LitePlacer.com forums

dmwahl wrote:I noticed that the solenoid gets pretty hot at 24V, has anyone had trouble with it overheating?

The valves I use also get hot, but not too much too touch. I have never had any failures so far. I bought a quality valve and the datasheet says it's rated 24V for 100% duty time so I hope they dimensioned the coil accordingly... I have not measured how much current it draws, might be interesting to see how much energy is wasted for heating.

dmwahl wrote:I'm putting together a quick PCB to simplify wiring of the pump and solenoid, so figured I might as well put a basic economizer circuit on it for the solenoid to keep heat down. The idea would be to initially apply the full 24V to energize the coil, then drop it down to 12V to cut the coil power by 75%. Perhaps I'm solving a problem that doesn't exist, but if the solenoid is on half the time that's a fair amount of heat (assuming equal time placing a part and going to pick up a new one).

Your approach seems very reasonable but I think it's overkill if you get a good valve. And if you buy cheap valves, just buy a few extra as spares...

IMHO the valve orrifice (how much air can go through how quickly -> air "resistance") suitability for negative pressure, seal material and opening / closing times are more relevant parameters to look at and optimize.

The circuit you describe *might* help with the latter; maybe even put a small booster to energize the coil with more than 24V and then quickly drop the voltage to not harm the coil... If you still want to build it, that might be an interesting experiment

If you want your machine to be really quick, maybe also consider a secondary air supply with positive pressure to equalize pressure inside the tube quickly.

dmwahl wrote:PCB will also include some diodes to clamp the flyback from the coil and pump, although again maybe not a problem.

That might be necessary depending on the MOSFET you use. Some already contain voltage clamping circuitry, others don't. I included flyback diodes no matter what - costs next to nothing and might save the MOSFETs from being fried.

Which valve did you use? I haven't found any good data on the skoocom one included in the kit, other than coil resistance which is way off what I measured (it's spec'd at 160+/-10% ohms and I measured 258 ohms).

The point of my circuit isn't to speed up the solenoid though, just to reduce the power and make wiring cleaner. It's easy enough to modify though, so if speeding up the close time of the solenoid has some benefits then I can try it.

mrandt wrote:If you want your machine to be really quick, maybe also consider a secondary air supply with positive pressure to equalize pressure inside the tube quickly.

That's the plan, I've got some nice medical air pumps that produce about 70kPa of vacuum and 100kPa of positive pressure simultaneously. Haven't figured out the plumbing yet, but that's half the fun

You may find a plan of my setup here:
https://groups.google.com/d/msg/openpnp ... S0xTAvAgAJ

And more details on pump used and storage tanks here:
http://liteplacer.com/phpBB/viewtopic.p ... =145#p1172

and here:
liteplacer.com/phpBB/viewtopic.php?f=11&t=175#p1532

From the top of my head, I don't know the type of valve I use - but I can check later if that still helps.

Brian Dorey also discussed his experience with solenoid valves and PnP - I found this helpful:
http://www.briandorey.com/post/SMC-Vacu ... ve-Upgrade

Just going through stuff on my desk, found a MAC 24V 3/2 valve that should be perfect. 6ms open time, 2ms close time, 0.17Cv.

Have you found giving the part a little puff after placement helps significantly? Increased speed? Something else? I'm not going to be using the machine for anything more than a few hundred parts at a time, so not a big deal if the speed isn't optimized. I just don't want placement errors.

dmwahl wrote:Just going through stuff on my desk, found a MAC 24V 3/2 valve that should be perfect. 6ms open time, 2ms close time, 0.17Cv.

Sounds good. Suitable for negative pressure?

dmwahl wrote:Have you found giving the part a little puff after placement helps significantly? Increased speed? Something else? I'm not going to be using the machine for anything more than a few hundred parts at a time, so not a big deal if the speed isn't optimized. I just don't want placement errors.

Honestly, I think I overengineered my system a bit... I wanted to mount the solenoids and related circuits onto the table and thus have quite a long tube running from solenoid to the nozzle.

