I know you are just speculating here, but I have spent a fair amount of time doing exactly this, program, configure and test a bunch of rather small boards. When it was early in the process it took well over 5 minutes to program and test each board the first time. We also did a 24 hour burn in and retest which went faster, but still nowhere near 1 minute.
Of course this was after the CM did their own visual inspection which would have taken a few minutes each as well.
This board is only 4.5 x 0.85 inches and they produce them on panels of 4.
I agree that the OP is likely being a bit optimistic about the need for a more rapid throughput in the oven.
Yes, but then is you ONLY populate a couple of the boards for baking you're wasting space in the oven. I don't want to build lots of design A, lots of design B, lots of design C, etc. Rather, one or two of A (*now*... the rest later -- or NEVER if I am unhappy with the results!) and one or two of B (*now*...)
All are "bad" for you! :> Cookies are hard to discipline yourself not to "sample" as another sheet comes out of the oven every 4.5 minutes for more than an hour ("Well, maybe just *one*... on THIS sheet...").
And, if you're baking 20+ dozen at a time, no one will miss one (or twelve!) I rationalize the *need* to test each batch (sheet) as I have learned NOT to aim for consistent results! (counterintuitive) If, instead, each batch is slightly different than the other batches (larger/smaller, thinner/thicker, crisper/chewier, etc.) then folks notice them more as they are eating them. Contrast with something ultra consistent (like Oreos) where folks just eat them mindlessly.
Cheesecake, OTOH, you can't sample a corner without it being VERY obvious that a piece is "missing" (I have learned to blame it on
*mice* -- with a STRAIGHT FACE!!). So, cheesecake is the "safer" experience.
[Unless, of course, you plan on keeping the entire cheesecake for yourself! In which case, be sure there's an AED on hand!!]
The same is true of coffee cake, etc. (hardest Rx for me to get right because there is no way to even *test* it until someone cuts into it!)
Building several different designs at the same time makes it hard to effectively "stage". I.e., I figure I can stuff the first N copies of design A and *place* them in the oven (to get them "out of the way"). Then, stuff the M copies of design B and add them to the oven. Etc.
When EVERYTHING is in the oven, turn it on and wait.
By contrast, if I stuff design A and start baking them, I can't effectively begin on design B without fear of being interrupted by design A's completion, etc.
As I said, it's more a personal style issue. Easier for me to keep track of where I am with everything if I don't have to multitask as much -- one set of drawings at a time. Then, one "bake cycle".
Note that I expressly did NOT ask for higher throughput. Rather, a larger batch size. So I can do the *entire* (small) build in one burn instead of having to reload the oven with the "next design". I don't gain anything (except equipment cost savings?) by doing more SMALLER batches. I'd just wait longer for the oven (for multiple small batches).
The advanced controller plus hand characterized toaster oven approach is tough to beat for that use case. The characterised aspect is rather important for best results.
For the occasional board, you might like to look at
formatting link
for their experiences using different techniques.
I've had some success with their "skillet" method, except I used
2mm of sand to spread/diffuse the heat, and monitored the sand/board temperature with a handheld non-contact thermometer.
For small single-sided SMD boards including 100-way mezzanine connectors, small 820uH inductors, leaded ICs with 0.3mm pads on 0.65mm pitch, I was /surprised at how good/ the results were.
I don't see any issues with using larger boards, other than the size of the saucepan.
As with all techniques (and probably different components) you'll need to hone your technique with a sacrificial example.
Den fredag den 5. september 2014 07.55.33 UTC+2 skrev Don Y:
In the real world throughout and batch size is the same thing because the only ones, except you, that ask for large batch size is those who need high throughput
if you don't need throughput, insisting on doing something in 5 minutes instead of 20 is just a waste of money, space and power
I'm "touched" at your concern for *my* money, *my* space and *my* power (electric bill). Just a bit dismayed that you aren't equally concerned over my *time* or my desired work style! :>
I don't believe there is any appreciable difference in POWER consumption. As we aren't on a "demand" tarriff, here, for our electricity, there's no penalty for burning 4000W for 5 minutes vs. 1000W for 20. Granted, the service size may differ -- but I had already conceded my willingness to accomodate this in my initial post (i.e., willing to repurpose a 220V drop for the oven instead of a typical 110V appliance circuit).
As to SPACE: if I am running a smaller over multiple times to achieve the same "total output", where do the boards awaiting their turn in the oven get staged? Do I only *populate* enough boards to fill the oven -- staging them *in* the oven prior to baking? Then, once those are done, remove them, toss them in a *box* ("wherever") and start building the next group of boards?
