Seed thread to start things off.
|Andrew Johnston||25/05/2014 00:02:39|
5726 forum posts
Looks a darn sight better than the boards I made as a kid, using transfers and those stupid pens, Day*** something?
If it's any consolation there were a few foul ups (not mine fortunately) on the professional board. The first time round the assembly company boo-boo'd and all the BGA balls were shorted together. Our client then changed to a new assembly company. Second time around when the boards came back the PCB company had 'helpfully' covered all the isolation gaps with small lozenges for copper balance. If you look closely you can see some of them remaining in non-critical errors. Strong words were had along the lines of it's my board and nobody, but nobody, mucks about with it without asking me first!
|Gordon W||25/05/2014 09:47:02|
|2011 forum posts|
Re making LEDs flicker- I understand this is to simulate oil lamps? If you overload and use old style components you will get oily black smoke as well, just a thought.
31 forum posts
Could you please explain "copper balance"..
31 forum posts
2904 forum posts
That link seems to focus mainly on PCBs warping but I suspect what Andrew is referring to is slightly different.
This link **LINK** summarises copper balance for both warpage (1) and plating thickness (2). In my experience, adding lozenges for copper balance is usually to address the latter.
Like him, I've had problems with unauthorised lozenges(!), as they are usually added by the PCB manufacturer, sometimes without asking.
31 forum posts
|John Stevenson||26/05/2014 15:00:42|
5068 forum posts
Two projects I'm working on that fall into this group but at the moment both are background projects because of time.
First is I have an electronic gear hobber as described in MEW 108, it's based on a Victoria U2 and has literally cut probably thousands of gears, was going to put hundreds but it wouldn't be true thinking about the quantity runs done.
Anyway it has a problem in that the gear has to be done in one pass as when its starts up and slows down you can see it get out of step. I say a Video on You tube of an electronic hobber built by Andy Pugh in which it didn't do this and was based on Linux CNC so I got a local Linux guru to build me the same box with the OK from Andy.
Test running is perfect but now need to transpose all the new gear onto the Victoria.
Second project is an open source [ at the moment ] CNC system that works on Windows, linux and some tablets that uses USB.
Called EasyCNC the link is here **LINK**
Bought all the gear to build two boards, one is built but as yet untested, time again, I can see this being a winter project. I don't know if once built and tested I will do anything with it other than possibly a special purpose machine or little router but I feel that the people who start these projects need as much encouragement as possible.
|32 forum posts||
Neil, thanks for the names of those 2 books on non-logic electronics. Do you or does anyone else have 1 or 2 books they could recommend to study the logic type of electronics please ? Thanks, StephenS.
|Neil Wyatt||30/05/2014 11:04:11|
18320 forum posts
What sort of projects are you looking at? The books I mentioned give a basic introduction to interfacing with and using 'logic' chips, but don't go into great detail. In practice there are three areas you might get into:
Basic logic gates (TTL/CMOS) these days very few people use these beasts unless they have to, but some time playing around with them is worth the effort as one day sticking a NAND gate or flip-flop into a circuit might save you a huge re-design.
'Bespoke' logic chips from simple counters to complex display drivers. For these you generally just need to know the rules for wiring them up (in Mike Tooley's book) and follow the example circuits in the datasheet -usually there's enough there for you to use them as building blocks. You may often find that to 'work' these chips you need:
Microprocessors - the computer on a chip, which is 'where it's at' these days.
The best way into microprocessors isn't a book, it's buying a simple project board, such as an Arduino, for which there will be masses of free tools, support and tutorials on the web, and start playing.
You can dive straight in with micros, but learning about discrete devices and other types of chip will be a big plus when you want your projects to interface with the 'real world'.
(There are other things like FPGAs, but life's too short, sadly).
|Russell Eberhardt||30/05/2014 15:01:23|
2600 forum posts
Interesting project John.
EasyCNC looks as if will be good when the bugs are ironed out. Less geeky than EMC2 (LinuxCNC) and cheaper than a Mach3 license. I don't have the interface but I've played with the software in simulation mode. Running it on Linux the toolpath display is blank, it's OK in Windows7 and runs programs with straight lines OK but programs I have tried with arcs throw up errors and won't run. Those programs were generated from dxf2gcode and run OK on EMC and on Mach3.
I like the idea of using an Android tablet keyboards are a pain in the workshop.
|Peter G. Shaw||30/05/2014 20:56:14|
1191 forum posts
I'm now going to upset a lot of people.