Of course, during pickup the whole tube needs to evacuated and vice versa for placement. This is where the positive pressure comes in handy.

I like to compare the pneumatic plumbing to electronic circuits; basically the tube adds "capacitance" to my circuit, and the "charge" needs to be drained quickly...
The "reservoir" tanks also work in a similar fashion (this time a desired effect) and help to "smooth" the flow of air and stabilize the pressure levels.

Instead of using positive pressure to "discharge" the whole thing quickly, one could also use longer dwell times. So main advantage is speed; although we're talking about 300 to 500ms only. Camera operations usually take longer; so pneumatics are not the critical path at the moment.

Releasing the part has not been a big issue so far. With most parts the solder paste also helps; as the parts tend to stick to the paste rather than the nozzle. However, the "puff" helps to make sure smaller parts stay on the board. On the other hand, it can't be to strong otherwise you'll blow parts away... Trade off as usual

If I were to build the pneumatics again, I'd experiment with putting a small but quick 2/3 valve onto the head itself - as close to the nozzle as possible to reduce "capacitance". I'd also use a very small diameter tube from the valve to the nozzle, so an alternative shaft would be needed - again to reduce the volume of air that needs to be "sucked out" with each pick. I think in that scenario it should be fine to just equalize the pressure level by opening the valve to atmosphere / surroundings and not using positive pressure at all.

It's also worth mentioning that there is much benefit in connecting an absolute pressure sensor near the nozzle; mostly to detect failed picks / dropped parts and monitor that vacuum works. The rmod fork supports analog readout via a modified pin on TinyG; or you could make your own vacuum control board and connect that via SPI to TinyG or a secondary RS232 to PC. I always wanted to do that but have not found the time yet.

So let's not get head over heels here... I'd suggest you start small with a 2/3 valve, vacuum pump and maybe a small reservoir - that will be fine for placing the first hundreds of parts. Once that works, you can always improve and upgrade

Have fun and good luck!

mrandt wrote:It's also worth mentioning that there is much benefit in connecting an absolute pressure sensor near the nozzle; mostly to detect failed picks / dropped parts and monitor that vacuum works. The rmod fork supports analog readout via a modified pin on TinyG; or you could make your own vacuum control board and connect that via SPI to TinyG or a secondary RS232 to PC. I always wanted to do that but have not found the time yet.

How would you tell the software that the pick failed? I actually have a bunch of I2C absolute pressure sensors that read from 0-50psia and should be perfect (these), but I'm a little unclear on how I would signal the software that the pick failed and to try again or move to the next component in the tape.

dmwahl wrote:How would you tell the software that the pick failed?

I could think of many strategies - theory:

- After lifting the head from the tape during pickup measure and compare pressure level to reference value; if pressure is too high, assume no part on nozzle and retry with next part.
- Before placement, measure to check if you lost part in flight; if you assume yes, retry with next part.
- Also you could test if part was "mispicked" during pickup or moved during travel (e.g. only blocking half the nozzle orrifice), pressure level would again be different from your reference measurement - in that case you could go to "drop off" location, drop the part there and retry with next part.
- Please note that the reference values will be different for each nozzle size and might also depend on the part you pick up.

Implementation:

Depends on the software you want to use. RMod fork of LitePlacer has some support for these; OpenPNP also does. Maybe this is something Juha is interested in to include in main branch at some point?

mrandt wrote:

dmwahl wrote:I noticed that the solenoid gets pretty hot at 24V, has anyone had trouble with it overheating?

The valves I use also get hot, but not too much too touch. I have never had any failures so far.

So do mine, and I have had failures -- two of them. I made two changes:

1. Used thermal epoxy to attach a small Enzotech copper heatsink to the solenoid

2. I believe the Liteplacer wiring instructions are set up so that current is required to keep the solenoid closed (i.e. not picking up). I rewired the hoses so current is required to pick up, and absence of current releases the part. This means the solenoid is not energized when the machine is idle.

I use this micro valves: