Poor Man's Digital Oscilloscope (now with VGA option and screen capture to SD)

Perry · 2011-03-25 19:33

I have been working on this for some time and need to share

This is a "Dual Channel Oscilloscope" with vector mode v key to switch mode)

it uses two sigma/deltas , a keyboard and tv on a protoboard

Adapted from "raman"'s dual channel ADC, the code posted below works. but there is a problem I can not seem to resolve.

main controls are:
space" ... start/stop capture
"v" ...... switch from normal to vector mode
"F7" ....... decrease sample time
"F6" ....... increase sample time
left ...... increase attenuation
right ..... decrease attenuation
up ........ move display up
down ... move display down

added shift and ctrl to individually manage Channel A or B for movement and attenuation

My problem is that in the main program at line 185 I have to restart the ADC cog, instead of just resetting the buffer address. (it works if I restart the capture cog, I have another project in mind that needs more speed)

in the main program (Dual_Channel_Oscilloscope.spin

    else
      adc.start(SampleTime, inpina, opina, nAverageA, nPeakA, inpinb, opinb, nAverageB, nPeakB)
'      adc.SetBuffer(@ScopeBuffer1)  ' This is how  it should work

in Dual_Scope_ADC.spin

PUB SetSampleBits(arg) ' use sampleTime not bits this one works
  long[@asm_cycles] := arg 

PUB SetBuffer(argh)'  this one does not work
  long[@Buffer] := argh

I have added two photos
the first is in XY mode showing the classic chaotic Chua's Circuit
and the second in standard display showing the inputs from a video source

Have fun with it, and suggestions/improvements are desired

Just added VGA option
to build you need Dual_Channel_Oscilloscope.spin and Dual_Scope_ADC.spin
then chose one of TV_Graphics.spin or VGA_graphics.spin

code online now is TV version, to use VGA just change to the OBJ declaration and basepin

April 10 2011

added config file
changed to lonesock's Easy_ADC
updated code to allow writing to SD ram

April 13 2011

I wimped out and used FloatMath and FloatString to show timing info.
also added a simply function generator that you can toggle on/off with the "f" key
saved all files needed in the archive.

It's been reported that this version won't compile with Propeller Tool( run's out of memory)
You need " Brad's Spin Tool" http://www.fnarfbargle.com/bst.html which has a compiler option of "Eliminate unused SPIN methods"

April 16 2011
Another upload. I think there is not much more that can be done with this. There is not much memory left and only one cog

I might rethink the input mechanism.

added two function generators that each use a cog

function_1 square wave direct toggling of any pin max speed is maybe 500KHz
toggle on/off with "q" key
set its period with a number and then the "p"

function_2 sine wave ......... using the duty cycle (use your sound output pin) max almost 4000Hz
toggle on/off with "f" key
set its period with a number and then the "t"

presently you have to manually stop/start a function to see the change.

Apr 16 2011
Another update

changed function generators to accept input in Hz instead of period
change trace display to voltages instead of values.

Humanoido · 2011-03-26 07:56

I'm very interested in this oscilloscope, where are the comments and instructions for connections?

Perry · 2011-03-26 08:39

Humanoido wrote: »

I'm very interested in this oscilloscope, where are the comments and instructions for connections?

right near the start of the main program

  'start ADC
  SampleTime := 127 
  InPinA     := 18   ' channel A pins
  opinA      := 17
  nAverageA  := 10
  nPeakA     := 1 
  InPinB     := 8     ' Channel B pins
  opinB      := 10
  nAverageB  := 10 
  nPeakB     := 1

Rayman · 2011-03-26 10:58

Good app Perry, I've been meaning to do something along these lines for a long time, but never got around to it.

I assume you're using the standard 2-resistor ADC circuit where there's a 150k resistor from the signal source to the InPin and then a 100k resistor from the Opin to the InPin...

Perry · 2011-03-26 11:31

I assume you're using the standard 2-resistor ADC circuit where there's a 150k resistor from the signal source to the InPin and then a 100k resistor from the Opin to the InPin...

not quite, I use two 470K resistors but I suspect as long as the two resistors are pretty close it will work.

I am fine tuning the display and functions, but still can't figure why the setbuffer function does not work dynamically.

HollyMinkowski · 2011-03-26 13:13

That's a very nice looking scope display!
Well done Perry :-)

I saw some tiny 3.5" LCD NTSC displays on ebay the other day for
22.00 USD delivered. They might go well with this to make a small portable
package.

Humanoido · 2011-03-26 14:03

Is this a sound oscilloscope? i.e. what is the frequency response?

Perry · 2011-03-26 14:52

Humanoido wrote: »

Is this a sound oscilloscope? i.e. what is the frequency response?