Or, do I set aside some shelf SPACE somewhere in which to accumulate the populated boards *before* I turn on the oven for the first group? If, instead, the oven is large enough for me to stage ALL of the boards (in this group) *in* the oven while I am populating the other boards (different designs, different components, etc.) then how is this using MORE space than a smaller oven PLUS staging area would require?
As to MONEY: I'm not "strapped for cash" so while I appreciate your desire to help *me* SAVE *my* money, it's not a selection criteria. What's the resale value of a tiny oven that has run N batches vs. a larger oven that has run N/4? If I donate it to a trade school, how many students could they support in a class if their *total* work output had to fit in a smaller oven? Or, do they queue up for small timeslots at the oven and waste the rest of the class period waiting for their turn?
I *walk* more miles annually than I drive. Should I get rid of my car and reclaim all that SPACE in the garage? All that money spent on insurance, fuel, maintenance, repairs? Perhaps start hiring taxis for the little bit of driving that I "must" do? Would *you* consider that a worthwhile trade (time/money/space/convenience)?
There are only two of us here. I.e., not a lot of "laundry" to do in any week. Or, over the course of a year. Consider how much space we could reclaim if we discarded the washer and dryer and just sent the wash "out" (or, went to a laundromat). I suspect it would take many years to pay for the appliances even at a dollar a wash (and another dollar to dry). Would you consider *that* a worthwhile trade?
Despite being just the two of us, we keep a "freezer chest" (i.e., "deep freeze") for foodstuffs. It takes a lot of space
*and* consumes a lot of power. Plus the initial expense to purchase. Why not get rid of it and, instead, make more frequent trips to the store? Would you consider that a worthwhile trade?
Etc.
I very much value being able to spend my time the way *I* want to spend it. I'm annoyed if I have to run out to the store because I don't have another pound of butter stashed in the freezer. I'd be *really* annoyed if I had to drive somewhere to get my laundry done (even if I could bring something along to occupy my time as efficiently as it would have been had I stayed home) -- I can't elect to do that at 3:00AM. Or, pause a few hours between "loads". Having to wait for a taxi to arrive would be maddening -- vs. just hopping in the car and going "wherever".
Similarly, having an *oven* dictate to me the process by which I can fab a particular group of boards would be irritating. Especially for such a short cycle time (like opening the kitchen's oven every
4.5 minutes to swap sheets of cookies -- vs the 50 minute bake cycle for the cheesecake!). Do I sit and watch the oven patiently waiting for it to *command* me to service it? Do I start stuffing the next few boards while watching the clock (so I am not interrupted by the oven)? Or, do I load it once and then MOVE ON to some other activity (sort of like leaving your clothes in the dryer even after the cycle has ended)?
I prefer having the choice as to how I spend my time instead of having those choices limited by a piece of equipment (especially one that *I* presumably selected!)
Well shallzbot. OK, a real reflow oven takes about the space of a car in the garage/workshop. But it is always "touch safe" if normal precautions a observed. Then next best thing for real temperature control is a very high temperature test chamber with CO2/LN2 cooling input to drive the temperature profiles required.
I found one listing for an oven that looked to be about twice as big in length and breadth, for about three times as much.
The number of different times that particular oven shows up on eBay makes me think that (a) you'll be lucky if it works, and (b) you'll be dead out of luck if anything breaks. $250 may be a good price just for the sheet- metal work, though, so maybe it's an OK deal.
--
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
I have a T-962A that I've used for several years. Probably put
500+ boards through it. I've lost count. Others bitch but I have not had a problem with it. The keyboard debounce is not great and a couple other minor issues but it does work.
It's not a $30K reflow oven so it does have some limitations. Temp control is OK but does tend to over shoot by about +10C when trying to ramp up. You can program 2 custom profiles.
Runs off 110V 15A circuit.
I have learned not to shove it full of boards. You have to subtract about 1" from the sides as keep out. If you go this route get the larger T-962A.
For me it was a buy vs build decision and buy won out. At the time when I bought it I simply did not have time to be screwing around with a toaster oven approach.
That's my story. Like I say, others bitch it's just Chinese junk and it does have some warts but it works for me.
I use Amtech LF4300 paste. If you are going to make more than just a few of the same board it is well worth the money to get a proto stencil.
