EEVblog Electronics Community Forum
EEVblog => The AmpHour Radio Show => Topic started by: LabSpokane on February 23, 2015, 06:24:52 pm
-
It's been nearly a week with absolutely no mention of this episode and interview. So, I'll start.
It was a great show, with a huge amount of insight into the electronics business. If you haven't taken the time to listen, you're really missing out.
I'd like to send a huge Thank You to Joe and Mark for their candor and sharing their expertise. It makes me want to buy another logic analyzer from them even though I already have one.
-
it was interesting, my take away was they dont quite know how to run a business based on the decisions made, now they have gone and raised prices through the roof. shrug. otherwise, was a good amphour to listen to.
-
Ya, know.
I was very surprised the neither Dave or Chris brought up the experiences that Dave had talking with and visiting that alarm company in AU that moved all their PCB assembly in house and had such success with doing that instead of outsourcing the work.
I would think that Dave might have gotten some real insights from that experience that would be relevant to the situation of Saleae.
EEVblog #684 - Ness SMT Manufacturing & Assembly Factory Tour
https://www.youtube.com/watch?v=bTij7Juj5qE (https://www.youtube.com/watch?v=bTij7Juj5qE)
And based on this podcast, I did decide to order a saleae knockoff from china to play with that is based on their older designs. I mean it was $15US.
If I like it I will consider buying one of their newer models.
I consider it a 'cheap trial version'.
-
I listened to that episode and was very surprised by two things.
Firstly, that they have nine people working there, and secondly that they didn't use a contract assembler for their products (until now). Now perhaps the two are interlinked to some degree (I am sure they are) but nine mouths are an awful lot to feed if I do the rough calculations of their sales volume they talked about over the time they've been going. In fact their volumes and price points are very close to mine, and I have the sum total of one employee in my business... and that's me! (Although I do take on "temporary staff", i.e. the family, to help fill in for the peak times).
Having listened to their trials and tribulations, I do however have a lot of sympathy: barely four years ago I did my first foray into the world of manufacturing and it's not fun, especially if the closest you ever got to manufacturing was in a factory 30 years earlier doing a holiday job. If you cold call contract manufacturers you'll be promised a lot of things, but it's only when they deliver (or perhaps worse, when they don't deliver) that you know. It really is a jungle out there. My best advice? Do what you can to get a personally recommended contract assembler, and cheaper doesn't necessarily save you one penny in the short, medium or long term. Also, unless you can afford to have someone work for you to manage the relationship at the assembler's location who knows the ropes and customs, use a local contract assembler.
One comment often cited about outsourcing is that you lose some control over delivery times, and QA. Both are true. While you can't do much about delivery times other than plan ahead or throw extraordinarily large amounts of money at the problem, perhaps more importantly, you can control the QA. After all, every time you have a unit returned and replaced, your margin's completely eaten up and then some, so how do you resolve that? The answer is to give the contract assembler a quick but thorough test procedure that will weed out the obvious failures, and then do the final assembly (e.g. putting boards into enclosures, sticking on labels and serial numbers, and final testing) yourself. That way the labels will be straight, and you _know_ that every unit worked when it left your hands.
My contract assembler is relatively small and conveniently also does PCB layout, and that's done by their chief exec. Although I do my own layout, it does mean that I do get really useful free advice from the horse's mouth on layout at the end of a phone: after all, it's to their benefit not to have failed boards.
-
I listened to that episode and was very surprised by two things.
Firstly, that they have nine people working there, and secondly that they didn't use a contract assembler for their products (until now). Now perhaps the two are interlinked to some degree (I am sure they are) but nine mouths are an awful lot to feed if I do the rough calculations of their sales volume they talked about over the time they've been going. In fact their volumes and price points are very close to mine, and I have the sum total of one employee in my business... and that's me! (Although I do take on "temporary staff", i.e. the family, to help fill in for the peak times).
Having listened to their trials and tribulations, I do however have a lot of sympathy: barely four years ago I did my first foray into the world of manufacturing and it's not fun, especially if the closest you ever got to manufacturing was in a factory 30 years earlier doing a holiday job. If you cold call contract manufacturers you'll be promised a lot of things, but it's only when they deliver (or perhaps worse, when they don't deliver) that you know. It really is a jungle out there. My best advice? Do what you can to get a personally recommended contract assembler, and cheaper doesn't necessarily save you one penny in the short, medium or long term. Also, unless you can afford to have someone work for you to manage the relationship at the assembler's location who knows the ropes and customs, use a local contract assembler.
One comment often cited about outsourcing is that you lose some control over delivery times, and QA. Both are true. While you can't do much about delivery times other than plan ahead or throw extraordinarily large amounts of money at the problem, perhaps more importantly, you can control the QA. After all, every time you have a unit returned and replaced, your margin's completely eaten up and then some, so how do you resolve that? The answer is to give the contract assembler a quick but thorough test procedure that will weed out the obvious failures, and then do the final assembly (e.g. putting boards into enclosures, sticking on labels and serial numbers, and final testing) yourself. That way the labels will be straight, and you _know_ that every unit worked when it left your hands.
My contract assembler is relatively small and conveniently also does PCB layout, and that's done by their chief exec. Although I do my own layout, it does mean that I do get really useful free advice from the horse's mouth on layout at the end of a phone: after all, it's to their benefit not to have failed boards.
Agreed. Especially in San Francisco!
What do you manufacture and sell? You have sold nearly 20000 units!? If so, thats very impressive for a one man business! :-+
Did you have bad experiences with contract manufactures in the USA? I have looked at getting small runs of my products done...I can get them done for 2$ a board in China, or 25$ down the road....given my products are also around the 150$ mark, I find doing it in the USA hard to justify....or maybe that manufacturer was just quoting me bad prices because they only want to do 1000s of boards.
