Author Topic: Essemtec Fox - thoughts?  (Read 29467 times)

0 Members and 1 Guest are viewing this topic.

Offline robint91

  • Contributor
  • Posts: 44
Re: Essemtec Fox - thoughts?
« Reply #75 on: October 03, 2019, 11:51:03 am »
I like it. It is very fast, currently we run it only at 50% of it rated speed because raw speed doesn't matter that much for us. We are building almost only prototypes. We are in the talks to upgrade our solderpaste setup, we are comparing the jetter vs a new screenprinter.

We did someboards  0.65mm BGAs and almost all had 0402 components, it does that flawlessly. The teaching of components is very intuitive. If you copy the numbers on the datasheet the vision system has it almost no issues in detecting a part, some parts require some manual fiddling to get it working 100%.

The machine did bring some changes to our design setup, we aligned the naming scheme of PCB libraries to the machine, and also we introduced some simple stock system with unified component names.

My goal is to get the setup time down to 30 min on moderate difficulty PCB's.

We have a few projects planned where we need to do 0.5mm BGA and 0201. But I expect no big problem with the pick and place, but more with the solderpaste.

Going from a Fox1 to a Fox2 is quite simple setup. And I would only do it if I really have speed problems and would spent more the screenprinter/dispenser.

I have a mix of HyQ feeders and CLM, If I would do it again, I would have taken more HyQ feeders. But then again, the CLMs are quite cheap compared to the HyQ and speed doesn't really matter for me.
 

Offline SMTech

  • Frequent Contributor
  • **
  • Posts: 883
  • Country: gb
Re: Essemtec Fox - thoughts?
« Reply #76 on: October 03, 2019, 02:55:19 pm »
As a CLM user I think you might find they have more parts that wear, the HyQ might not be as expensive as you think once its a few years in. As an interesting point of reference I was talking to a guy selling Fuji machines the other day and a single 8mm lane for one of those is ~700GBP. Not all that different to the per lane cost of a CLM950 cassette.
 

Offline robint91

  • Contributor
  • Posts: 44
Re: Essemtec Fox - thoughts?
« Reply #77 on: March 30, 2020, 09:53:42 am »
Finally took the time to film our machine running.



After 9 months I'm still very happy about it. It does 0201 flawlessly. And even 0.35mm pitch BGA (NRF52840 in WLCSP93) is no problem. But the difficulty for that lies in the stenciling.
 

Offline Ice-TeaTopic starter

  • Super Contributor
  • ***
  • Posts: 3148
  • Country: be
    • Freelance Hardware Engineer
Re: Essemtec Fox - thoughts?
« Reply #78 on: March 30, 2020, 05:31:26 pm »
Robin, eh? ElecGator, eh?

*Connects the dots*  ;D
 

Offline Reckless

  • Frequent Contributor
  • **
  • Posts: 367
  • Country: us
Re: Essemtec Fox - thoughts?
« Reply #79 on: March 31, 2020, 07:37:08 am »
Four nozzles (fox4) seems so yesterday.  When are they going to put dual gantry or dual heads (attached to same gantry) on that machine?  I did like their software with huge touch screen.  I never saw one run 100% speed effectively.  The most I saw was between 60-70%.  I really hope essemtec keeps tweaking and improving their software.  I don't see why AOI/SPI can't be added to the machine.  The 110V pick n place sounds like the way to go along with the neoden in6 tiny oven for protos. 

Personally I ended up with 2 fuji cp642 machines and still not fast enough.  I love the simplicity (and cheap) mechanical feeders.  I managed to get 1000 feeders for less than $2/feeder.  Somehow I can't get myself to spend $100/feeder anymore.  I know feeders typically are very expensive and no one wants quality pneumatic feeders but electronic feeders are too expensive and less durable.
 

Offline SMTech

  • Frequent Contributor
  • **
  • Posts: 883
  • Country: gb
Re: Essemtec Fox - thoughts?
« Reply #80 on: March 31, 2020, 08:17:40 am »
4 nozzles is all all you'll ever need on standalone machines, they're already quite quick enough for the applications where they are relevant. The more nozzles on the machine the bigger the head, the heavier it is to accelerate and decelerate and with it a decrease in flexibility. I'm not sure you'll see a dual gantry Essemtec machine any time soon but if you did it would most likely be on a larger machine, and even then I suspect it would be an inline only machine as the gantries would get in the way of manual loading a PCB/component tray. Gantries make loading a machine more complex as you have to load them in such a way both heads are kept busy, this mixes badly with the low volume high mix & high(ish) feeder count these machines can handle.

