SRT 1000, WHAT ELSE?

For Dean and the rest of you interested,

I have reconciled many of my thermal issues with this machine and its peculiarities. The bottom side heater seems adequate, at 750 watts (not going to spend 10 grand extra for 500 more watts no matter what) and the top heater is behaving pretty well and all this on our infamous 16 layer boards using 432 permiter BGA's.

What may interest some of you is we use no solder paste on this particular board type. Not suprising to most of you, we have a few minor problems as non-wetting or ball flow on occassion because of co-planarity and other issues. We cannot until certain design features are corrected.

What I would like to discuss now is the SRT's "bogus" force setting software. It does in no way communicate with the system's mechanical devices. In other words, it doesn't work with the mechanism in place as a linear spring "lightening" the pickup tube and top heater array, and the hall effect sensor (and others) to change, in any way, the force applied to the placement sequence.

If any of you find this interesting, I would appreciate knowing your "fixes" and their effects. I spent the entire day, and more, adjusting and readjusting stops, springs, switches and the like until I came up with an almost workable solution while using solder paste and how far into it, with how much down speed/force was applied.

There is much more. I hope it is all correctable, coupled with the tray positioning cylinder mounting "kluge" that often stops, or not, the tray where specified.

This is not a negative message concerning the machine. It merely is a comment on fact as our machine now occupies space and performs reasonably well while being capable of so much more.

Earl Moon

reply »

| For Dean and the rest of you interested, | | I have reconciled many of my thermal issues with this machine and its peculiarities. The bottom side heater seems adequate, at 750 watts (not going to spend 10 grand extra for 500 more watts no matter what) and the top heater is behaving pretty well and all this on our infamous 16 layer boards using 432 permiter BGA's. | | What may interest some of you is we use no solder paste on this particular board type. Not suprising to most of you, we have a few minor problems as non-wetting or ball flow on occassion because of co-planarity and other issues. We cannot until certain design features are corrected. | | What I would like to discuss now is the SRT's "bogus" force setting software. It does in no way communicate with the system's mechanical devices. In other words, it doesn't work with the mechanism in place as a linear spring "lightening" the pickup tube and top heater array, and the hall effect sensor (and others) to change, in any way, the force applied to the placement sequence. | | If any of you find this interesting, I would appreciate knowing your "fixes" and their effects. I spent the entire day, and more, adjusting and readjusting stops, springs, switches and the like until I came up with an almost workable solution while using solder paste and how far into it, with how much down speed/force was applied. | | There is much more. I hope it is all correctable, coupled with the tray positioning cylinder mounting "kluge" that often stops, or not, the tray where specified. | | This is not a negative message concerning the machine. It merely is a comment on fact as our machine now occupies space and performs reasonably well while being capable of so much more. | | Earl Moon | My friend...I'm not exactly sure of the problem but I'll go out on a limb and look foolish (just for you)

My understanding of the placement force is as follows. 1. Zero placement force is not possible. It (force) can only be measured with a load cell . The zero-force actually equates to about 30 grams of force. The pickup tube assembly consists of a main spring and a counter ballance spring assembly. The counterballance is tuned to allow the mechanics to rest on the bottom stopper. When contact is made, the main spring elongates (upward) and the break-away sensor is activated (typically 0.004 inch). The distance from the rersting stopper to the "active" sensing position of breakaway is measured and stored in some .ini file. This is a compensation value used as a reference to begin calculating force (in other words hard contact is made). Your spring is characterized over a range of stepper motor steps. example: 10 steps past break-away may equal 10 grams, 40 steps = 40 grams etc. This is measured (and calibrated) using a load cell. In fact the info is kept in the Force.ini file as a 'look-up" table. Aside from the mechanics..the software sequences must be set to use the force.ini with the command "pickup force" . This command references the force setting in the part description of each individual job (program). True, the machine doesn't have active feedback for a force measurement. It only knows the breakaway sensor is active therefore drive down a certain number of motor steps to achieve some characterized force. I think this is very cleaver and cost competitive solution. I have verified my force settings only once and they were dead on. the only need to calibrated is if the spring breakes. Which seems rare since it operates in a very linear range. I hope this helps, Earl.