Exactly, by waiting less you get higher throughput. Regardless, the point is this is not an important thing to optimize in the grand scheme of things. Once you are set up to run it won't matter so much whether you run one, two or three batches. They take what, 5-10 minutes each? Chump change in the process involved.
From what you have said, it seems to me you would be more interested in keeping the unit small and if there are any advantages to controls that would also be important. But I'm sure you know your mind better than I do. :)
Argh!!! ------------------^^^^^^^^^^ s.b. "panellized" (I had to paLLetize a bunch of kit for shipment this week so got the wrong word stuck in my head! :< )
This seems a bit larger than the other unit mentioned in this thread -- closer to 1 sq ft. Still a bit smaller than I would like, though. (buy two??)
Is this a typo? "LARGER t-962A" (isn't that what you were describing already?)
I don't want to build. I don't want to use paste gun and Leister. I don't want to be in the board fab business. I don't want to save pennies and waste hours.
OTOH, I want to be able to assemble boards *when* I want them without having to send off to a fab house (esp for small quantities). But, I don't need to construct an "out building" just to house a small fab!
Most of my designs are small quantities - five or six pieces. Biggest (quantity) one is ~75 pieces. I am currently trying to merge designs and/or support differential stuffing to cut down on the number of layouts, etc. and drive the number of instances up. But, I am *sure* I will never get to JUST one or two designs... there are too many physical constraint differences as well as cost, performance, etc.
(optimizing designs also "wastes" my time, beyond a certain point. Easier just to settle on a reasonable compromise and let the quantities be what they may be)
You get higher POTENTIAL throughput. If you then wait six weeks before powering it up again...
I'm laying ceramic tile, here. I can *rent* a wet saw. Or, *buy* one (and "dispose" of it later). In the rental case, I have to be ready to use it for *all* my tile cutting needs in a single rental period to economize on costs. In the buy case, I can use it as needed -- work on one room at a time (and let the saw idle whenever I am drawn to other activities).
"Why buy? You don't have the sort of demand (throughput) that would justify that sort of expense? You're only using it for a few hours several times over the course of months..."
"Ah, but I can choose to use it for *one* hour today and another hour next week. And, save myself a second rental fee, second round-trip to the rental store, etc."
This approach served me well in my career. Buy the *right* tools to enable you to work as you desire. "Employees" have to work the way their *employer* desires.
And where am I staging all of these stuffed boards queued up awaiting their time in the oven?
A tiny oven means the *practical* way of using it is to stuff as many boards as will fit in a single batch (group). Bake them. Test them. Rework, etc. Then, start the process over with the next group of boards.
A larger oven lets me stuff all of the boards that I want to fab on that day. Bake *all* of them. Test *all* of them. Rework. Then be *done* (until I've programmed and deployed the boards and am ready to move on to the next design -- or, more instances of another previous design).
Size (within reason) isn't an issue. As I said, if you don't have space *in* the oven to stage the boards you are about to bake, then you need that space somewhere else.
Controls aren't an issue because I'm doing *one* bake cycle instead of multiple.
What I want is what I asked for in my initial post. I can always drive over to an old client (or a fab house) and use *their* oven (off hours, weekends, etc.). Heck, even get *their* staff to inspect the finished boards for me! Takes *zero* space. Costs *zero* dollars. But it is not the way I want to work.
One advantage of doing the whole bunch at once is that it avoids the temptation to fiddle with parameters. If your solder paste is stale or your temperatures are at tad too high for those connectors, the entire batch will be marvelously consistent.
I have an extra one as a spare. Total cycle time for a board/panel can be in the 10 minute range with cool down. The good thing about the Amtech LF-4300 is the long work time. I can get >6 hours out of it. I have thought about using both ovens but I dont feel like running another 20A circuit. I could have one baking while the other is cooling down. The problem I found was not the time in the oven. It's the time to paste and stuff the board or large panel. I have a small air powered paste dispenser with a foot pedal. For dense and or large boards that gets old quick. I'm more than happy to spend the $100 or so for a stencil. Your time can go from 30 minutes squirting pads to 3 or 4 minutes. Sometimes it takes longer to clean up the stencil than to actually paste the boards.
Wish I had the volume to justify a setup like JL has out there in earthquake country. That's a sweet line.
I mean the larger T-962A vs the T-962.
Mine is on a small roll around cart in the garage. You will stink up the house if you use it inside (any oven technique). Need ventilation. For
I think if you (I) avoid anything "fancy" (conveyors, etc.) most repairs (for a low usage device) should be pretty easy. As you said, getting the mechanical structure is more than half the battle.