Electronic feeders might be expensive but they are also (in most cases) more flexible. I don't need to do anything to use a 24mm lane in any number of 4mm indexes/component outside of defining how that component is packaged it also doesn't matter where I put it on the machine. I could load it with 10Ks put it in slot3 on the machine, use it, not need it and swap the feeder for another, then need it again and load that same feeder into slot7 no input would be required from me, the machine would happily simply start picking them from that location.

I run our Paraquda at full speed without issue, a Fox2 probably moves slightly quicker with its linear motors when working at full speed and would have a secondary advantage of having smaller distances to travel.

A decent vision head for AOI or SPI is not a small thing (assuming at least 2.5D for meaningful results) nor is it a cheap thing to develop. Best to leave both of those features in the printer or dedicated machines.

Their software hasn't stopped moving on, I have a quite prehistoric v11, I think the current version is something like 24, they tweak all sorts of things, allegedly the newer versions have even improved the head travel path to get a bit more speed. Of course there are also minor quirks and glitches, improvements to the SQL (if you use it) etc.
 
The following users thanked this post: Reckless

Offline robint91

  • Contributor
  • Posts: 44
Re: Essemtec Fox - thoughts?
« Reply #81 on: March 31, 2020, 05:56:07 pm »
I really don't need the speed. Currently I do about 3000 cph on my fox 1 with 50% speed. at 100% is goes to 4200. But I rarely run it at speeds. I the time I had this machine I never produced a board in two different production runs. This is just very high mix, and in that situation you don't really need a 12 headed monster with dual gantries. I need something that is flexible and easy to setup for a prototype.

If I don't need to teach in any footprint. It takes only 15 minutes to go from generating the pick and place file in altium to having a programmed machine. Then depending on the BOM it takes one or two hours to kit the machine.



I use an aluminium plate with double sided tape to place my cut tape strips on. The standoffs are just the high that everything is in focus.

 

Offline SMTech

  • Frequent Contributor
  • **
  • Posts: 883
  • Country: gb
Re: Essemtec Fox - thoughts?
« Reply #82 on: March 31, 2020, 06:39:25 pm »
Interestingly that is slower than I would have expected, literally no quicker than the Paraquda would run if I disabled 2 nozzles.

I think its the speed from BOM to recipe that is one the software's key points, I have no idea what the exact process must look like on some other machines but it leaves me with no requirement to use 3rd party tools like PCBSynergy, the only thing I need to do sometimes is merge BOM with centroid data  if that data comes separately.

If I was looking for a feature on future Essemtec machines it would be force feedback on the placement this would help with slightly uneven boards (for whatever reason) or component heights on short runs that are incorrect and is a a nice feature you find on high-end machines. You can see an extreme example of why this can be useful in Europlacer videos.
 
The following users thanked this post: Reckless

Offline Reckless

  • Frequent Contributor
  • **
  • Posts: 367
  • Country: us
Re: Essemtec Fox - thoughts?
« Reply #83 on: April 06, 2020, 10:55:20 am »
Came across this, no plan for dual gantry:


I like Essemtec products alot (own 5 essemtec 3 zone ovens/sold 3 essemtec manual screen printers) and always looking for an excuse to buy more of their equipment.  Currently, I've gotten used to old Fuji chip shooters (tier 1) otherwise if I had a need for a flexmounter I think I'd go with a PUMA/FOX.  Although I don't know how well they would age, I like to get 20+ year service life for my money.

I'm suprised it takes so long to program and setup.  With their impressive software I would have thought programming from altium would be immediate and using smart feeders to kit super fast (under 20 mins). 
« Last Edit: April 06, 2020, 11:14:04 am by Reckless »
 

Offline SMTech

  • Frequent Contributor
  • **
  • Posts: 883
  • Country: gb
Re: Essemtec Fox - thoughts?
« Reply #84 on: April 06, 2020, 01:14:38 pm »
If all the data coming out of Altium was perfect and consistent with the engineer using the correct/the same part numbers as the ones used on the machine and steps had been taken to make sure 0degrees is consistently defined the same way by both systems, yes instant. However in real life that doesn't always happen and if its someone elses designs almost certainly never happens.

