Pick and Place Machine TVM802A / TVM802B

[sedelman]

0.5 MPa is only a minimal pressure requirement of under 10 psi. See the attached picture that Kimi sent indicating the size of the compressor that they are using.

I have been using California Air Tools compressor which is very quiet and I can operate it in my lab and still hear myself think.

[sedelman]

A few more details about the CMHT530P and then I'll stop hijacking this thread. The 530P has the same mechanical construction as the CMHT48VA and therefore it has encoders on all steppers. The maximum Z-axis travel is 20mm, so the tallest component it can place is under 10mm. If you're placing 10mm tall parts, you need to place them last and from right to left so that the head with part don't bump into the parts already placed. The other difference from the CMHT48VA is the fact that the maximum board size is 350mm x 250mm (as oppossed to 350mm x 350mm for the CMHT48VA). This is still acceptable to me, because I only typically use letter size panels (A4 would work also).

[uncle_bob]

@bob unless someone has a CHMT48VA who can provide facts, I'm not sure if it makes sense to start a new thread. @ServoKit is doing an excellent job reviewing the TVM802A machine. I am actually re-thinking whether it is a good thing to have embedded Linux on the CMHT48VA because it will complicate the firmware update process (for the end-user) and therefore software updates (bug fixes, feature addition, etc.) are likely less forthcoming from the manufacturer than if it was a Windows based solution.

Also, as others have pointed out, the encoders on the stepper motors are useful to detect missteps, but it would have been better if the machine utilized linear optical encoders so that the machine actually knows the actual position at all times. The encoders on the steppers add marginal value.

It looks like Charmhigh came out with a new unit that utilizes pneumatic feeders. The model is CHMT530P and appears to be the same price as the CMHT48VA. These look like Yamaha feeders.

http://www.smtpickandplacemachines.com/sale-7565529-high-speed-smt-pick-and-place-machine-with-pneumatic-feeder-vision-system.html

I just confirmed with Kimi that it indeed uses Yamaha feeders. Here is the current pricing:

Unit price: $5160 ;
Shipping to Canada door to door: $760;
Total price: $ 5180+$790 = $5920

Feeder is standard Yamaha feeder, please refer pictures as attached. Feeder
price is as follows:
8mm-- $65
12MM--  $95
16MM--   $95
24MM--  $260

CHMT530P Working video: 

Hi

My only real issue is that a lot of very useful info on this or that machine gets buried (and tangled up) in other threads. Even without buying one, you obviously have done a lot of research on their machines. Sometimes even a "this came out and then that came out and now they make this other one" sort of trail can be really useful. Zigs and zags adding this or that feature don't bother me a lot. Going to an approach that likely makes the old firmware obsolete .... that worries me. Often a decision gets made (hopefully years from now) to drop support for this or that "old feature". Yes, that's just an observation, but it does apply to the feeder change they just made.

Bob

[ServoKit]

@sedelman: Follow up on the nozzles, here's an adapter for Juki nozzles:

http://www.robotdigg.com/product/517/Nozzle+w/+Adapter+to+5mm+Hollow+Shaft+Stepper

(from a post on the Neoden thread)

Not sure if this works without tools (thumbscrew). Also no clue if it's really airtight. Given the shape of the nozzle's head I think you would need at least an O-ring or so.

Regards, Axel

[uncle_bob]

@sedelman: Follow up on the nozzles, here's an adapter for Juki nozzles:

http://www.robotdigg.com/product/517/Nozzle+w/+Adapter+to+5mm+Hollow+Shaft+Stepper

(from a post on the Neoden thread)

Not sure if this works without tools (thumbscrew). Also no clue if it's really airtight. Given the shape of the nozzle's head I think you would need at least an O-ring or so.

Regards, Axel

Hi

One of the reasons auto-change nozzles work in the first place is that they really don't need to be airtight. You design the tubing and the pump size to accommodate the level of leaks you expect to get. Having fancy seals makes the change process much less reliable (the seals hang up and wear out). Bumping the tubing up one or two sizes is cheap. Routing the bigger tubes is a bit of a hassle that you would  rather avoid, but you do it. The bigger pump (and bigger valves and ...) do cost real money.  Most quick change setups make the same sort of decision for the same basic reasons.

Bob

[ServoKit]

Presuming that the big-boy machines also have pressure sensors to detect missing / failed parts, how do they compensate for a leaky nozzle? At least with my machine the sensor is so sensitive that it will usually signal a leak when I have a larger nozzle with a larger part, probably because neither the surface of the nozzle nor the surface of the part are completely flat.

