EPE Theremin Design

Ok - I have just run an ISIS simulation on this..

The oscillators natural simulation frequencies are quite high - over 1Mhz.. In reality however, these frequencies will be lower (never trust simulation for determining frequency, unless you have added all the capacitances on the board etc).

Amplitude from these oscillators is high (due to the resonant tank, higher than the supply voltage - this is usual) - about 20V P-P

Plugging these values into a simulation of the mixer (oscillator 1 = 800k Voltage = 7VRMS, oscillator 2 = 800.5k Voltage = 7VRMS) I get a modulated (sum+difference) waveform peak voltage = 3V, lowest voltage = 2.45V. (on TR3 Collector)

The audio output is a clean 500Hz audio waveform with nice distortion - sounds really good for such a simple circuit!

The simulation demands are quite light.. the oscillators need a kick-start by setting the base of their transistors to 5V ( a little above mid supply) when the simulation is started.. this gets them going quickly.

You should be able to copy the simulation schematics into LT quite easily.. I did not use LT because I find ISIS much better when I actually want to hear what a design sounds like.

Fred

I have just put the whole design through simulation - only takes a couple of minutes to run.. Not easy to hear it, because one does not have control over output frequencies directly.. makes it difficult to select looping points.

One thing is noticable - these oscillators are EXTREMELY sensitive to capacitance change.. the simulation produced about 1kHz difference for a change of 1pF.. So, when running your simulations, stick to using generators! In real life, you really will need a lot of patience and/or a good frequency counter to set this Theremin up to work!

Fred.

[b]***** Files can be found here ---> [/b]
a simple theremin analysed: EPE 1995 : simulation.zip (http://www.element-14.com/community/docs/DOC-23483/l/a-simple-theremin-analysed-epe-1995-simulationzip)
This .zip file contains two .pdf's and one .wav
[b]<----[/b]

Thank you Fred, I've seen the zip file and the circuit you've posted and it's really nice.

I tryed to replicate it up to the C14-R14 stage, before putting the diodes into. I used the same values as you did in your schematic, but I can obtain only a little improvement, since after the C14-R14 stage I see a signal oscillating between 175V and -175V. Thus it is still too low to pass through a diode. The frequency of the TR3 collector signal, before the C14-R14 stage, is about 3,9 KHz (I've calculated it putting the cursor on the simulated waveform, as you've said, findind che waveform period and then just calculating f=1/T).

I cannot find an antenna model in ltspice, so I just put the 33p capacitor, as you did, in parallel to C10. Then I cut C5 and C1 as well as R8 and R4, following your schema.
What's wrong?? :-S

Oh, I'm sorry, obviously the signal I've mentioned after C14-R14 oscillates between 175mV and -175mV (lol) :-)

I, for one, DO enjoy Fred's long posts! I always learn something from him.

The engineer I replaced here at work had a saying "hardware doesn't lie" ... he had little faith in simulation! I just have little patience and not enough time to do accurate simulations for my "hobby" stuff.

[i]"I cannot find an antenna model in ltspice, so I just put the 33p capacitor, as you did, in parallel to C10. Then I cut C5 and C1 as well as R8 and R4, following your schema.
What's wrong?? :-S" [/i]

You will not find an antenna model - there isnt really such a thing.. What is an antenna? In the case of a Theremin it is (for every practical purpose) only a capacitor... For real 'RF' there are a whole set of 'antenna' parameters you could specify - but there are special simulators costing many thousands $ for these sort of apps..

Don is right! What have you got.. 3 transistors, a couple of coils, diodes. Rs + Cs... What are you doing? You are building a virtual Theremin, debugging this - even when you get it working perfectly, you dont have a Theremin! ... Build it in hardware, debug it in hardware, and you will have a working Theremin.

I love simulations, and I think they are also a good way to experiment and get into electronics.. but I think that a beginner needs to play with both simulations and reality - flipping from one to the other - to get the best educationally from it.

Please feel free to email me your LT files (email is in my picture <---) and I will quickly look them over for you.. [b]But please check the obvious first! be sure all component values are correct (for example 10n not 10u !) [/b]- If it turns out to be a careless mistake, I will insist you make a $10 donation to www.streetkidsrescue.org! ;-)

But my main advice now ? Pick up your soldering iron!
;-)
Fred.

[b]-- In reply to Don -- ""hardware doesn't lie" [/b]

Hi Don,
Alas, this is not always true! My use of simulations is mainly to overcome the problems related to putting a project into production.. Knocking up a "one off" rarely justifies simulation - I would never have run a simulation for something like the EPE 1995 design - one can more quickly mess about with values on a board and get a simple circuit to work.