A smart feeder doesn't inherently make it quicker to physically load a strip or reel into it, and in fact you could describe Essemtec CLM feeders are moderately fiddly to load. MyCronic & Europlacer do address that directly, but they are more expensive machines and even then if you really tried to load a tape every 15 seconds into a feeder on those, you have to accept working like that will introduce errors. You have to allow time for proper care to be taken and the loadout to be checked. If you are doing small batch, you also need to accept you'll be spending some time adding covertape, or footers onto strips so they work in the feeders reliably or setting up your trays of strips. Of course if you're lucky enough to have a dedicated kitting team who loads the parts into the feeders, then loading the feeders into the machine really is a job of minutes, slide it in, check your offsets/pick heights done (it can sort of do this automatically but on v11 at least can't be quite relied on with anything in black tape).

 

Offline Reckless

  • Frequent Contributor
  • **
  • Posts: 367
  • Country: us
Re: Essemtec Fox - thoughts?
« Reply #85 on: April 06, 2020, 03:14:12 pm »
How durable are the CLM feeders?  After my Universal experience I judge pick place machines by feeder reliability. 
 

Offline SMTech

  • Frequent Contributor
  • **
  • Posts: 883
  • Country: gb
Re: Essemtec Fox - thoughts?
« Reply #86 on: April 06, 2020, 04:00:37 pm »
I would say they have their weaknesses.
 
The covertape is wound around a spool which consists of two FR4 discs (one with teeth) held together with magnets sandwiching teflon sheets and a nylon spacer, the whole thing acts like a very simple clutch. During use bits of tape glue and other crap get into that clutch mechanism and changes the tension applied to the covertape. In addition the surfaces of the magnet and the steel insert in its opposing disc get too close each other given enough time and use and start to wear, at this point the slip tension can become enough to indexes the feeder forwards revealing  a part when it shouldn't or at least moving it out of sync (at worst spilling crap loads of your parts onto the floor). Because they work in banks of up to 10 lanes per feeder one covertape motor drives all lanes at once so a dodgy tension in spool 10 can mean components being picked in lanes 1-9 make lane 10 slip.
The nylon spacer has a groove it it holding a small metal toothy but to give some grip on the cover tape, over time that causes the spacer to split and the tooth to drop out. (I'm not saying it happens a lot).
That teflon sheet of course wears, easy enough to replace but on some of mine I replaced it with a random plastic shim and that lasted longer, I wonder about the material choice. Like most custom Essemtec parts, official spares are frankly inexcusably priced.

Banked feeders like CLM mean most of the lanes are inaccessible without taking them apart, now and then tapes will curl up inside the feeder instead of feeding through (plastic or spliced typically), they are also use seriously complex mechanism for the feeder advance. I haven't dared attempt to take that section apart, but I have only had one or two go faulty, nevertheless I suspect its still a job best left to someone with a manual & a jig. While not a durability issue, banked feeders in general are not ideal, as you run through multiple jobs your parts start to get spread out across the feeders becoming less and less optimal and if you need to service one because of faulty lane you lose 10 while its out of commission.

In theory HyQ feeders are being made cheaper as their prohibitive price was keeping CLM alive when it was really supposed to be a legacy option for existing Essemtec users, in fact even the "budget" CLM feeders were becoming rather closer in price per lane to some expensive rival systems than you might expect. On a machine like a Puma or Paraquda using CLM feeders costs you ~25% placement speed as they can only advance 4mm once every 0.7seconds or so, the head can obviously pick much quicker than that, HyQ are 3 times quicker.

Now my CLM feeders are getting a little longer in the tooth I certainly wouldn't trust them with 0201 and we haven't worked them that hard, assuming a roughly even usage across the whole lot they have fed only ~150k/lane. Not even the sales or maintenance guys believe they can handle the claimed 01005.

They connect via 25 pin D-Sub, not really something intended for multiple insertions, those are connected to removable ribbons but on the machine side they are not fun to get at if they need replacing.

Plusses :
Really dense lane count
Actually pretty good at handling short strips even tho' I don't think its by design.
I've only had one total failure and two erratic function issues in 8 years so they can't be terrible. (one was a sensor failure, I'm not sure the erratic one could be explained and the other I haven't had repaired because sending erratic things off is a crapshoot in any industry)
The per lane cost (IIRC) stays roughly the same regardless of the lane widths in the cassette, by contrast rivals gun feeders tend to get more and more ridiculous in price the more uncommon the width.