Regards, Axel

[uncle_bob]

Presuming that the big-boy machines also have pressure sensors to detect missing / failed parts, how do they compensate for a leaky nozzle? At least with my machine the sensor is so sensitive that it will usually signal a leak when I have a larger nozzle with a larger part, probably because neither the surface of the nozzle nor the surface of the part are completely flat.

Regards, Axel

Hi

Again, it's part of the design. Rather than using a super sensitive sensor with no calibration, they use something they can actually measure changes on. When the thing is in the "up and empty" situation they can (but may not always) calibrate the sensor. When they change heads from giant to next to no diameter at all, they can calibrate for that head. They might store that value or measure it each change. Storing it lest them detect the "oops missing head / wrong head" problem. Yes it's a bit more firmware. Yes it works better than the alternative. There are likely as many variations on how to do it as people doing it.

It still is not perfect, but no solution ever is. The smallest head does still set a limit on the leaks. You can't have more air through the leak than through the head. A screw down gasket fitting has a *much* lower leak rate than the flow through that small head. You can have a lot more than a 1x10^-8 cc/atm/sec leak and still detect a missing part just fine.

Bob

[ServoKit]

Ah, ok, thanks. Perhaps a bit too much to ask from my $4k machine ;-)

Regards, Axel

[sam512bb]

That seems to be the manufacturer's page, QiHe. I'm still in the market for a printer and a larger oven, can anyone comment on those offers:

Printer ($260): http://world.taobao.com/item/525291626153.htm

Oven ($1050): http://world.taobao.com/item/525259351947.htm

Addendum: Aha, the devil is in the details. They quote free shipping but only to mainland China. For overseas destinations it's 90yuan/kg (at least for the printer and oven, for the TVM802A it says contact the seller; at 60kg and this rate this would come to Y5,400 = $830)

Regards, Axel

Good day Axel,

I purchased one of those ovens (T-960) and overall it seems to be a pretty solid unit.  Given its price my original plan was to install my own control hardware and simply reuse the rest...  However, I did some basic tests and it actually works pretty good.  That being said, I will be qualifying the oven just to be sure and can report back if you or anyone is interested.

Cheers,

Sam

[ServoKit]

Sam, very interested in this. Please keep us updated.

Regards, Axel

[ServoKit]

I added a few high-res photos of the p&p head with the cover removed:

http://servokit.com/blog/2016/03/14/a-closer-look-at-the-pick-and-place-head-of-the-tvm802a/

Regards, Axel

[uncle_bob]

I added a few high-res photos of the p&p head with the cover removed:

http://servokit.com/blog/2016/03/14/a-closer-look-at-the-pick-and-place-head-of-the-tvm802a/

Regards, Axel

Hi

The pressure pump either is a blow off / clean pump (low pressure) or it's a bit higher pressure to run an optional dispensing head. You see applications where adhesive is dispensed to hold down big parts in addition to the ever (un) popular solder dispensers.

Bob

[ServoKit]

No dispensing head, the pressure is far too low, must be blow-off. So far each part came of voluntarily...

Regards, Axel

[uncle_bob]

No dispensing head, the pressure is far too low, must be blow-off. So far each part came of voluntarily...

Regards, Axel

Hi

One of the "features" you need with a blow off pump is a "scrap cup". Often this gets overlooked in the first pass design. Blowing stuck parts off randomly all over the machine is *not* a good idea. Since you can't reliably release the part, placing it in the proper location is highly unlikely. The right approach is to move the head over the cup and "spray" the part into it. Designing the cup so the part does not bounce out is a bit interesting. (Hint: think of a milk jug as a good design model and a dinner plate as a really rotten one).

Bob

[ServoKit]

The machine has one of those; it's at the end of the tray and when the pressure sensor detects a dangling (or missing part) the head travels there and drops the part. Could be that the machine actually blows at this moment to make sure the part does indeed come off, difficult to tell because it happens really fast. Re. the design, the cup is formed like a ramp, apparently some thought went into this.

Regards, Axel

[uncle_bob]

The machine has one of those; it's at the end of the tray and when the pressure sensor detects a dangling (or missing part) the head travels there and drops the part. Could be that the machine actually blows at this moment to make sure the part does indeed come off, difficult to tell because it happens really fast. Re. the design, the cup is formed like a ramp, apparently some thought went into this.

Regards, Axel

Hi

The normal process when it heads over the cup is to give the thing a blast of reverse pressure. It does not take long at all.

Bob

[ServoKit]

More tests and observations re. max. usable part height:

http://servokit.com/blog/2016/03/16/more-about-the-tvm802a-maximum-part-height/

Also, we've finalized and ordered our PCBs; I hope to have the first prototypes sometime next week for some more realistic tests.

