Analog Calculator
- Thread starter cala
- Start date
OK......why? I mean, what is your goal? And do you mean calculator in the modern sense, with a keypad and display? Analog computers were around before all of those bit-bouncers, and internet searches can call up circuits for each of the functions you list. So....why?
ak
ak
I used a variety of them for years in my job. They worked very well implementing the four basic operations and the final result was shown by two, three (or more) instruments with central zero. Sorry, no differential equations solved. (That I did by hand in my spare time) 
Their generic name (from a well establish brand) was "loadicator". Their use was to calculate displacement, trim, bending moment and sheer forces of your vessel.
Manual input (coarse and fine) of weight for EACH cargo space or tank (and there were/are many in big bulk carrier vessels). Kind of 7 holds plus around 20 tanks, some of them meaning 2 knobs.
Chief Officer was the only one to use it and was usually kept on during loading to follow the operation and check (almost on line) the process.
They used lot of power. Had the chance to open the back plate of one of the biggest I ever used and see all the op amps with a crazy amount of presets.
The operator just messed with front panel commands.
Now, a dedicated PC that nobody should use for any other purpose, is compulsory to have, certified by the classification society.
I am sure is not hard to get a simple one running. Avoid being ambitious when starting. Make sure you get good advice on what op amps to use, because offset means a lot here so you could go away without compensating it, at least in the first steps.
Elektor run an article on a basic implementation some time (years) ago.
Now that I think of it, op amps were initially built for that...!
Buena suerte.
Their generic name (from a well establish brand) was "loadicator". Their use was to calculate displacement, trim, bending moment and sheer forces of your vessel.
Manual input (coarse and fine) of weight for EACH cargo space or tank (and there were/are many in big bulk carrier vessels). Kind of 7 holds plus around 20 tanks, some of them meaning 2 knobs.
Chief Officer was the only one to use it and was usually kept on during loading to follow the operation and check (almost on line) the process.
They used lot of power. Had the chance to open the back plate of one of the biggest I ever used and see all the op amps with a crazy amount of presets.
The operator just messed with front panel commands.
Now, a dedicated PC that nobody should use for any other purpose, is compulsory to have, certified by the classification society.
I am sure is not hard to get a simple one running. Avoid being ambitious when starting. Make sure you get good advice on what op amps to use, because offset means a lot here so you could go away without compensating it, at least in the first steps.
Elektor run an article on a basic implementation some time (years) ago.
Now that I think of it, op amps were initially built for that...!
Buena suerte.
Ohh yes - the name "operational" amplifier is an indication of the purpose to perform various mathematical operations.
Random thoughts...
In WWII, all services used mechanical analog computers. The bombers had the Norden bombsight, artiliary had ballistics calculators that factored in different elevations, wind speed and direction, etc., and the navy has something similar for torpedo aiming.
For mechanical ACs, one of the easiest things was multiplication, and the hardest thing was integration. But for electronic analog circuits, integration is one of the easiest functions, and accurate multiplication is a bear.
For accuracy, both vacuum tube and solid state analog computers used a chopper-stabilized inverting amplifier as the basic building block. Differential amps with both inverting and non-inverting came later. I rebuillt an analog computer replacing each chopper and a wad of germanium transistors with the ICL7650. Instant 10x performance improvement. This is the deep end of the pool, and you certainly don't ahve to start here; basic opamps will do everything you want to investigate the concepts. But if you want analog multipliers to give you anything useful over a wide range, you have to sweat the error budget from the beginning.
http://en.wikipedia.org/wiki/Analog_computer
http://en.wikipedia.org/wiki/Chopper_(electronics)
http://electronicdesign.com/analog/chopper-stabilized-op-amps
ak
In WWII, all services used mechanical analog computers. The bombers had the Norden bombsight, artiliary had ballistics calculators that factored in different elevations, wind speed and direction, etc., and the navy has something similar for torpedo aiming.
For mechanical ACs, one of the easiest things was multiplication, and the hardest thing was integration. But for electronic analog circuits, integration is one of the easiest functions, and accurate multiplication is a bear.