On the robot side, I don't see any reason so far why it couldn't last 20 years, its a solid platform and I don't think there are any obvious points of failure you couldn't fix by simple replacing the part as long as you had access to a calibration kit. What would catch you would be availability of electronics just like anything else, if one of the drive controllers or something went EOL and yours died, that could be a pickle. I did ask if there was an expected EOL date for the Paraquda but apparently they tend not to do that to their machines, they remain supported as long as you can pay to fix them. Go back to pre-FLX era Essemtec machines and I can believe they don't last 20 years so if you found one of them mouldering somewhere, I'd avoid it.
« Last Edit: April 06, 2020, 08:52:41 pm by SMTech »
 
The following users thanked this post: Reckless

Offline SMTech

  • Frequent Contributor
  • **
  • Posts: 883
  • Country: gb
Re: Essemtec Fox - thoughts?
« Reply #87 on: April 15, 2020, 09:26:12 pm »
http://www.shawline.co.uk/product.php?id_product=116 Here's a 2nd hand Paraquda, its a G2 version with 2 nozzles so about the same speed as a FOX2 but without the linear motors. No idea what changes between a G1 and G2, I can see in the spec sheet they toned down some of the speed claims they made for G1 and at least one optional feature moved to being standard (Automatic Placement constraints).
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6570
  • Country: nl
Re: Essemtec Fox - thoughts?
« Reply #88 on: April 15, 2020, 09:27:46 pm »
How much is it going for?
 

Offline SMTech

  • Frequent Contributor
  • **
  • Posts: 883
  • Country: gb
Re: Essemtec Fox - thoughts?
« Reply #89 on: April 15, 2020, 09:37:09 pm »
If you make an account (which shouldn't result in much or any pestering) it says £27.5K, but if you asked for a quote you might get a higher number in which they include getting it to you or at least that has been my experience.

This looks like its the same machine https://www.equipmatching.com/used_equipment/2/35/409690.php that reveals another little snippet. "Posted by bangus" - which suggests Elite7 listed it there, they have moved one on before recently to make way for a new machine (they are Hanwha dealers), but this one says its a a factory closure.

Few other Essemtecs on Adopt https://tools.adoptsmt.com/equipment/search/, although I know the Paraquda has been on there for ages so who knows ho reliable that list is, an HLX 8100 is a bit of a unicorn tho', pretty sure they are Versatec C5s with a shiny Swiss shell on.
« Last Edit: April 15, 2020, 10:02:14 pm by SMTech »
 
The following users thanked this post: Reckless

Offline robint91

  • Contributor
  • Posts: 44
Re: Essemtec Fox - thoughts?
« Reply #90 on: April 17, 2020, 07:39:31 pm »
Made a second movie, running at 50% and 100%. It is having fun with WS2813C-2020
100% still amazes me. It's about 5000 cph.

 

Offline MR

  • Regular Contributor
  • *
  • Posts: 234
  • Country: tw
Re: Essemtec Fox - thoughts?
« Reply #91 on: April 23, 2020, 06:15:58 pm »
Hi,

can you make some more pictures of the pick/place head, so someone can see what it is made of?
Are they using linear motors for the Z Axis?
 

Offline VladaAca

  • Contributor
  • Posts: 21
  • Country: cs
Re: Essemtec Fox - thoughts?
« Reply #92 on: July 03, 2020, 08:15:04 am »
Hello,
maybe not right place for this question but....
If anybody have FLX (or  similar) machine with laser alignment. Can you observe is there lit any red led on laser board? On my board led signed as "1" is on. What does it mean?
Aleksandar
 

Offline MR

  • Regular Contributor
  • *
  • Posts: 234
  • Country: tw
Re: Essemtec Fox - thoughts?
« Reply #93 on: July 04, 2020, 04:35:12 am »
I wonder is there some components per hour counter included in the Essemtec software? How much CPH can you get on an average project?
"Average" of course is also very relative ... if the project is just an LED panel and picking up components nearby the CPH is of course much higher than if there's a high mix of many components from all the feeders.

Our machine (from a shitty manufacturer from Poland), can only do around 1700cph (wild mix of components) while it is rated at 3200 - 4000 cph. A faster camera could speed it up a few 100cph more but it will never be possible to get over 2000 CPH with our single head. This also includes nozzle changer time.

And how long does it take to change the nozzles with Essemtec? I think on our side it's around 3 seconds plus 2-3 seconds for picking the component.
 

Offline VladaAca

  • Contributor
  • Posts: 21
  • Country: cs
Re: Essemtec Fox - thoughts?
« Reply #94 on: July 04, 2020, 07:29:06 am »
...older Essemtec machines with CLM feeders are limited with feeder advance tact which is about 940ms. And yes, in software there is counter which show cps for current board/job.
 