This goes way beyond audio, you can see the equalization pulses in the TV horizontal synchronization signal.
I have found one problem where I forgot to put # in front of a jump address

have uploaded replacements on first page

jazzed · 2011-03-26 15:11

That display looks pretty good. Congrats.

davidsaunders · 2011-03-26 15:20

Darn. You started this thread 1 day to late. A friend just bought me a BK1476A, to replace an even worse scope. Yours is a good bit better than the old scope (a home mad job [using a very old BW TV CRT], by another friend).

Toby Seckshund · 2011-03-26 15:33

Hi Perry

Are you using the QPF style of props or is the DIP40 good enough?

I have a couple of the quad sort but I always prototype with the DIP40s, There has always been comments on how the QUADs are really the only ones that hack the Sigma/Delta stuff, due to lead lenths etc.

Perry · 2011-03-26 16:07

Toby Seckshund wrote: »

Hi Perry

Are you using the QPF style of props or is the DIP40 good enough?

I have a couple of the quad sort but I always prototype with the DIP40s, There has always been comments on how the QUADs are really the only ones that hack the Sigma/Delta stuff, due to lead lenths etc.

I think that problems with sigma/delta have been much less than described. I can't see why they should not work on a DIP40, make sure you have all the bypass capacitors on every power supply.

My protoboard is quite "shop worn" in fact my "B" channel has had to get a little thread of wire to replace part of a PCB run that got burnt. Keep the leads to the feedback as short as you can, once you begin dealing with the input side of the input resistor you should have few concerns.

IF you can keep you board clean you should be able to sigma/delta circuits.

davidsaunders · 2011-03-26 16:28

Beings as this is at 2.5Ms/s, I am wondering if it would be somehow possible to use oversampling to get even higher sample rates? What would need be taken to consideration for such a circuit? Mostly thinking aloud, though also hope some one with the spare time plays with the idea a little.

Perry · 2011-03-26 17:03

davidsaunders wrote: »

Beings as this is at 2.5Ms/s, I am wondering if it would be somehow possible to use oversampling to get even higher sample rates? What would need be taken to consideration for such a circuit? Mostly thinking aloud, though also hope some one with the spare time plays with the idea a little.

The sigma/delta ADC code will not actually work at 2.5MHz as documented, with the extra code I added it will do about 1MHz.
Because you use clock counts to get the ADC values the higfer the frequency attempted the lower the bit resolution obtained.

If you had a real ADC chip instead is sigmal/delta you would get better resolution and speed, should not be very hard to adapt this program for that.

davidsaunders · 2011-03-26 17:57

K, (you should be able to get 5Ms/s at 3 bits resolution, though that is pushing the limit with op timing). I was actually thinking have multiple ADC cogs for the same signal, with the timing offset by an appropriate factor (depending on the number of ADCs). This would require some way to completely isolate an analog signal, and that could be a bit tricky.

davidsaunders · 2011-03-27 16:29

Taking a day of play here. Does any one know how to isolate a analog voltage in time, using a fairly simple circuit? This idea of a inexpensive oscilloscope has me intrigued. If anyone can get this to somehow work reliably >30Ms/s we would have a very usable digital oscilloscope, that could be built for about $100 including the display.

Duane C. Johnson · 2011-03-27 16:44

What do you mean by "isolate a analog voltage in time"?

Duane

Phil Pilgrim (PhiPi) · 2011-03-27 16:46

I'm not sure what you mean by isolating a voltage "in time". Is this meant to be different from galvanic isolation? Can you elaborate?

-Phil

davidsaunders · 2011-03-27 18:11

Phil Pilgrim wrote:

Can you elaborate?

To sample the voltage of a analog signal at one point in time and some how hold that level (separate from the source witch should be aloud to change).

Phil Pilgrim (PhiPi) · 2011-03-27 18:42

You're referring to a sample-and-hold circuit. A cap, an analog switch, and an op-amp buffer are all you need.

-Phil

davidsaunders · 2011-03-27 19:35

Thank you Phil. I should probably play with the analog side a bit more. Now off to do some math and see if my theory holds water. I figure that it should be possible to make this Poor Mans Digital Oscilloscope work at 6.75Ms/s.

Duane C. Johnson · 2011-03-27 20:45

You guys have been describing "Direct Sampling Oscilloscopes".
Their basic limitation is in how fast one can write the ADC values into memory.
The other limitation is how fast the ADC can operate.
Sure, you can increase the speed by using 4 8-bit ADCs and storing the 4 samples in a 32-bit word.
But that is about as fast as you could go.

However, there is another configuration that can operate at much faster speeds.
These are the "Time Domain Sampling Oscilloscopes" or the "Frequency Domain Sampling Oscilloscopes".
The heart of these devices is a very fast sampler, or sample and hold.
There's always a trade off, in this case you take only 1 sample point each time the the waveform comes by.
So one must have a repeating waveforms.