For accuracy, both vacuum tube and solid state analog computers used a chopper-stabilized inverting amplifier as the basic building block. Differential amps with both inverting and non-inverting came later. I rebuillt an analog computer replacing each chopper and a wad of germanium transistors with the ICL7650. Instant 10x performance improvement. This is the deep end of the pool, and you certainly don't ahve to start here; basic opamps will do everything you want to investigate the concepts. But if you want analog multipliers to give you anything useful over a wide range, you have to sweat the error budget from the beginning.
http://en.wikipedia.org/wiki/Analog_computer
http://en.wikipedia.org/wiki/Chopper_(electronics)
http://electronicdesign.com/analog/chopper-stabilized-op-amps
ak
Hola AK,
I recall seen in use a mechanical calculator? (not sure of the actual name) used to aim the gunnery in the former USS Phoenix (sunk as General Belgrano in Malvinas).
It was fascinating to see how they implemented integration (if IIRC, a mechanical "cursor" going up / down on a cone. I am not sure what was constant if the rotation of the cone or the speed of the "cursor" going up / down. Seen that once, more than 40 years ago. Not sure why I remember that.
Am I right?
I recall seen in use a mechanical calculator? (not sure of the actual name) used to aim the gunnery in the former USS Phoenix (sunk as General Belgrano in Malvinas).
It was fascinating to see how they implemented integration (if IIRC, a mechanical "cursor" going up / down on a cone. I am not sure what was constant if the rotation of the cone or the speed of the "cursor" going up / down. Seen that once, more than 40 years ago. Not sure why I remember that.
Am I right?
DickCappels
- Joined Aug 21, 2008
- 10,180
You can add and subtract using an opamp as a summing amplifier. If you don't already have Opamps For Everyone, you might want to download it.
www.ti.com/lit/an/slod006b/slod006b.pdf
A shorter book on the topic is the Burr-Brown HANDBOOK OF OPERATIONAL AMPLIFIER APPLICATIONS. Less material, but more concentrated on what you want to do.
www.ti.com/lit/an/sboa092a/sboa092a.pdf
There are a number of way to multiply. Analog multipliers based on Gilbert Cells and such are nice for AC signals, but they tend to drift, even when using very good monolithic multipliers and expensive trimming components. You may want to look at this device from Analog Devices
http://www.analog.com/en/special-li...ltipliersdividers/ad633/products/product.html
If you can get by with a two quadrant multiplier, (one input is always positive) then a pulse height-width multiplier would give a more stable output. A pulse height-width modulator is basically a circuit that generates a pulse whose height is proportional to one input (x) and whose width is proportional to the other input (y) and that pulse is passed through a low-pass filter to get the product (z). This is a minor variation on pulse width modulation.
http://en.wikipedia.org/wiki/Pulse-width_modulation
To divide, take either multiplication technique and put it into the feedback path of a summing amplifier.
You can also multiply and divide by adding and subtracting respectively the logs on input voltages then taking the antilog of of the sum. See section three of this publication:
www.ti.com/ww/en/bobpease/assets/AN-31.pdf
And of course, there is a lot more on the web.
www.ti.com/lit/an/slod006b/slod006b.pdf
A shorter book on the topic is the Burr-Brown HANDBOOK OF OPERATIONAL AMPLIFIER APPLICATIONS. Less material, but more concentrated on what you want to do.
www.ti.com/lit/an/sboa092a/sboa092a.pdf
There are a number of way to multiply. Analog multipliers based on Gilbert Cells and such are nice for AC signals, but they tend to drift, even when using very good monolithic multipliers and expensive trimming components. You may want to look at this device from Analog Devices
http://www.analog.com/en/special-li...ltipliersdividers/ad633/products/product.html
If you can get by with a two quadrant multiplier, (one input is always positive) then a pulse height-width multiplier would give a more stable output. A pulse height-width modulator is basically a circuit that generates a pulse whose height is proportional to one input (x) and whose width is proportional to the other input (y) and that pulse is passed through a low-pass filter to get the product (z). This is a minor variation on pulse width modulation.
http://en.wikipedia.org/wiki/Pulse-width_modulation
To divide, take either multiplication technique and put it into the feedback path of a summing amplifier.
You can also multiply and divide by adding and subtracting respectively the logs on input voltages then taking the antilog of of the sum. See section three of this publication:
www.ti.com/ww/en/bobpease/assets/AN-31.pdf
And of course, there is a lot more on the web.
Before designing an electronic circuit it is advisable to do some corresponding simulations in advance.
I think, also in this case it could be helpful.
And for an analog calculator it makes sense to use a simulation package based on a block diagram approach (rather than simulating circuit diagrams on a parts level).
For this purpose and with good success I have used "VISSIM" (from Visual solutions.com).
It is very easy to call up all needed blocks (amplifiers, summers, integrators,...) and to combine them for solving many tasks (implicit equations, differential eqations,..) in real time.
