Design + Development : Rules and tools.

This posting will only be of any interest to those interested in engineering practice etc. Please, no howls from hobbyists and wanna-be's who are happy messing about with "tab" type circuits, but have no real interest in how things work or in good quality design.. Parts of this "article" are NOT specifically about theremins, it is the type of article which seems to have got up at least one persons nose recently - so RS, if you are reading this, stop reading it now! ;-)

General waffle:

All the GOOD technical stuff you play with came about through a process of INTELLEGENT DESIGN - Design based on knowledge of the art, design founded on correct understanding of the principles of physics etc, design where mathematics is used to calculate and verify tolerances and performance etc.

All intellegently designed 'products' are 'created' at least twice - First they are created in the minds of those undertaking the creation (this is the absolute minimum "first" creation - for only the simplest ideas can one move from this directly to the "second" creation where one actually makes the 'product'), they are then created in a documentation form (diagrams, schematics etc), they may be created as a 'model' (reduced scale mechanical model testing functions or machanical / electronic simulations) before they move to actual physical prototyping for further testing / evaluation / simplification / improvement, and then to creation of the product.

The other method of creation is evolution - we cannot use this method, it takes too long and is too expensive (and the results are far from elegant!) - The fundamental difference between design and evolution is that, with design, the end product is imagined first - with evolution there is no knowing what the end product will be!

One dilute form of "evolution" is practiced by some electronics hobbyists - this is to take a designed item, and modify it without really understanding the design.. Happy accidents can rarely occur, but mostly this approach ends up with something inferior to the original design.

As TW is embarking on development of a theremin, I saw this as an excuse to make another verbose posting!  ;-) LOL

Electronics Development. "Rules":

1. Define the objectives and "base-line" Specification.

This is a bit like a "business plan" - Companies (like mine, LOL ;-) who do not have a clear business plan written up, are statistically much more likely to fail. With electronics, lack of a clear minimum specification and clear objectives and targets (including factors such as costs etc) almost always fail.

2. Central Managment by a competent individual / group:

A load of engineers with different specialisms all working on a project, will stumble about in confusion unless their activities are managed. An individual (or small group) with a broad overview / understanding of the various disciplines needs to be in place to act as an interfacing point.. this individual / group does not need to be specialist in all (or any) of the disciplines, but does need good general understanding and the ability to weigh up conflicting ideas and choose the best path to achieve the objectives.

3. Split tasks and assign these to competent individuals:

5. Tight revision control:

It is frustrating to be working on your section of a design, only to find that the section it interfaces to has been changed and you have not been consulted or notified.

All documents should have a central repository where changes / revisions are automatically logged (this is available at E-14) and one is notified automatically when an update is made, and previous revisions are available for viewing, but the latest is always at the top of the pile.

Tools:

Schematic editor:

There are some essential tools required - the main one is a schematic editor. In a collaborative project, loads of time is saved by being able to exchange schematics in a common form.. But there are problems - everyone has their favorites. I have worked in companies where two teams used two different editors because each refused to change to the other - this meant that every schematic was hand re-drafted, with the resulting errors which arise through this process.

I think that the free editors available should be looked at - I think that Eagle and LT-Spice both have usable free editors - Eagle is size restricted, but possibly just enough for a modular theremin, LT is unlimited, but doesnt do PCB layout - Eagle doesnt do simulation..

There is also DesignSpark which I have not used, but looks too good to be true! - Unlimited, Free, Schematic and PCB design, and interfaces to LT Spice ..

PCB Editor:

Only required by those designing PCB's - needs to take input from the chosen schematic editor. Eagle and DesignSpark look like the best free options

Simulation:

This is a complex tool.. It alows one to "build" reasonably accurate models of circuits, and 'run' these to emulate what one could expect to see from an actual circuit.

Complex tools take a lot of learning if one is to use them correctly, and this tool, in the hands of someone who does not have reasonable understanding of electronics, can give misleading results - however - even an electronics novice can, if they start with simple circuits, learn a great deal from this tool - it enables safe experimentation without the need for expensive test equipment.

In the hands of a competent electronics engineer, this tool can aid in testing circuits under a range of conditions - component values can be 'swept', temperatures can be 'swept' and conditions observed which enable optimum component parameters to be chosen.

I am sure that Lev will have used the design tools at his disposal (log tables, slide rule etc) - Simulation is just a tool.

There is only one worthwhile free simulator IMO, LT-Spice.

