In-sourcing electronics manufacture
Discussion
I've been outsourcing electronics manufacturing to two different suppliers over the last 10 years, but they have both let me down repeatedly with lead times. From talking with other similar businesses, this seems to be a common problem with CEMs. We've put a lot of effort into trying to ensure that they don't run out of our stock, but it looks like they are about to run out again soon and leave us with no product for several weeks.
In the first 10 months of this financial year we have spent over £75k on PCB assembly and a fag packet calculation reckons that the CEMs run on approximately a 35% margin. I've started looking at bringing manufacturing in house and using that £26k to buy the necessary machines and go towards the employee costs of running them. We only produce around 500 boards per year, so the volumes are very low. If it goes well, we could sell the remaining capacity to do some low volume prototyping work for others.
Has anyone got any experience of setting up a low volume SMT manufacturing facility that they could share? I know that I will need to buy a stencil printer, SMT assembler and reflow oven. New or used? There seems to be a wide variety of machinery out there. Anything else I'm going to need? I have a workshop to put it all in, but will this need to be clinically clean and separate from other working areas?
I do like the idea of getting this off the ground and adding more value and capability to the business. But only if it's fairly straightforward to do. If it's going to be a nightmare or money pit then I guess I should put my efforts into sorting the suppliers out, if that's possible.
In the first 10 months of this financial year we have spent over £75k on PCB assembly and a fag packet calculation reckons that the CEMs run on approximately a 35% margin. I've started looking at bringing manufacturing in house and using that £26k to buy the necessary machines and go towards the employee costs of running them. We only produce around 500 boards per year, so the volumes are very low. If it goes well, we could sell the remaining capacity to do some low volume prototyping work for others.
Has anyone got any experience of setting up a low volume SMT manufacturing facility that they could share? I know that I will need to buy a stencil printer, SMT assembler and reflow oven. New or used? There seems to be a wide variety of machinery out there. Anything else I'm going to need? I have a workshop to put it all in, but will this need to be clinically clean and separate from other working areas?
I do like the idea of getting this off the ground and adding more value and capability to the business. But only if it's fairly straightforward to do. If it's going to be a nightmare or money pit then I guess I should put my efforts into sorting the suppliers out, if that's possible.
For the amount of boards you're looking at, find a decent supplier, I'd start with Jaltek in Luton though I've not done any business with them in a while.
The alternative is all the kit and someone who understands it to care for it and run it. Consider that 5000 components an hour is easy for a mid spec machine then if each of your boards has 250 components then you'll only keep it occupied for 50 hours a year.
Room cleanliness will depend on how small your parts are, ie 01005 resistors and fine pitch BGAs are much more sensitive to a bit of muck in the solder compared to a few DIPS and 0805 parts. If you're doing much with BGAs you'll also need an X-ray machine.
The alternative is all the kit and someone who understands it to care for it and run it. Consider that 5000 components an hour is easy for a mid spec machine then if each of your boards has 250 components then you'll only keep it occupied for 50 hours a year.
Room cleanliness will depend on how small your parts are, ie 01005 resistors and fine pitch BGAs are much more sensitive to a bit of muck in the solder compared to a few DIPS and 0805 parts. If you're doing much with BGAs you'll also need an X-ray machine.
Thanks, we don't have any BGA components - the bigger ICs tend to be LQFP or SOIC and the smallest passives are 0603. The boards are single sided and there's only one through-hole device which is the connector. So they're not difficult to assemble.
I've tried several CEMs and they all let us down. I just don't think we take enough volume to figure as an important customer. So we get bumped down the priority list and 4 week lead times become 10 weeks. We can't live with that. I would hold higher stock but if we need a design tweak then it can be months before it gets through to production.
I'm still going to look into it, I know someone who used to work for a CEM that can help advise on what we need and do our programming. I could hire them on a day rate to make product too. I don't mind having the machines sat there idle most of the time, as long as I can fire them up and get some product made in a few days it'll be worth it. For our volumes we can use a small bench top reflow oven that runs off 240V rather than a massive 5 zone 3 phase behemoth. The only thing that worries me is guaranteeing the quality of the parts.