Im in a similar position and currently searching for contract assemblers. I am looking to have them build the PCB's, and then final assembly will be done at home.
Thanks!
-
I am UK based.
My first boards were contracted as 100 to a Chinese company and 100 to a UK company. At that time I had no idea what the demand was going to be. The board was about 1"x2" with 110 parts, mostly 0402, a sprinkling of SOT-23s and three fine pitch QFNs. The BOM was 30 line items, all but two of the parts on the top side with a crystal on the bottom and an edge connector that needed hand soldering. Otherwise it was all SMD.
For the UK manufacturer, I supplied PCBs and parts (aka "free issue"). They assured me they could do 0402. They had four weeks. The day they were due to deliver, despite interim asurances, they sent me about half a dozen boards hand soldered and my parts back, saying they couldn't do it. They were charging about $8 per board. Lesson one: never trust anyone in manufaturing based on cold calling.
The Chinese outfit insisted that they would supply the boards and parts, and assured me in writing that they would use the exact parts on the BOM, from the suppliers specified (eg, Farnell, Digikey etc). They also had four weeks to deliver, and they did. However there was a 50% failure rate because a dodgy 1.2v regulator they'd used dumped 5v onto the most expensive chip on the board, blowing that up too. They also had the brass neck to complain to me that a comnector didn't fit the board, well that's what happens when you don't buy the connector specified on the BoM. They charged $5 per board for assembly, and parts and PCB prices were on a par with what I paid anyway. Lesson two: my concept of an assurance is not the same as some others, there are fundamental cultural differences, although at the time I wasn't quite so charitable with my language.
Then the Chinese New Year hit, I was left with no domestic manufacturer, and one crap Chinese one with a minimum lead time of six weeks.
By this time, I realised that I had a hot product, each time I updated the inventory with another 100 or so it sold out in under a minute. I ended up hand placing amd reflowing nearly 200 units myself.
At this stage I was about to give up, I did not see my career as a human pick and place machine going anywhere.
Then out of the blue, I had a call from an assembler who'd been recommended to me about an hour's drive away, and who I'd asked to quote before but they were too busy at the time. I asked them when they could do 500 boards, "next week" was the answer, $12 per board, I supply PCB and parts. After biting their arm off, I jumped in my car to see them. After all, at this stage contract manufacturing to me seemed like a cowboy's convention. I pressed the flesh, saw the whites of their eyes, and with some trepidation left them with the parts for 500 boards that day. A week later, my boards turned up, all assembled and tested. That was four years ago, and I still use them for all my assembly, we have an excellent working relationship. Yes, they are 2.5 to 3 times more expensive than the Chinese guys, but I have peace of mind and don't have to employ someone to look after my interests overseas. And if there is ever a problem, like parts shortages, I can jump in my car and correct it within an hour.
-
I think your own pick and place machine doesn't make much sense, if you only have 4 products like Saleae. Might be useful for Adafruit, because they have so many different PCBs that just-in-time manufacturing is necessary, because they might sell just 10 boards per month of one type, so the setup costs, logistics etc. would be too high if they would outsorce it. But it doesn't matter for 4 products. You can order 1,000 boards a time, then the other costs doesn't matter.
And right, a local PCB manufacturer and assembler can be less expensive in the end, too, if you don't sell millions of boards. I ordered 250 populated PCBs for my Kerberos cartridge from Hupperz (http://www.hupperz.de) (less than half an hour distance from where I live). Luckily I knew the company already, because another company I worked for ordered their (much more complex) boards from them. And if there are problems, you can phone or visit them, no cultural differences (and good for me, because my spoken English is terrible). But so far no problems with the boards. Only two not working, and I didn't bother to check why, maybe just a cold solder joint or faulty part.
-
I was very surprised the neither Dave or Chris brought up the experiences that Dave had talking with and visiting that alarm company in AU that moved all their PCB assembly in house and had such success with doing that instead of outsourcing the work.
I would think that Dave might have gotten some real insights from that experience that would be relevant to the situation of Saleae.
Totally different.
One is a huge long established company, the other is tiny.
One has hundreds of products, the other has less than a hand full
One invested millions of dollars in a whole suite of gear, the others investment was tiny.
One bought an existing assembly company and staff that made the product and moved it in-house, the other just bought a machine and had to figure out how to use it.
One has contracts to make stuff for other companies to make use of any spare capacity, the other doesn't.
-
And right, a local PCB manufacturer and assembler can be less expensive in the end, too, if you don't sell millions of boards.
Indeed. And unless you are in a market that is a race to the bottom in pricing (think Arduino clones for example), you simply build that price into your product.
IMO if you aren't selling millions, and you are forced to use China for assembly pricing reasons, then you probably aren't in a market with a good future.
-
Right. I even had to increase the price of my product by 10 Euro after the crowdfunding batch was shipped, otherwise all the extra work I still have to do, like drilling and deburring the cases for the LEDs and the DIN-connectors, hand soldering some through hole parts, all the bookkeeping, and packing and shipping, wouldn't be profitable. But I still sell a few cartridges per week.
For my next product I'll outsource everything for production and just do a final test and ship the product (maybe outsourcing the shipping process, too). I talked to my assembler and they could do this. With crowdfunding projects it is no risk: If people don't want to pay the higher price, it simply doesn't reach the goal and nobody loses anything, except me a bit because I had to develop a prototype, but that is my hobby anyway :)