In my opinion, the amateur who generally does not have fancy test & design equipment, should restrict themselves to discrete component designs, eg transistors, resistors, capacitors etc, and the simpler ic's such as the CMOS 4000 series, 741 op-amps and CMOS 555's. The advantages of this approach are that a) you know how it works; and b) it does not depend on specialized components which may not be available a few years down the line. Other advantages are that you can then tailor the design to suit your parameters/requirements and not what someone else has decreed, and that you can generally use a higher working voltage.
The downside is that performance may not be as good but for a one or two-off circuit does that really matter? Another problem is that of having to learn how to design using what some would describe as obsolete technology and here I would recommend two books, both out of print as far as I know, but should be available through the library service: these are T. K. Hemingway's Electronic Designer's Handbook and the companion Circuit Consultant's Casebook.
Personal experience, ok from the mid-'70's, is that TTL is best left alone, whilst standard 4000 type CMOS is easier to use. BiPolar 555's can produce very heavy current spikes which the CMOS version overcomes, and 741's are reasonably easy to use. Using CMOS allows for voltages up to 15-18V which makes design very much easier. Transistors can easily be obtained which will work, within reason, at any voltage you care to use. In the mid '70's I was using them on plus & minus 50V, mainly the latter, in an environment which was pretty harsh electronically. (Huge spikes flying around.)
I have a 9 speed belt driven lathe. Barring complete motor failure, this lathe is likely to be still running when I fall off the perch. I also have a variable speed electronically controlled milling machine. Should the electronics fail on that, unless I can decipher the circuit and either repair or redesign, then an electronic failure may well result in an expensive door stop. And before anyone says that electronics don't fail, I would point out that this particular series of milling machines is/was noted for electronic failure (and plastic gear failure).
I'm well aware that there are people on this forum who are into electronics professionally - I'm simply suggesting that the amateur without the benefit of formal electronics training and/or facilities should either stick to discrete component designs that can be readily understood, or go out and buy professionally made items.
For what it's worth, I wouldn't even attempt an electronic design, even with discretes, without my trusty oscilloscope to analyze the circuit when it fails to work as hoped.
Peter G. Shaw
|Andrew Johnston||31/05/2014 15:19:36|
5726 forum posts
A few random thoughts:
1. In some ways FPGAs are easier to use than the modern processors; smaller datasheets for a start. Smaller FPGAs can be useful for soaking up functions that can't be done in software, but require more than simple logic. The last FPGA I used implemented a clock distribution scheme and synchroning logic for voltage measurements. There were 12 channels of measurement, all isolated. For each channel I designed a very simple open loop, ie, on/off, at a fixed duty cycle and no feedback, power supply. To prevent large current pulses if all the supplies switched together I used an FPGA to create a time shifted clock to each supply. The FPGA also implemented steering of the SPI channels and some other random logic. At the lower end the design software for FPGAs is free and you can design in logic diagrams or write the code in VHDL or Verilog.
2. I would think that standard TTL is dead in the water. For general purpose stuff I use the HC family; TTL functions, but runs over the wider voltage range with low quiescent currents and uses CMOS voltage levels. There are also some useful single gate logic functions available which are useful for getting one out a hole when there is a conflict over logic levels, or you need to level-shift or tri-state a signal.
3. I would think that the Art of Electronics would cover all one needs to know about the theory and practise of logic design unless you're going to get involved in much faster and more complex systems.
4. There are way better opamps around than the 741. Why design with both hands tied behind your back? For el cheapo, and bomb proof, devices use the LM2901/2/3/4 series. Dual and quad opamps and comparators. Not particularly good specs or fast, but cheap and robust. Opamps running from 1V to over 100V are widely available with specs that are orders of magnitudes better than the 741. A lot of old design techniques for opamps are simply there to get round limitations of the old ICs, they're just not needed now.
5. I wouldn't be seen dead using a 555.
6. I suspect that a lot of the problems associated with 'hobby' level electronic motor control is down to poor design, rather than the reliability, or otherwise, of the electronics. Two things that will kill electronics are heat and over-voltage. Good design can solve these issues, as well as taking account of faults, like the motor stalling, and designing the circuit to cope with that.