I've tried several CEMs and they all let us down. I just don't think we take enough volume to figure as an important customer. So we get bumped down the priority list and 4 week lead times become 10 weeks. We can't live with that. I would hold higher stock but if we need a design tweak then it can be months before it gets through to production.
I'm still going to look into it, I know someone who used to work for a CEM that can help advise on what we need and do our programming. I could hire them on a day rate to make product too. I don't mind having the machines sat there idle most of the time, as long as I can fire them up and get some product made in a few days it'll be worth it. For our volumes we can use a small bench top reflow oven that runs off 240V rather than a massive 5 zone 3 phase behemoth. The only thing that worries me is guaranteeing the quality of the parts.
It really depends on what you're doing know and where you want to be in the future, ignoring the Capital expenditure as you say you're looking at a lot of training and expertise coming in house which is good if you plan long term expansion but generally these days companies want to outsource be as the retention of that expertise will cost much more and give you more headaches than the capex. Do you free issue the parts or do you currently outsource purchasing too? Do you have a test included in he service or is that done in house?
If your volumes are low and your designs fixed find an ecm that is starting up or specialises in low volume high cost stuff, like all trades and businesses the good guys cost more and relationships take time, where in the country are you? I know a few companies but they're all busy.
Unless you've got ambitions to really grown the business in that direction i wouldn't consider it to be honest for 500 pcbs a year.
If your volumes are low and your designs fixed find an ecm that is starting up or specialises in low volume high cost stuff, like all trades and businesses the good guys cost more and relationships take time, where in the country are you? I know a few companies but they're all busy.
Unless you've got ambitions to really grown the business in that direction i wouldn't consider it to be honest for 500 pcbs a year.
FredClogs said:
Do you free issue the parts or do you currently outsource purchasing too? Do you have a test included in he service or is that done in house?
If your volumes are low and your designs fixed find an ecm that is starting up or specialises in low volume high cost stuff, like all trades and businesses the good guys cost more and relationships take time, where in the country are you? I know a few companies but they're all busy.
We currently outsource all the purchasing and testing is included in the service - but it's our test rig that's used to test the parts. We developed it and we can take it back to test in-house.If your volumes are low and your designs fixed find an ecm that is starting up or specialises in low volume high cost stuff, like all trades and businesses the good guys cost more and relationships take time, where in the country are you? I know a few companies but they're all busy.
As you say, you know a few companies but they're all busy. I guess that's exactly the point. The good companies are busy and we don't bring enough business to get their attention. I have tried the smaller and more specialist CEMs but they grow and then you become a small fish in their bigger pond. We have been with two main suppliers, but I reckon I have tried more than six CEMs of various types over the years.
I do want to get into this, especially as I now know I can hire an experienced person in on a day rate to set up and run the machines on an ad hoc basis. That does make a big difference to the viability. I want it to be a growth area for us and enable us to develop and build new products more quickly. I also hope our current volume of 500 units will increase in the future.
The general feeling seems to be don't do it, but I'm not sure why? Cost, quality, hassle?
What markets are you in, in terms of the quality trail and certification etc... if you aim to sell anything into automotive, aero, military, telecoms etc... You've got a real up hill battle showing quality, the people who e been doing it years know all the right shortcuts and things to say auditors.
So you're going to need a buyer, an smt engineer (or two) some test and rework resources, a quality guy, a few hundred sq ft of clean space and a big capex.
A company like Kaisertech.co.uk will price you a lab refurb, some tools for rework and an esd safe set up.
Then get on alibaba and see what you can get for your money. To be honest the UK is not flush with smt expertise, it went east in the noughties and most people like me who started in electronics production moved into development roles, but the people are there who will consult. I'll have a look through my linked in tomorrow and see who i can dig up for you.
So you're going to need a buyer, an smt engineer (or two) some test and rework resources, a quality guy, a few hundred sq ft of clean space and a big capex.
A company like Kaisertech.co.uk will price you a lab refurb, some tools for rework and an esd safe set up.
Then get on alibaba and see what you can get for your money. To be honest the UK is not flush with smt expertise, it went east in the noughties and most people like me who started in electronics production moved into development roles, but the people are there who will consult. I'll have a look through my linked in tomorrow and see who i can dig up for you.
