Please HELP! Need to purchase a few machines

[LAPSKRFC]

I’m setting up an RF and mm-wave lab in a university. My idea is to purchase a set of machines using which I can do pcb prototyping, solder pasting, picking and placing the components (the smallest footprint would be 0201), and soldering (oven). In addition, we may occasionally need to fabricate a multilayer board (up to 5). We are not gonna do mass production however I also do not mind buying something automatic. We may do one design in a month!

Equipment that I have already purchased: Protolaser U4 and Protomat S104.

Now I have about $130k and I need to allocate the budget wisely.

For multilayer designs, I’m planning to buy MultiPress S and Contac S4 from LPKF. These two will cost about $30k

For solder pasting, pick and place, and oven, I’m thinking about DDM Novastar SPR-45VA for stencil printing, LE40v for automatic pick and place, and GF-12HC-HT for reflow. These three will cost about $45k

Furthermore, I’m thinking of buying a rework machine. I’m thinking about SV560 from Precision PCB Services. This is gonna cost about $22k.

Could you please tell me what do you think about this? Is there any of them that you think may cause trouble? Or can I buy something better with this budget?

I would sincerely appreciate your advice on this,

Best regards,

[Gribo]

Why do you want to fabricate your own PCB? Even going with top end US manufacturers will be cheaper and get you the desired results (impedance controlled boards, tight thickness tolerances, multi layers).

[Kean]

I understand the desire for a nice shiny lab of equipment, and preference of capex over opex when spending some funding... But I have to agree that with the volumes you're talking about that you would be better off outsourcing the PCB manufacture and assembly.  Even in low volume, you'd likely be ahead financially at least for the few years this way, considering the additional cost of having someone learn to run and maintain all of this equipment.  And the actual PCBs will be consistent with commercial offerings.  Maybe there are other needs you have relating to your research that mean you can't use standard outsourcing options?

[LAPSKRFC]

I totally understand your points. This is a budget for production equipment and I need to either spend it or lose it. In addition, students will be able to do the fabrication and get some experience of that. Assuming that I'm gonna spend the money, I would appreciate if you kindly provide your feedback on the machines I listed above?

[mairo]

Furthermore, I’m thinking of buying a rework machine. I’m thinking about SV560 from Precision PCB Services. This is gonna cost about $22k.

...

I would recommend you instead this machine to go with products from Ersa such as IR/PL 550 , or HR600/2. Pace are also great, such as IR3000. These machines use IR as a top heating system, which means you will not need to buy different nozzles. if you do go the top convectional heating way, make sure you have (or can have - buy, or manufacture yourself) all the required nozzles.

[mrpackethead]

Another Use or Loose Budget for thigns that are just going to be a waste of money..   Seriously, pcbs are only 2-3 days away from ordering, and you'll spend so long making them and remaking them.. its just not worth it.

[tszaboo]

Another Use or Loose Budget for thigns that are just going to be a waste of money..   Seriously, pcbs are only 2-3 days away from ordering, and you'll spend so long making them and remaking them.. its just not worth it.

Yeah, but the point is that university students could see the equipment before they are hired to do their first job (and be clueless). It matters. We had a complete 4 layer lab in the uni, nothing spectacular, but at least we could see it how it is done.
To be honest I dont think this is the right forum to ask the question, what to buy. You should ask it from the guys where the students end up being hired. For us it was this automotive company, they had very good idea what we needed, because it was what they had.

[mrpackethead]

Another Use or Loose Budget for thigns that are just going to be a waste of money..   Seriously, pcbs are only 2-3 days away from ordering, and you'll spend so long making them and remaking them.. its just not worth it.

Yeah, but the point is that university students could see the equipment before they are hired to do their first job (and be clueless). It matters. We had a complete 4 layer lab in the uni, nothing spectacular, but at least we could see it how it is done.
To be honest I dont think this is the right forum to ask the question, what to buy. You should ask it from the guys where the students end up being hired. For us it was this automotive company, they had very good idea what we needed, because it was what they had.

Maybe the smart students will use the experience to know that trying to manufacture multilayer PCBs by cnc routing is something they should avoid in the future. 

[mairo]

a bit as a joke, but with some truth in it: one great thing about people buying such equipment and end up never using it is that we can later acquire it for a fraction of the price (altho if this is an University, its our tax money that we are wasting when they purchase these..). The job descriptions "distributor", "reseler" largely exist because of this.

[Kean]

I totally understand your points. This is a budget for production equipment and I need to either spend it or lose it. In addition, students will be able to do the fabrication and get some experience of that. Assuming that I'm gonna spend the money, I would appreciate if you kindly provide your feedback on the machines I listed above?

Getting back to the original question, I can't comment specifically on any of those items.  I've used some LPKF gear, and found it useful, but I never pushed it hard and I know of others who can be somewhat vocal about the poor experience they had with LPKF.  YMMV in USA of course.

