DIY Chronograph, am I doing it right?

[rp181]

1) it is built into 1in PVC pipie, so sunlight wont do it good. I will be chronoing inside, so im fine with IR LED's.

2)Vapor cloud wont be a problem, this was built mainly to messure velocities of my high power multistage coilgun (1KJ ea stage).

3) i did not want ot use the peak and negative spikes to calculate speed, mearly to verify speed.

4) my sound card works fine with slow things. I can poke my finger in, leave it for 5 sec and pull it out and still see it if i want too.

I wont bother trying to boost the power unless i see a problem.

heres a picture:



taken with webcam, so crappy, you can see the parts though.

The black blocks are what hold the LED's (there actually lego bricks, led's fit tightly in them, allows me to replace if need be). The spacing is so the LED end is right outside of the PVC outer wall, so it wont get in the way of the projectile.

the yellow and thicker black wire is LED power, the other one is the mic jack wire.

On a off opic note:
The other thing in that picture is a flyback, i assume you know what that is. Do you think using a tone from the computer at 20Kilohertz, square wave, and a mosfet i can drive the flyback? it turns out the 555's i ordered are SMD =/

[jimmy101]

I would'a built it into a piece of modified pipe, really no need for a completely enclosed detector, but whatever works for you ...

On a off opic note:
The other thing in that picture is a flyback, i assume you know what that is. Do you think using a tone from the computer at 20Kilohertz, square wave, and a mosfet i can drive the flyback? it turns out the 555's i ordered are SMD =/

That's an interesting idea, it would be easy to fiddle with the frequency and wave form (square wave, sawtooth etc.). IIRC flybacks are usually designed for something like 38KHz. That would be out of the range of the soundcard but the flyback might still work at 20 KHz.

I don't know a lot about MOSFETs but don't they expect a positive voltage on the gate to trigger? The sound card will be putting out an AC signal that is probably symmetrical about the PCs chassis ground. So you might have to half- or full-wave rectify the soundcard's output.

If you full-wave rectify a 20KHz sine wave you get a distorted 40KHz wave which might work.

[rp181]

MOSFET's change resistance based on the gate voltage (i think) its whats used for the 555 flyback driver. I think it puts out DC, i get a clean 1.5v on my DVM, it fluctates if its AC and im mesuring on DC.

[jimmy101]

rp181:
Ya, the 555 puts out DC so that must be what the MOSFET is trigging on.

You can get 555 and 556 (two 555's on a single chip) at RadioShack to replace the one that died.

[rp181]

I got 10 555 timers from Texas Instruments. I got NA555 timers, whats the difference to NE555 timers?

What fps is expected from a rubberband? I tried it with audacity and got 4 spikes (im assuming its because its a loop) and the 2 farther ones gave me .000317 (i got >3000fps from that number). What did i do wrong?

trying rubber band again now.

EDIT: I tried again, but im unable to get it through (fat rubberband). Which number do i want? I highlight it and at the bottom it says:

Selection: *number here* - *number here* (*number here* min:sec)

i was using the 1st number, but now im looking at it and thinking its the 2nd number

[SEAKING9006]

I would expect between 20 and 70 fps for a rubber band, but that's a vague guess.

[TurboSuper]

I got 10 555 timers from Texas Instruments. I got NA555 timers, whats the difference to NE555 timers?

I just compared the datasheets, and it seems like the NA555 has a higher operating temperature range.

[SEAKING9006]

Sweet! I've always wanted to time how long it takes for muffins to bake, but my old NE555's always melted!

[jimmy101]

EDIT: I tried again, but im unable to get it through (fat rubberband). Which number do i want? I highlight it and at the bottom it says:

Selection: *number here* - *number here* (*number here* min:sec)

i was using the 1st number, but now im looking at it and thinking its the 2nd number

The first number is the time of the left side of the selected part of the recording (from a "click-n-drag" selection). The second number is the right side of the selected part. You'll notice that the first and second numbers match the numbers on the time axis at the top of the window.

The number in parenthesis is the time width of the selection, that's the number you want.

So, click and drag from the first peak to the second peak. If you miss one of the peaks by a bit you can drag the edge of the highlight to get it where you want it. The time between the peaks is the number in parenthesis.

I prefer to change the default time display using [View Menu][Set Selection Format ...][sec (snap to samples)], I hate the minutes:seconds format, I prefer just seconds and I want the selection to snap to the actual data points.

If the peaks are far apart, and you want maximum resolution, you can zoom in on one of the peaks then single click it. The time will be displayed at the bottom of the window, write the number down. Now pan to the second peak and single click on it and record the displayed time. Subtract the two numbers (duh).