Fred.

100% agreement here.

Define the objectives and "base-line" Specification.

So far, the only objectives I see that we mostly agree on is hetrodyne based and modular. I'd also like to suggest the boards be single sided for cost, service ease, and home builder simplicity. I guess I should add that to the design goals thread.

Schematic editor:

PCB Editor:

I use Eagle now. It has board size limits, but I think it can do large schematics. I've already got some of the boards designed for a theremin I was working on before we started this project.

Simulation:

I downloaded LT-spice. It's very different from Simetrix, but I'll figure it out eventually. I was, however, designing and building electronics the old fashioned way when computers ran off steam. It's a cool tool to play with, but the truth comes out when the solder hits the iron. Also, I have yet to find a realistic 12AX7 spice model. That seriously limits it's usefulness for vacuum heads like me.

Tight revision control:

For the last eight years, I've worked for a small electronics company building prototypes and refining them into products that we sell. Documentation, revision control, and communication between departments is very VERY important. E14 is a great place to file them.

Central Managment by a competent individual / group:

Lets draft Jason for that one.

This is his web site. His reputation and that of this site could be impacted for good or ill by what we produce.

We can advise, recommend, whine, and so forth, but I believe the final call should be his.

Split tasks and assign these to competent individuals:

Let's hear from some volunteers.

I'd be thrilled to lay out some boards. I can do bread board prototypes, but my time is limited so some help there would be nice.

Design Spark

I have downloaded this and been playing with it for the last few hours - As a free package, it seems unbelievable.. I imagine that some companies selling  schematic / PCB design packages will be extremely upset by this product.

Ok, I havent got deep into it, so take this as a preliminary assessment, and I may change my opinion - but so far:

Schematic editor:

Schematic design is quite straight forward - component selection is a bit of a pain, but once one has placed a few basics (resistors, capacitors etc) one can copy / paste these and alter their values.. not much hassle, and as easy as many expensive packages.

A nice feature is that one can add value parameters to components - as well as the usual (value, footprint etc) one can embed things like order codes etc - all useful stuff for collaboration.

Another nice feature is that, hovering over a net (wire) the net name is shown - this is extremely useful as one can refer to net names when collaborating, rather than saying "the junction of R7 and base of Q6" for example.

This latest version (3) also has a great feature that one can "group" sections of the circuit - this "grouping" is passed to the layout editor, and keeps grouped components togeter when auto-placing them.

The Layout editor, after some brief playing with it, looks extremely good - There is an auto-place function - I have this feature on other packages, but this one seems far better than my others. One would place things like connectors and potentiometers which need fixed positions, then invoke the auto-placer.. It places the remaining components trying to keep grouped sections together. When this is done, one can (and for serious designs, should) hand optimise the placement.

Then there is the auto-router.. Well, these can (and usually do) make a mess of designs where signal routing is important (as in, one needs to keep tracks intellegently positioned to minimise coupling etc - important for the stuff in many sections of a theremin) - But this router did a damn good job of wiring my test circuit.

Have not tested back annotation or any of the important stuff where a schematic change needs to update a PCB.. I also have not tried the simulation interface to LT Spice.. I am sure that the cross-probing available in the LT Spice schematic wont be available from the Design Spark schematic - But probing after simulation should still be possible and easy because net names are easy to see in Design Spark - so entering these names into the waveform viewer at LT Spice will effectively enable cross probing.

I will certainly be playing extensively with Design Spark over the next week or so - I have had difficulty many times because I needed to collaborate with other engineers and clients, but each used different design software, often locked to specific PC's, and costing thousands of pounds so obtaining a license for all collaborators was impossible. This package looks like a comprehensive solution to that problem.

Oh, one other feature - untried - but it claims that you can import Eagle files and libraries..

Fred.

Spark installed.

Looks promising....

Can I howl? ;)

Any time Amey! With a voice like yours, any sound you could make would be a pleasure! ;-)

Dang skippy!

Let's send the first prototype to Amey.

i am game !!! YEAH!!!

"Let's send the first prototype to Amey." - w0ttm

No - never send the first prototype to anyone! - that is a big, nay - huge mistake! In my desperation to complete my theremins for HO-2010, I was forced to put my prototype into one of the 16 units .. It had some modifications I had not implemented in the 'production' units.. And somehow my documentation went wrong - never recovered those mods.

But the first reliable production unit, Yes! I agree wholeheartedly!

Fred.