Offline MR

  • Regular Contributor
  • *
  • Posts: 234
  • Country: tw
Re: Essemtec Fox - thoughts?
« Reply #95 on: July 04, 2020, 07:48:20 am »
I'll go a little bit further, do they have a simulation eg. are they able to give the operator an estimate time how long a board will take to assemble?
I'm just picking up some ideas for our pick and place application.

Can someone take a detailed picture of the pick and place head, maybe so someone can see how they implemented the vertical actuator?
« Last Edit: July 04, 2020, 08:12:06 am by MR »
 

Offline SMTech

  • Frequent Contributor
  • **
  • Posts: 883
  • Country: gb
Re: Essemtec Fox - thoughts?
« Reply #96 on: July 04, 2020, 11:57:15 am »
Why are people so obsessed with seeing how they do the Z actuation. Essemtec put that under a cover, even if you remove the cover there isn't that much you can easily deduce from a picture without completely taking it apart. The head is a very well built and very complex piece which must cost a significant chunk of the cost of the whole robot. I have posted a picture previously, on the outside you can see guide rails and a threaded outer, inside that is a sleeve of some sort full of bearings and running inside that again is the bit you can see with the nozzle mounted on the end you can see in the video at the very top is some kind of drive. None of it makes any noise you can discern over the vacuum and the tolerances are really really good, no slop anywhere. For servicing you can disconnect the hosing and flush it through with IPA, you can change the SMC filters and if you are really careful you can slightly dismantle some of that assembly and grease it all. However I stick to the manual recommendations and let the guy doing the annual service do that bit , some of how it all fits together inside can come disconnected if you lose tension and that doesn't sound like fun. Any further than that and things might move and a full calibration would be needed using the glass plate.

The software gives a live account for the CPH on the current panel, it starts that once it has loaded tools and picked the first part, personally I think that's cheating so you would need to add time for conveyor movement (or you opening the lid,securing a board and closing it again) and first tool load. I think that adds about a minute per panel which for what we typically build is very significant. The software also does have time estimates in its optimisation screens (which runs simulations to suggest where to load each part), on our version at least they are total crap, you'd do much better using your own knowledge of how long certain builds take. With a machine that can hit 6kcph reported (but more typically 5k) I find it easiest to tun that into a time based on 3600cph actual once you factor in all the other bits and bobs that are not included and the odd issue here and there. However issues are almost entirely: new shapes that need training, short/spliced tape, deep pockets or part in tubes. You encounter issues of some sort with those on all machines, except maybe the deep tape.
« Last Edit: July 04, 2020, 12:09:20 pm by SMTech »
 

Offline olkipukki

  • Frequent Contributor
  • **
  • Posts: 790
  • Country: 00
Re: Essemtec Fox - thoughts?
« Reply #97 on: July 04, 2020, 12:22:12 pm »
Can someone take a detailed picture of the pick and place head, maybe so someone can see how they implemented the vertical actuator?

I would say that you might have a bigger chance to get more info from digging into such as google patent search rather than observe some pictures...
 

Offline MR

  • Regular Contributor
  • *
  • Posts: 234
  • Country: tw
Re: Essemtec Fox - thoughts?
« Reply #98 on: July 04, 2020, 01:14:51 pm »
Quote
Why are people so obsessed with seeing how they do the Z actuation.

educational purpose / curiosity - nothing else.

Quote
The head is a very well built and very complex piece which must cost a significant chunk of the cost of the whole robot.

That's what I'm interested in yes! However those kind of fixed multihead designs are basically not so extremely difficult since the z-axis of each pickup nozzle is fixed.
I find the spinning starheads or some Fuji chipshooters most interesting, those I'd say are really complex.
I've done some tests with closed loop steppers, you can already achieve reliable turnarounds within 100 milliseconds (of course they're a bit bulky).
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6570
  • Country: nl
Re: Essemtec Fox - thoughts?
« Reply #99 on: July 04, 2020, 08:29:49 pm »
My best bet the actuator is a small linear motor. They are in many regards better than a pneumatic cylinder.
I have one from an Assembleon machine , I posted that a few years ago:

https://www.eevblog.com/forum/projects/mystery-pp-part-found-need-help-identifying-purpose/msg1398882/#msg1398882

« Last Edit: July 04, 2020, 08:32:36 pm by Kjelt »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf