one of them there is in one end written 220V 150W and in the other end 220V 100W. Which one is correct? I measure the current the oven took during heating, it was 1,37A, this means 300W. According to the specification and the label in the back it should be 800W.

I made a test reflow without a board. All seemed to go fine. There is in the display drawn the selected curve and a cursor is showing, how the reflow is moving on. The cursor seems to follow the curve very well. I measured the time with my clock and the whole sequence took 12 minutes. The clock in the display showed 7,5 minutes, that is right according to the curve. I made another test with the same results. During the heating phase the display clock is slowed down noticeably. This means that to give the user a picture that the oven makes the reflow according to specification for the paste, the computer calculates the horizontal axis time from the temperature using the curve "backwards".

So the oven is not 800W as promised, it is only 300W. Based on this low power the heating period is a lot too slow. To "compensate" the low power the internal clock, which is used to show the horizontal axis values, is slowed down so that the cursor follows the curve and the clock shows the "correct" reflow time when the process ends. This all seems good, but obviously will not result in any amazing boards.
#157168
Believe it or not, a simple toaster oven does a great job. When we first started assembling surface mount boards we used a simple toaster oven that we bought from Sears. It did a great job. Once you notice everything has melted, you just open the door and have a small fan blow over the board to cool it off quickly. For simple prototypes it works great.

Today we use a $90,000 reflow oven. But the concept is the same. It's just heating up the air around the circuit board. IR reflow was abandoned in the 80's and hot plates have very uneven heating. Trust me. Use a toaster oven.