Another thing to think about if you build in house is component sourcing. Potentially with lower volumes you may be paying more for components as you may not have the benefit of the CEM who would buy standard parts in higher volumes for multiple end customers which means better pricing and stock on hand.
Also MOQs of components could mean you would need to store them until you used them up.. Lead time also come into effect when you buy components, unless a distributor has stock on the shelves then you would need to think about 12 weeks+ potentially.
I think generally the UK is good at low volume complex manufacturing of boards.
Also MOQs of components could mean you would need to store them until you used them up.. Lead time also come into effect when you buy components, unless a distributor has stock on the shelves then you would need to think about 12 weeks+ potentially.
I think generally the UK is good at low volume complex manufacturing of boards.
Edited by mini95 on Sunday 5th February 22:14
I used to be Purchasing Manager for a small PCB assembly firm.
It's a PITA, mostly because the customers will always need something yesterday....hence you getting bumped.
The key to getting it to work successfully is investing in quality control at the beginning of the process, if you are only managing your build that will be simpler, but when you branch out it is the re-work that kills it. Most will aim for 35% profit but it is more likely 20%.
You can source components pretty cost effectively through Farnell etc, you will probably want to buy a years worth in 1 go to save cost.
Don't forget your hazardous material certs. WEE disposal and your ISO certification if you want to sell your services, you will also need some decent software to manage stock (batch control), do MRP and ERP.
It's a PITA, mostly because the customers will always need something yesterday....hence you getting bumped.
The key to getting it to work successfully is investing in quality control at the beginning of the process, if you are only managing your build that will be simpler, but when you branch out it is the re-work that kills it. Most will aim for 35% profit but it is more likely 20%.
You can source components pretty cost effectively through Farnell etc, you will probably want to buy a years worth in 1 go to save cost.
Don't forget your hazardous material certs. WEE disposal and your ISO certification if you want to sell your services, you will also need some decent software to manage stock (batch control), do MRP and ERP.
Bertrum said:
I used to be Purchasing Manager for a small PCB assembly firm.
It's a PITA, mostly because the customers will always need something yesterday....hence you getting bumped.
The key to getting it to work successfully is investing in quality control at the beginning of the process, if you are only managing your build that will be simpler, but when you branch out it is the re-work that kills it. Most will aim for 35% profit but it is more likely 20%.
You can source components pretty cost effectively through Farnell etc, you will probably want to buy a years worth in 1 go to save cost.
Don't forget your hazardous material certs. WEE disposal and your ISO certification if you want to sell your services, you will also need some decent software to manage stock (batch control), do MRP and ERP.
This All over- i used to work for quite a big contracting firm (well biggish- £25m turnover) who manufactured PCBs off 6 lines. It's a PITA, mostly because the customers will always need something yesterday....hence you getting bumped.
The key to getting it to work successfully is investing in quality control at the beginning of the process, if you are only managing your build that will be simpler, but when you branch out it is the re-work that kills it. Most will aim for 35% profit but it is more likely 20%.
You can source components pretty cost effectively through Farnell etc, you will probably want to buy a years worth in 1 go to save cost.
Don't forget your hazardous material certs. WEE disposal and your ISO certification if you want to sell your services, you will also need some decent software to manage stock (batch control), do MRP and ERP.
Purchasing was the biggest problem- we had 6 people in the purchasing team and constantly ran into problems. Getting stuff "just in time" was a necessity as most customers wouldn't pay for over ordered stock, but didn't want to wait if there was a delay.
Frankly it was a horrid business- really really hard to do well at the margins that customers want to pay.
I know of one good company in Andover if that can help.
OP i would think very veyr long and hard before going down this route- I am pretty sure that you will swap product delays for a hundred other problems. You are much much better trying to solve your issue in another way- perhaps hholding stock on a call off. I don't believe you need to raise a new BOM for every product revision, so perhaps you can find a supplier who will hold stock, with you knowing that you might have to pick up some over stock occasionally, but with the certainty that you will have supply.
Gassing Station | Business | Top of Page | What's New | My Stuff