Researching placement machines - where do I start?!

>1. can you load cut tape onto the feeders, or only reels?

You can, but you have to use cover tape extender (pretty cheap, though)

>2. how about parts on trays (like Atmel AVRs in the QFP package)?

You need tray feeder. But you can "make" your own - just use some kind of support to raise tray to the pickup plane and set it somewhere aside. Then teach machine feeder's position (simple).

>3. is it possible to add glue and/or solder paste dispensing to the head?

There's almost identical dispensing machine (GDM), but I am not sure if you can combine dispense and placement heads on the same machine.

4. does it take a lot of fiddling and tuning to keep it working?

Honestly, no. The only thing we have take care about is the propper room temperature (22-26C) and occasionally lubricate linear slides and spindles. I also bought calibration kit (without it you can't replace cameras) since one of cameras died. I bought 3 cameras (just the camera, not the whole assembly) on e-bay for 50$ each, so I am safe for now...

If you can choose, take 4-spindle Flex heads. They are much simpler and require almost no maintenance. Also two ULC's are better than one. We have 4-mil front and 2.6mil rear cameras. For 0.5mm pithc that's OK. The only thing we have a problem with are some QFN packages with Palladium plating. Pins are VERY reflective and due to the camera construction look invisible.

Regarding machine block levels, only L and later are upgradeable by UIC, but forget the upgrades if you buy an
used machine.

We have I-block and USOS ver. 1.41. It seems old, but software is so well done that it can do almost anything.

One thing to look for is Ethernet card - get one if you can and then you can connect GSM to your LAN and simplify data transfer and backups. I've configured our GSM with OS/2 to act as Microsoft LanMan client so I can access shared folder on our office server.

Also, for standalone operation you will need SMEMA manual interface to simulate board-in and board-out commands (which normally come from the other equipment). Without it you won't be able to load the board into the machine.

There's a lot of other small advices I can share. If you buy GSM, we can switch to e-mail as otherwise we will flood this forum :-)

- Deni

reply »

Before you buy a manual screen printer try to do it by hand. We hand stenciled boards with 19mil pitch parts for three years with decent results, with your parts it should be simple. lay the board on a piece of ESD bubblewrap, the kind with small bubbles, and lay the stancil over it, either buy a hand squeege from Permalex or someone or make you own from any squeege blade you can get off ebay or you can even use a putty knife. At least try it, it works fine for simple boards. Put your money in the placement machine and oven to start. You can usually get one of the reps for oven profiling systems to come out and do a few demo profiles for you and that should get you started with the oven without spending any money on a profiler if the oven does not have that capability. You can also just used a meter and some long thermocouple wire.

The GSM is an excellent machine with a great history of reliability and there are a ton of used parts out there.

reply »

You can buy a couple of conveyors for loading and unloading machine or a real half aced way is to jump SMEMA cable pins with tweezers to get boards in and out. With that method, you could end up with one on the floor at some point with a smashed, delaminated corner but it works.

reply »

I would not disagree with you...GSM has 64 feeder slots...in my mind not the right machine for you...if I had something for you I might be a little pushy myself...we make machines for exactly what you need but I think your expectations of price and what I have to offer are not in sync. As a start up venture I still recommend getting on board with the OEMs...even if that means finding the right machine elswhere and then going to the OEM...they will want you pay up (machine registration) but in the end they will look at you as someone who will upgrade sooner or later. BTW...my company does not charge for machine registration.

reply »

Hi Jeffrey,

I bought my first used p&p last year, and I wholeheartedly agree the process of picking out one of these things is daunting! I am a small OEM with a strong tendency toward vertical integration. I'm just not satisfied unless I control as much of the process as possible. I bought a well used (~15,000 hrs) Philips CSM84VZ (~1992) which holds up to 84 feeders and has top and bottom vision. It weighs in at almost a ton and was clearly built to last. By making an inquiry on an auction site I wound up with an equipment broker who was involved in an upgrade where this machine was to be taken off of a running line when a buyer was found. Knowing the thing was running helped with the purchase decision. The broker was very helpful with my questions and got the thing shipped very promptly. The unit turned out to be very simple to set up and operate and I was running the first pcbs in just a few days. The machine actually paid for itself in about four months doing just one pcb that was being partially assembled by a sub. From my modest first experiences I would suggest full vision is a must-have, and don't be bashful about getting a machine that holds a lot of feeders. The slots fill up faster than you might think, and the more parts you can keep resident, the less you have to change when you switch between products.

Good luck in your search!

reply »

Regarding more info on equipment - what you need is driven by your product and process requirements and tolerance for rework, labor amt, operator dependencies ..etc. EMPF offers a bootcamp for circuit card assembly, SMT mag publishes SMT 101 guide to get cover the basics, Clyde Coombs "Printed Circuits Handbook" is a good reference. You need to understand the process before selecting equipment - maybe look for a consultant or bring on a process engineer to get things going.

reply »

I'm assuming that 18k components per month represents less than 100 boards per month. I've worked for both OEMs and CEMs and if your design is stable and your volumes are relatively stable I would highly recommend going with a CEM given the low volumes. I'm all for vertical integration as I was hired to bring SMT back in-house, but only if it makes economic sense to do so. There are many hidden costs that you will have to absorb such as training, maintenance, cost of floor space, money tied up in inventory, utilities, registrations, etc. There are decent CEMs out there, preferably local.

The hard part for you (IMHO) will be understanding the PROCESS of assembly (depositing solder paste, thermal profiling for reflow, rework, inspection, etc.), not merely choosing a particular machine (that's the easy part). Its hard enough to deal with daily issues when you have a solid background, but your skills lie in other areas. Stick with your core competency (design) and you won't be sucked into a pit.

If you absolutely must assemble in-house, I'd take a well built older machine (assuming parts are available) than a new "Yugo" type. Make sure service is local. Looks like you'll need a machine with at least 100 8mm feeder slots. I've disqualified many machines just on this alone and make sure the matrix trays don't take up feeder spaces. Every machine has pros and cons...none will be perfect. Reflow ovens and screen printers will also be part of your decision...but that's for another day. BE WISE!

reply »