If you read german, we have a thread about this machine over at:

https://www.mikrocontroller.net/topic/371594

Regards, Axel

[rwb]

So it can place 10mm tall parts if you use the part trays + your other components that are on the PCB are no taller than 4mm?

Plus a 10mm part has to be raised up 15mm from the base of the machine to get picked up properly correct.

Good info to know.

[uncle_bob]

So it can place 10mm tall parts if you use the part trays + your other components that are on the PCB are no taller than 4mm?

Plus a 10mm part has to be raised up 15mm from the base of the machine to get picked up properly correct.

Good info to know.

Hi

One issue with large parts is likely to be weight. In some cases bigger is also heavier. At some point a given size nozzle simply runs out of "pull" to reliably lift the part and control it while it is moved. On a machine with only two nozzles, there may not be much use for the giant nozzle on other parts. With no fast change capability, it may be faster to hand place a single large part and let the nozzles each handle a range of things that covers everything else.

Bob

[ar__systems]

With stock s/w it CAN'T place anything higher than 5mm.

With custom s/w it should be possible to place parts up to 10mm from tape and tray.

I too own TVM802B.
I used it to assemble around 160 boards with total component count around 15000.
I'm not very happy.
It has multiple issues in h/w and s/w. I intend  to fix all of those. I building new feeder, peeling mechanism. I want to replace the vacuum pump (not powerful enough), replace pressure sensors (it is probably mechanical and not very reliable). Possibly steppers drivers in future. I also want to build a nozzle changer.

We already started taking apart their network protocol. After just 3 days we can control almost all binary functions (like turn on and off pump etc) and actual movement is a matter of couple more days. To get to the point of having minimally usable s/w will probably be another month or two.

[rwb]

AR_Systems do you have any video you can share of your machine running?

[ar__systems]

Yes, I do... I just need to upload it.

Btw, to those that bash belt driven machines... vs closed loop with encoders or with lead-screw drive. From my experience the belt system even on this machine is sufficiently reliable. Judging by the lack of any kind of drift of pick up locations, I can say there is essentially no missed step and therefore no loss of accuracy - from the movement.

[protoneer]

ServoKit ... Very nice write up on your blog.

ar__systems can you comment on the reel feeder system and how it worked over 15K part placements? Do you have more details or even a Github repository with the progress you have been making on the network protocol?

I have one of these machines heading my way. Was also wondering if a vibration feeder can be added... (Not sure how the software would deal with the reel position)

[ar__systems]

ar__systems can you comment on the reel feeder system and how it worked over 15K part placements? Do you have more details or even a Github repository with the progress you have been making on the network protocol?

I have one of these machines heading my way. Was also wondering if a vibration feeder can be added... (Not sure how the software would deal with the reel position)

vibration feeder would be easy to add even with stock s/w. you just define a tray 'stack' with size of one, and it will pick up parts from the same location every time.

Reel feeders on this machine SUCK. The main problem is the peeling mechanism. It just can't be made to work reliably. It constantly changes from pulling on the cover tape too much or too little. I mean, if you have only one or two parts it is fine, you can set it up. I used almost all of the 8mm feeders. It does not happen on all of them at the same time, but there were several that just would not work consistently.  If it pulls too much it moves the carrier tape! It needs constant supervision. Other problems is the feeder block itself. the metal part that holds the tape in place is 1mm steel laser cut into shape. It scrapes some black plastic tapes and the shavings get in the way of the pick up. I would unscrew it and clean it up, but all the screws (30 of them!) are locktited and would not move. Also, it does not like clear plastic tapes - I had several of those with bigger ceramic caps. Because the tape is only held at the ends of the feeder channels and those plastic tapes are not very rigid, sometimes instead of pulling the tape through it just mangle it out of shape, bending it into upside down U. Naturally it creates a lot of mess.

The cover film from the feeder is spooled onto the smaller reels. it is routed over a rod. Works ok for 8mm tapes, but for one of my 16mm tapes the cover tape was kind of sticky and was sticking to the rod. We tried to twist it but then the head catches onto it as it moves. Oh well, we ended up wrapping a small roll of paper around the rod fixed the problem.

So there are dozens of small oversights and the result is that machine requires basically constant supervision to fix mispicks, peeling issues, and what not.

Also, don't run the s/w on W10. the docs says XP is not supported but we actually ran it on XP and it worked fine. On W10 it has some connectivity issues resulting in some commands not executed properly. I first attributed it to trying the machine over wifi, but no. I ended up with a broken nozzle and bent pick up needle as a result of trying w10.

[rwb]

AR_Systems Excellent info!

Please post video and pictures of what ever you can about this machine. I was on the fence about buying one.