Here is the datasheet for the Na555 and Ne555;
http://focus.ti.com/lit/ds/symlink/se555m.pdf
The only difference I see is the same as what Turbo sees, different temperature limits.

[rp181]

Well i stupidly accedentily connected a 9v to the infrared LED's (i thought they were the 555 circuits wires ), so they burned. I made another setup, its open, 6in long. Its also easy to change the width for different size projectiles.Heres a pic (again with the webcam)




I tested it with a toy spring loaded gun, i get consistance results at 15FPS. ile try it with my pneumatic tomorrow.

BTW, i saw a how too on the method of sound on spudfiles, ide be willing to make one for optical if theres enough interest.

[jimmy101]

rp181, that looks pretty good. You should mount the battery, resistors etc. on the Legos so the only wire coming off is the audio cable to the sound card. If you have a cheapo meter (perhaps a 100uA current meter or a cheapo voltmeter) laying around, wire it in series with the detectors. That way you can use the meter reading to check the voltage (current) levels, alignment etc. Stick a stick in the barrel to help bore sight the alignment between the barrel and the detector.

Here is the detector I built for my small bore (~0.18") coil gun. The detectors are slot optoisolators (basically what you built from LEDs + photodiodes). The stick is "bore sighting" the alignment between the barrel (dark red) and the two detectors (black).

From here.

My setup also records the voltage across the power cap. Here's what the data looks like (0.177" BB at 42.5 FPS). The first peak is the voltage on the cap (320V dropping to ~0V in 0.33mS), the second two peaks are the BB moving through the two detectors.

[rp181]

Nice, i like the idea off seeing cap voltage versus projectile speed, easy to see if pulse length is good.

Two questions:
1) why do i need resistors? works fine without them.
2) is it possiable to power the LED's from the mic port with enough power?

I do have a meter, but i just like using the PC ocilliscope to check everything.

How did you connect it to the power cap without ruining the photo detector or PC sound card?

[dongfang]

Hi Jimmy,

OK, good that you can get it to work

The CDS cells I have played with must have been much slower than yours. Those things took hundreds of milliseconds to detect anything at all.

Semiconductor detectors, like phototransistors or -diodes are still 1000s times faster, I hope we can agree.

Regards
Soren

[rp181]

the way i see it, speed doesnt matter, each sensor takes the same delay, thus equaling out.

[jimmy101]

Nice, i like the idea off seeing cap voltage versus projectile speed, easy to see if pulse length is good.

Two questions:
1) why do i need resistors? works fine without them.
2) is it possiable to power the LED's from the mic port with enough power?

I do have a meter, but i just like using the PC ocilliscope to check everything.

How did you connect it to the power cap without ruining the photo detector or PC sound card?

The resistors are just to limit the current a bit. They are not absolutely required, especially if you are taking power from the MIC input instead of a battery. If you use a AA battery and a phototransistor (PT) the PT will draw ~0.1 Amp in full sun, that'll kill the battery pretty quick. THe resistor is just to drop the total current draw on the battery.

You probably can't power the LEDs from the MIC input. My MIC input sources 2.5V at ~1mA. Most LEDs require about 1.4V at ~20mA, so the MIC input can't source enough power for the LEDs. The MIC does supply enough power for PTs or photodiodes or photoresistors.

To monitor the photocap the LED in the optoisolator (slot optodetector actually) is just wired in series with a 22K resistor across the the launch coil. The photocap is charged to ~300V, 300V across 22K is about 14mA, which is about what the LED wants. You get a ~1.4V drop across the LED and the rest of the voltage drop is across the 22K resistor. The current draw of the LED (~20mA) is miniscule compared to the current through the launch coil (hundreds of amps), mostly because of the huge difference in resistance in the two parallel circuits. The schematic below shows how my 3 optocouplers are wired. Note that there is a big-ass diode across the launch coil that is not shown in the diagram. That diode is required to keep the LED from seeing a negative voltage.


The voltage detector opto should work with your O-scope setup. It's an easy way to isolate the HV from you scope's inputs. IF you get it wrong you fry the opto, which is a lot better than frying the input to your storage O-scope.

Dongfang
Are you sure they were CDS cells and not solar cells? Solar cells are really really slow. CDS are just slow. But CDS is fast enough to be used on camera flash units where they are watching for a ~1/1000 second flash.

rp181
That is true if the detectors aren't too slow (and if they are reasonably well matched). If the detector is really slow then it won't give a signal at all and won't work.