7. A 'scope is definitely useful, particularly for things like switch-mode power supplies which can have a habit of going do-lally, while still providing an output voltage.
|Peter G. Shaw||01/06/2014 16:36:15|
1191 forum posts
A lot of what you have said simply proves what I am saying: that for the amateur, a lot of modern day stuff is of no use at all. What's an FPGA? Clock distribution? Synchronizing logic? SPI channels? VHDL? Verilog? I don't know, and frankly I don't want to know.
Look at it this way. What proportion of people in this country understand anything about subjects beginning with electr*******, comp******, tele*****. The answer is that a very high proportion can use this stuff, but have no idea how it works. Just look how the wool was pulled over some peoples eyes with digital TV. The small proportion of people who do have an idea are immediately hamstrung by the lack of information about what's inside the little black boxes: they simply have to either guess or, if a data sheet can be found - and that's not always easy to do - they are then confined to the black box designer's design parameters. It's all right if you have all sorts of design aids readily available as a professional, but does the amateur? This is why I say that the amateur is best sticking with discrete components because generally the design rules are perhaps well known, more flexible, and one can design to ones own parameters.
I understand that yes there may well be "better" op-amps than the 741 available, that some people wouldn't be seen dead using a 555, but strangely, both are still available so someone must be using them. Don't know about TTL as I haven't even looked at them for nigh on 40 years.
I've just had a quick look at a well known mail order amateur electronics firm and note that the venerable TO3 2N3055 is still available although some of the other transistors I have used in the past are not.
I agree that good design should be able to guard against failures, but if it is not a good design, and the circuit diagram is not readily available, what then? I understand that some boards have even had the part numbers removed from the active elements, presumably to stop people attempting a repair. Expensive doorstop?
In respect of switch-mode power supplies (SMPS), as an amateur, I would not use them if possible as I much prefer the simpler linear system of transformer (which also acts as an isolator), rectifiers, smoothing capacitors followed by, if necessary, a discrete transistor regulator. I am aware that SMPS are supposedly more efficient - but when I feel the heat coming off my laptop supply, I begin to wonder - but they are more complicated, and I understand that elsewhere in the world, there have been reports of an increase in catastrophic failures of electronic equipment caused by the failure of the SMPS. Which does indeed make me wonder about the economics of them. Let's face it, it's not allright saving a few poundsworth of energy if at some point down the line the complete item has to be scrapped because it's been blown up!
I think, Andrew, that if you are a specialist designer using fancy design tools and can have specialist PCB's and IC's knocked out by the million or so, then that is one thing. But when you are an amateur without those tools, then you are stuck with either using someone else's designs, eg the Arduino kit, or reverting back to simpler basic components that don't rely on specialist boards or items and for which spares are likely to be available for a long time to come. I suspect that there are a number of people around who can indeed understand discrete component circuitry, but are completely hamstrung when it comes to specialist stuff.
An interesting aside for you Andrew. I learn't how to design switching circuits using transistors in the 1970's. In 1980 I was promoted and promptly stopped doing anything like that. In 1982 I bought myself an oscilloscope for use at home, but never actually used it that much due to other constraints, mainly time. Two years ago, I had occasion to use it and discovered it was faulty. Now this is an analogue device, yet by studying the circuit diagram, and careful noting of what did, and didn't work, I was able to get to within one transistor of the one which had failed. It turned out that the casing had cracked, maybe it had been like it from new, dunno. Anyway, surprisingly I found some identically labelled transistors seven miles from where I live. I bought 10, set to with variable voltage supply and multimeters and managed to sort out some matched pairs at dc conditions. Replaced the broken transistor and its matching partner and now the oscilloscope is working again. Obviously I can't check it's high frequency response etc, but at least it is working. Speaking to the importer later, I discovered that the modern equivalent is totally ic. What chance would I have had of repairing that? Probably none - unless they still use medium power output transistors in a balanced push-pull output stage.
Peter G. Shaw
|John Haine||01/06/2014 20:50:35|
|3425 forum posts|
Well, some of those amateurs will be young people - should we condemn them to using obsolescent components and techniques? Do you think they would be at all interested? Just how long do you think 4000 series CMOS will be around? Or maybe they should use valves?
|Neil Wyatt||01/06/2014 21:44:01|
18320 forum posts
Surely it all depends what you want to do.
No one is forcing people to use modern chips if they don't want to. Thing is, it's so easy to do parametric searches these days you can find a 'best fit' op amp or transistor in seconds on Farnell, for example.
I started off with three things: Antex iron, Dad's Heathkit scope and a high-end Eagle multimeter. A mate and I cracked the boot sequence of the Amstrad PCW with a 555 toggling the reset line and inspecting the data and address lines one by one on the scope.