The Time Domain type takes it's samples after delays from the trigger point.
The Frequency Domain type uses a phase locked oscillator synchronized to the trigger.
The Propeller is a natural with this second type because it has the Numerically Controlled Oscillators.

Once the sample is made an ADC can read it at a reasonably slow rate. Read this to mean
it can do it with high resolution.

Here is a "Idea For Design" from "Electronic Design" September 18th, 2000.
http://www.redrok.com/Circuits_1GHz-samplig-Oscilloscope-Front-End.pdf

Duane

MacTuxLin · 2011-03-27 21:42

I'm quite poor so I do really needed this but, is there a way to convert the NTSC signal to PAL?

Perry · 2011-03-28 08:41

MacTuxLin wrote: »

I'm quite poor so I do really needed this but, is there a way to convert the NTSC signal to PAL?

It should be doable, it uses the Parallax standard TV driver.

Just before you start the TV driver try this:

  tv_mode |= 1

  tv.start(@tv_status)

''  _______
''  tv_mode
''
''    bit 4 selects between 16x16 and 16x32 pixel tiles:
''      0: 16x16 pixel tiles (tileheight = 16)
''      1: 16x32 pixel tiles (tileheight = 32)
''
''    bit 3 controls chroma mixing into broadcast:
''      0: mix chroma into broadcast (color)
''      1: strip chroma from broadcast (black/white)
''
''    bit 2 controls chroma mixing into baseband:
''      0: mix chroma into baseband (composite color)
''      1: strip chroma from baseband (black/white or s-video)
''
''    bit 1 controls interlace:
''      0: progressive scan (243 display lines for NTSC, 286 for PAL)
''           less flicker, good for motion
''      1: interlaced scan (486 display lines for NTSC, 572 for PAL)
''           doubles the vertical display lines, good for text
''
''   bit 0 selects NTSC or PAL format
''      0: NTSC
''           3016 horizontal display ticks
''           243 or 486 (interlaced) vertical display lines
''           CLKFREQ must be at least 14_318_180 (4 * 3_579_545 Hz)*
''      1: PAL
''           3692 horizontal display ticks
''           286 or 572 (interlaced) vertical display lines
''           CLKFREQ must be at least 17_734_472 (4 * 4_433_618 Hz)*
''
''      * driver will disable itself while CLKFREQ is below requirement
''

Rayman · 2011-03-28 08:47

I think I've heard that the PAL output doesn't really work... But, most PAL TVs take NTSC as input...

Anyway, if you can't get that working, maybe another way is to use the USB serial connection to send the scope data from the Prop to the PC.
Just need to come up with some program or terminal window to display the data...

MacTuxLin · 2011-03-28 08:56

Thanks Perry. No harm giving it a try.

Hi Rayman. Thanks. I don't know about that but if my memory serves me right, I had distorting images when I connect the other way around. I'll definately be working on the at the sideline along with my urgent projects.

Cheers!

Perry · 2011-03-28 17:05

Rayman wrote: »

I think I've heard that the PAL output doesn't really work... But, most PAL TVs take NTSC as input...

Anyway, if you can't get that working, maybe another way is to use the USB serial connection to send the scope data from the Prop to the PC.
Just need to come up with some program or terminal window to display the data...

You got me thinkin' about your PrintScreenVGA, with slight modification and simple_serial you can do a hex dump of the screen with only 180 longs.

It should be easy modify this 'scope to run on VGA, I just don't have VGA connector on my proto board , so I can't test the changes !

MacTuxLin · 2011-03-28 17:47

Perry wrote: »

It should be easy modify this 'scope to run on VGA, I just don't have VGA connector on my proto board , so I can't test the changes !

:O that would be great!! I can test for you.

Perry · 2011-03-29 08:58

MacTuxLin wrote: »

:O that would be great!! I can test for you.

I just upgrade the front page of this thread for VGA, could some one try the VGA option.

MacTuxLin · 2011-03-29 09:11

Perry wrote: »

I just upgrade the front page of this thread for VGA, could some one try the VGA option.

Cool! I'll test this tomorrow morning when I'm back office. Don't have the board with VGA at home

MacTuxLin · 2011-03-29 18:10

Perry wrote: »

I just upgrade the front page of this thread for VGA, could some one try the VGA option.

Hello Perry,
I've tested but displayed nothing. The changes I've made were

'  gr    : "TV_graphics"
   gr    : "VGA_graphics"    'Use for VGA screen.

and

  'basepin  = 12                'Setting for Demo Board TV.
 'basepin  = 24                'Setting for Hydra TV.
 basepin  = 16                'Setting for Demo Board & Hydra VGA

Are there anything I need to change as well? On the hardware side, I'm using prop proto-board with mounted VGA/PS2 KB/Mouse connector and base 16 was tested OK.

Thanks a lot.

Poor Man's Digital Oscilloscope (now with VGA option and screen capture to SD)

Comments