reply »

| | For Dean and the rest of you interested, | | | | I have reconciled many of my thermal issues with this machine and its peculiarities. The bottom side heater seems adequate, at 750 watts (not going to spend 10 grand extra for 500 more watts no matter what) and the top heater is behaving pretty well and all this on our infamous 16 layer boards using 432 permiter BGA's. | | | | What may interest some of you is we use no solder paste on this particular board type. Not suprising to most of you, we have a few minor problems as non-wetting or ball flow on occassion because of co-planarity and other issues. We cannot until certain design features are corrected. | | | | What I would like to discuss now is the SRT's "bogus" force setting software. It does in no way communicate with the system's mechanical devices. In other words, it doesn't work with the mechanism in place as a linear spring "lightening" the pickup tube and top heater array, and the hall effect sensor (and others) to change, in any way, the force applied to the placement sequence. | | | | If any of you find this interesting, I would appreciate knowing your "fixes" and their effects. I spent the entire day, and more, adjusting and readjusting stops, springs, switches and the like until I came up with an almost workable solution while using solder paste and how far into it, with how much down speed/force was applied. | | | | There is much more. I hope it is all correctable, coupled with the tray positioning cylinder mounting "kluge" that often stops, or not, the tray where specified. | | | | This is not a negative message concerning the machine. It merely is a comment on fact as our machine now occupies space and performs reasonably well while being capable of so much more. | | | | Earl Moon | | | My friend...I'm not exactly sure of the problem but I'll go out on a limb and look foolish (just for you) | | My understanding of the placement force is as follows. | 1. Zero placement force is not possible. It (force) can only be measured with a load cell . The zero-force actually equates to about 30 grams of force. The pickup tube assembly consists of a main spring and a counter ballance spring assembly. The counterballance is tuned to allow the mechanics to rest on the bottom stopper. When contact is made, the main spring elongates (upward) and the break-away sensor is activated (typically 0.004 inch). The distance from the rersting stopper to the "active" sensing position of breakaway is measured and stored in some .ini file. This is a compensation value used as a reference to begin calculating force (in other words hard contact is made). Your spring is characterized over a range of stepper motor steps. example: 10 steps past break-away may equal 10 grams, 40 steps = 40 grams etc. This is measured (and calibrated) using a load cell. In fact the info is kept in the Force.ini file as a 'look-up" table. | Aside from the mechanics..the software sequences must be set to use the force.ini with the command "pickup force" . This command references the force setting in the part description of each individual job (program). | True, the machine doesn't have active feedback for a force measurement. It only knows the breakaway sensor is active therefore drive down a certain number of motor steps to achieve some characterized force. I think this is very cleaver and cost competitive solution. I have verified my force settings only once and they were dead on. the only need to calibrated is if the spring breakes. Which seems rare since it operates in a very linear range. | I hope this helps, Earl. |

Yes PHD, it does help, but I already knew much of what you say. Interestingly, Ernie got wind of my posting and called me today after my experimentation with the "spring" - his brain child.

I explained that I was not being negative about the bogus comment but was attempting to garner responses such as yours. I too think the setup is kind of neat.

I simply would like to see more two way interface from software to machine and back. Also, I would like to see "0" get closer to itself that 35 grams. What the hell, I can't have it all but a nice photoelectric cell/sensor located pre-place level would be a nice touch.

Just thought I'd pass on one more thing as we placed parts today using a roughly centered glue dot after solder paste printing. We ran the structure through the glue cure oven, then reflowed the parts on the SRT. The results were interesting in that all our problems with bridging to way too close traces and pads went away as we now had a fixed clearance and coplanarity between the part and board before and after reflow.

What we ended up with were columns as in CCBGA's. We might do some reliability (HALT/HAST) testing to see if this approach has any practical application. Just thought I'd throw that one in.

Dean, thanks again for your timely and informative response.

Earl Moon

reply »