Hameg 50Meg scope, quite a lot of nice analogue multimeters (rarely used as the Clarke digital is more accurate and has higher input impedance), home-made DDS and analog frequency synthesisers and my trusty 15W antex iron. Home made UV light box and these days I drill pcbs on my 6-speed bench drill (yes, really!)
Extra to that for microcontrollers I've got an STK500 and now an AVRISP which runs off USB (leaving precious serial ports free for projects). These cost peanuts compared to typical model engineering tools.
I have never actually touched an Arduino -(I did see one at Harrogate!) The simpler AVRs are ideal for learning on and have no more pins than a 555.
I have another project, rarely used, called Beeb Scope it's a box with a fast A2D that wires up to the user port of a BBC computer and just dumps data as fast as the beeb asks for it. A special free-run mode gets 256 samples at about 500Ksamples (if I recall correctly). I managed on this as my scope for about seven or eight years.
Electronics is fundamentally cheap, and ingenuity is worth more than the latest gear. What components you apply your ingenuity to, is up to you.
2904 forum posts
If you look at how technology has exploded over the last 50 years, it's difficult to imagine what will happen in the next 50. However, you can be sure it isn't about to suddenly come to a complete stop just because we (here) can't imagine what will happen next. The way it will continue is by the next generation(s) starting where we left off. Our kids grew up with computers, powerful software, smartphones, internet etc and they take them almost for granted. It's come to the point where no single individual can claim to understand every aspect of a modern product. Even the software has got to that point. Modern cars have between 50-100 microprocessors in them (you possibly didn't realise that). It's not about to stop...
I started out playing with discrete transistors like the OC71, AC128 and was messing with switching power supplies back in the late 70s. Now I develop smart motor drives, chargers etc that have dedicated SMPS controllers, microprocessors, comms ports, displays, diagnostics, data logging, thermal protection etc. Things have come a long way! SMPS and other switching power electronics products can be very reliable indeed, not least because we understand a lot more about failure mechanisms nowadays. Modern components are effectively defect-free as far as product reliability is concerned. Generally, failures are the result of poor design or inappropriate application. Might sound controversial that statement but that's the long and the short of it.
Today's engineers have to design with highly integrated devices, like it or not. There is still the need for discrete devices in almost any product but there's very little these days that you can do with a handful of discretes alone. Apart from inverted snobbery or nostalgia, there's really no benefit from sticking to decades-old technology. That's not to say I'm not nostalgic myself (I have "several" boxes full of nostalgia) but what has always really, really excited me is thinking about what we could do if we build on what we have - and going out and making it happen. In this context, that's perhaps the difference between an engineer (tackling new challenges) and a technician (sticking with the same)? No disrespect meant to technicians, BTW.
|Geoff Theasby||02/06/2014 09:12:52|
|613 forum posts|
All very interesting. I still play with small projects like regenerative radio receivers, sirens, simulated steam whistles, for a bit of fun in an evening. However, in the real world, an acquaintance of mine is working on rewiring a Land Rover with single-bus electrics run by an Arduino. I am inclined to go the Raspberry Pi route, having read the Haynes Manual about it. My radio amateur self wants to get a Funcube Pro Plus dongle, a wideband miniature computerised radio receiver!
|Neil Wyatt||02/06/2014 09:21:24|
18320 forum posts
When championing my AVR projects, i should mention my 'other' project on the go is a 'Fuzz Face' using teh original 1960s circuit and vintage germanium transistors
|Russell Eberhardt||02/06/2014 09:23:37|
2600 forum posts
TTL? New fangled stuff. Why not stick to using Mullard Combi Blocks for logic? Personally I'll stick with red spot and white spot transistors.
|Gordon W||02/06/2014 15:42:11|
|2011 forum posts|
Why wire a Landy with single -bus electrics? ( I'm assuming a proper one)It must be cheaper to use decent gauge wire, and a lot easier to fix on a cold wet night. Still recovering from trying to wire trailer socket on these new fangled things.
Please login to post a reply.
Want the latest issue of Model Engineer or Model Engineers' Workshop? Use our magazine locator links to find your nearest stockist!
You can contact us by phone, mail or email about the magazines including becoming a contributor, submitting reader's letters or making queries about articles. You can also get in touch about this website, advertising or other general issues.
Click THIS LINK for full contact details.
For subscription issues please see THIS LINK.