I was hoping folks would comment on *actual* experiences with specific products -- instead of arguing with me to change my specifications (makes me wonder how these same folks do with clients... "lecture" them on their misunderstanding of *their* markets?? :< ).
Joe (& now Wouter) seem to be the only ones chiming in with
*actual*, non "do-it-in-your-kitchen" experiences. I'll
*gently* (I don't want to get sucked back into any "working relationships"!) try probing past clients to see what their specific recommendations are on the subject. Maintenance costs, problems encountered with usage, performance, etc. I figure "empirical evidence" is worth more than anecdotal! :-/
E.g., tradeoffs of Ir vs. convection technologies, *practical* aspects of control sophistication, etc.
Given their comments I think I may put one of those on my short list. On the one hand I've got some products to build (on which I'll probably lose money, but it's more than half charity anyway), on the other hand my one current hardware design product features a board that's 260mm long, so I couldn't use this oven for that.
The small boards I'm doing are using TFLGA packages, fine-pitch processors and SO-88 transistors. I'm getting by using hot-air reflow for the really challenging parts and the "slobber and solder wick" method for the fine- pitch parts -- but I really ought to do it right.
--
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
Have you had a need for it in that capacity? I.e., something breaking in the "primary" unit that necessitates calling on the "secondary"?
Yes. But you have to have a place to "accumulate" each stuffed board while awaiting their turn in the oven. I'd prefer to just use the oven's tray for that purpose. And, not have to *attend* to the oven by putting it "in use" before I've finished stuffing all of the boards.
[Returning to my cookie/cheesecake analogy: it would be A LOT easier to have 10 cookie sheets and an oven that could accomodate all of them (cooking each with the same general thermal characteristics) than to have to feed a new sheet into the oven -- and clear off the recently withdrawn sheet -- every 4.5 minutes! Or, shove them through on a CONVEYOR!!]
Yes. My current goal is to try to get the number of different artworks down so the "overhead" per design doesn't get silly. E.g., making qty *1* of a board starts to leave you scratching your head re: the costs involved.
Currently, I have only one design that *requires* qty 1. And, it's functionality/feature set is generic enough that I can probably
*buy* something COTS that is more cost effective from top to bottom (as long as I can get complete docs so I can code on bare iron).
But, there are still some designs that are qty 5 or qty 3... Amortizing a $100 stencil over 3 boards -- each with $20-50 of components -- just never feels "right"!
I don't want to build boards for a living. That's a *job*! :>
Likewise, I'm not fond of doing layouts. But, it's a lot easier than having to hover over another engineer's shoulder advising him of mechanical interference issues. Or, having to abandon a particular approach because "things don't fit" -- instead of being able to make a design change on-the-fly.
Rather, I will build boards (and do layouts) of necessity to be more agile in my design/fab process. Once "done", dispose of all the associated kit -- having extracted all the necessary value from it.
Ah, OK. I found a T962c which is apparently larger still. (I hadn't noticed the size difference between 962 and 962A -- I assumed the former was just a typographical omission)
I have a small room off the garage that I will probably repurpose for this use. Venting to the outdoors will be relatively easy. And, it is effectively isolated from the interior of the house. Easy to get necessary power. Etc.
Or, if that doesn't work out, set it out in the yard where the drop for the kiln is located. (i.e., it would be really nice to just be able to load up a drawer *ONCE* and carry it to the oven)
Did I mention "that's a JOB"?? :> OTOH, you get to see the real manufacturing problems that your design and layout pose. Far more helpful than offloading those problems to a fab house and never realizing that "a little tweak" could significantly improve the process/product.
My smallest boards probably barely fit the 100 pad criteria.
1.25" x 2.5" components double-sided. But, I need three (different layouts) or four to complete "one device/design" (i.e., I'm severely constrained on the overall device envelope)
And, there is a fair bit of followup work involved (thru-hole parts, socketed devices, etc.)
Mine (leister) is more than 20 years old. Bought it for one-off prototype builds and occasional rework. Well worth the investment. OTOH, not something I would want to make a career out of!
Yes. This only gets worse with age (vision, tremor, patience, etc.)
In my case, I may build "many" of a particular design, but seldom "all at once". E.g., it's one thing to buy enough bare boards to build 75 (or even *10*) of something. Quite another to also commit to the components and *assembly* before you have thoroughly evaluated and characterized the design!
OTOH, building *one* prototype of each is hardly worth the effort and expense.
I wonder if *any* "laser cutting service" could do a similar job?
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.