It is the ability to run sweeps for component tolerances, to verify that nothing in a design is 'borderline' so one can safely go to production which makes simulation, in my opinion, vital.

Hardware can (and often does) lie .. something simple like the Hfe of the transistor you put in your prototype can make a circuit work - but this transistor may have an Hfe of say 300, when the devices have a range of 150 to 400.. A careful (good) designer is unlikely to make such a gross error - but it is easy to miss the obvious in a complex design.. When a lot of money (and reputation) rides on things working without problems, simulation is a good investment.

The other area I use simulation for is to quickly test new (often whacky) ideas.. And here it is a wonderful vehicle for exploration - one can create 'unreal' components and engage with a hypothetical world and build hypothetical devices.. Many of my best ideas have come by doing this - throw away the constraints - do something amazing - then find a way to achieve the same thing using 'real' components.

The biggest problem with simulation is just how fussy it is - it takes years to master SPICE setting so as to get the right balance between convergence and accuracy for a given evaluation - it becomes instinctive (Most of the time I can not explain why I tweek parameters I do, in order to get simulations to run).. And one really needs good knowledge (and expierience) of electronics so that one can instantly see when something doesnt look right, and is likely to be a lie.

Beginners get best results by learning basic electronics from simulations - A Theremin is well outside of beginner territory for simulations, IMHO!

Fred.

[b] -- EDIT -->[/b]
[i]"after C14-R14 oscillates between 175mV and -175mV"[/i]

What do you get on the collector? is it the same sort of AMPLITUDE as on my simulation? - leave C14 and R14 out and just concentrate on getting a good amplitude on the collector..

@Fred ...

I said it was his saying, not mine. I have mixed feelings. Certainly for complex designs and verifying the effects of component tolerance and drift simulation is well warranted. I have used analog simulators various times have had mixed success. We have PSPICE at work and I have yet to get it up and running, since most of my work is pure digital here. At home, I have an older version Multisim. It is easy to use, but seems a bit "wounded" here and there.

The theremin model I built in Multisim seemed to do what it should ... the hardware did not! I was able to build something that works ... just not well enough to my liking. I have a "higher power" to answer to ... my violinist wife who is very picky about how her instruments play. :)

[i]"The theremin model I built in Multisim seemed to do what it should ... the hardware did not!"[/i]
Oh, it is so annoying when that happens! I must admit that I dont get on with Multisim, PSpice or EWB .. I have had that sort of thing happen a lot with any of those simulators.. For 'exploration' my favorite is Spice-Age (http://www.spiceage.com/Spiceage/prices.htm) - This is the first simulation package I got into (cost hundreds £ years ago, now almost free) but it does not have anything close to a useful library - one needs to create almost every model oneself.. Creating models 9or linking SPICE files) is too time consuming to make this package really useful now.. but if one wants to create the strange, this package is awesome!

Then there is LT-Spice (http://www.linear.com/) which is free, and works really well - has a lot of features not found in any other simulation package, and is as good as any other simulator I have played with.

Then my main package is Proteus (http://www.labcenter.com/index.cfm) - This is a complete package with brilliant schematic editor, Fantastic SPICE simulator, VSM (an interactive 'workbench' type simulator with support for microcontroller simulation), A fantastic PCB design program Ares... This is my 'workhorse' and I can move designs from simulation through PCB layout without needing to leave it... But, to get the whole package at Pro level is expensive (a staggering £1200 for the version I have, and £1800 if one adds top-end auto-router).

[i]" I was able to build something that works ... just not well enough to my liking. I have a "higher power" to answer to ... my violinist wife who is very picky about how her instruments play. :)"[/i]

Alas, I have the opposite problem - My wife would be happy if my Theremins played only "the" 4 chords required for MOR pop.. Anything outside this genre is, to her, amusical.

My life certainly lacks peace, and the Theremin has not helped!
;-()

Fred.

This is a quick update..

Daniele emailed the LTspice files to me, and there is some simulator related problem..

I have suggested that the LTspice Yahoo users group may be best to give a solution - but this does highlight the problems someone starting electronics can have using simulation.. You can do something right, and the simulator makes you think you are wrong!

So, some egg on my face.. Don's advice was good advice!

Fred.

[b]-- Edit --> [/b]
Re-entered schematic, changed the sim card, and it all works fine now. Just a bit of fine tuning the simulator was required, I think... But one needs to know roughly wat you expect to see, so you can adjust the simulator to see it! LOL!

Yes, that is one of the "devil's in the details" aspects about a simulator. If you have never used one ... the very situation for a beginner in electronics ... you have no idea what settings you need, what you want to optimize, etc. So you make the bold presumption that the simulator vendor has taken this into account and "pre-set" parameters that should work for simple circuits. That is not always the case.

Don