If anything, I'd suggest putting more of the funds towards assembly and rework gear and not the PCB manufacturing side.

[coppercone2]

Am i the only person that actually appreciates the thought flow that comes with rapid prototyping?

I feel like im jerking off and then i gotta wait 3 days to cum. Look ill pay you to fonish sooner so i can think strait instead of putting everything into storage buffers in my head.

Fucking turn based working. I want warcraft not civilization. are you buying goverment bonds with the lab money too?

I can do other things with my thoughts then act like a god damn sequencing machine. Way to be bored.

Teach your students that shit ignore the haters they might actually renember something thatw way.

I actually find delays to be highly irritating and lead to disorganized thinking and creativity loss.

This would kick ass in a class.

[SMTech]

I'm firmly in the camp of don't manufacture the PCB and think this is even truer for your application, where some very weird and wonderful PCB types can used to achieve the desired result. I would sell that kit and move the cash to your other requirements but that too has some issues. If you push down to 0201 that makes any manual process fiddly/difficult, not a task anyone will thank you for being given. At 0201 even many automatic machines will struggle, 0402 is about where cheap Chinese machines stop, they might have a stab at smaller but the lack of linear encoders, flimsy belt drives and the accuracy of their replica component feeders doesn't make this reliable. The same is probably true of machines from DDM & Mechatronika they do claim 0201 capability but how good it is might not be what you expect as lower end machines really like to push the bounds of what they can do. To reliably do 0201 I think you would really be looking at a nice used big production machine from this side of 2010, these should be pretty affordable and unlike the little ones have some relevance to what they might find in the real world.

[coppercone2]

Do group projects and take turns doing soldering. Ee needs socialization.

[Kean]

Am i the only person that actually appreciates the thought flow that comes with rapid prototyping?

I'm a big fan of rapid prototyping, and have multiple CNC machines, 3D printers large and small, laser cutter, desktop PnP, and so on.
10 years ago I was regularly making double sided PCBs on the CNC, but I very rarely do today.  They're so cheap, arrive in less than a week, and I have more than enough work to get on with in the meantime.  If there is a particular part of the circuit I'm not sure about, then I will usually just prototype that on matrix board.

RF is of course a bit different, and I know many RF PCB engineers rely on their copper tape and a sharp knife.  Gets a bit tricky at mm-wave scale, so a high accuracy CNC like the LPKF tools make sense.  But they take a lot of effort to learn & maintain, and cost a lot in consumables.  More importantly, the results don't then always directly translate to something traditionally manufacturable.  But you can iterate (and fail) fast... 

[MR]

I'm firmly in the camp of don't manufacture the PCB and think this is even truer for your application, where some very weird and wonderful PCB types can used to achieve the desired result. I would sell that kit and move the cash to your other requirements but that too has some issues. If you push down to 0201 that makes any manual process fiddly/difficult, not a task anyone will thank you for being given. At 0201 even many automatic machines will struggle, 0402 is about where cheap Chinese machines stop, they might have a stab at smaller but the lack of linear encoders, flimsy belt drives and the accuracy of their replica component feeders doesn't make this reliable. The same is probably true of machines from DDM & Mechatronika they do claim 0201 capability but how good it is might not be what you expect as lower end machines really like to push the bounds of what they can do. To reliably do 0201 I think you would really be looking at a nice used big production machine from this side of 2010, these should be pretty affordable and unlike the little ones have some relevance to what they might find in the real world.

0402 is already tricky, 0201 needs a stable temperature, if you just miss that a little bit you will run into big problems with it.
And the setup is just shit...

Mechatronika is pretty lame, I have written several reviews about their top of the line product already.
They pretty much don't care if you run into problems with their machine.
And worst of it is that they would like you to go to a training to avoid the obstacles which they cannot handle with their software.
Interestingly we have reverse engineered the RS232 controlling commands and our application is able to handle all the stuff they can't with their own machine.

Their machine has a high potential, their software is rubbish, and I have no trust that they will ever change (as they had over 10 years to eg. implement stuff like multiselect, but they are happy that the operator of their products spends endless time on clicking the shit out of their application).

But if we write about "their" machine, we better write about Delta Servo drives, and Delta Servo Controllers (from Taiwan, yes they are good quality and after studying the configuration just plug and play), or off the shelf sub 10 USD CCD cameras from China. Whatever they did by themselves (electronically wise) is really low quality. The mechanical build itself is okay but stuck since they aren't developing it further anymore (asked for a multihead upgrade - not planned, asked for a tray feeder due to the limited space - not available)

Since customers now have the choice to go for Chinese PP machines, they aren't earning that much anymore and increase the price of several parts by 50-100% (nozzles a few years ago 30-40 EUR, now 2019.. 90EUR+; a machine service upgrade initially 9800, then 10800, now 14800 EUR).

The end of the story Mechatronika is not worth the money they want to charge for their products in 2019.