Lasersaber strikes again. A joule thief king ?

[TinselKoala]

That's a good idea. I can test the component temperature hypothesis easily enough with some "cold spray".

My own DSO, when used at slow timebase settings, is also correspondingly slow to refresh the display, so you don't see anything like the "live" display as shown on the analog scope. But a full scan would show, as you state, "bunches" and "voids" where the slow-period oscillations happen.  Somewhat as below probably, from a different circuit entirely.

[scifi123]

That's a good idea. I can test the component temperature hypothesis easily enough with some "cold spray".

  Another way to test the "thermal oscillator" hypothesis is to put the transistor (and/or the LEDs) on a big enough heat sink with a good thermal contact (maybe using thermal paste).
  Since the heat sink has a much bigger thermal inertia than the transistor case, the oscillation should disappear.

[scifi123]

My own DSO, when used at slow timebase settings, is also correspondingly slow to refresh the display, so you don't see anything like the "live" display as shown on the analog scope.

  Most of the digital scopes have a display refresh rate of 20-60Hz (even the Chinese ones) so you should be able to see it, I think.
  Here are some pointers:

1. Set your timebase to around 3-10 times the duration between spikes. For example, if the spikes have 1ms between them, you can set the timebase to 10ms.
  You can use the analog scope to measure the duration between spikes.

2. Set your sampling period to be 100-1000 times smaller than the timebase. In the example above, the timebase is 10ms, the sampling period should be 10-100us, i.e. the sampling frequency should be 10-100kHz.

3. Set your trigger level to be somewhat smaller then the peak of spikes. If the spikes peak at 5V, a good triggering level would be 4V.

  That should be a good starting point for visualizing the "shrinking and expanding" on a DSO. If your DSO is fast enough, you can decrease the sampling period to be 10000 times smaller than the timebase, i.e. 1us for our example.

  Happy hunting ! 

P.S. What is the model of your DSO ? Maybe I can take a quick look at its specs to be more helpful.

[TinselKoala]

Thank you for your advice. I am always willing to learn more about the use (and abuse) of oscilloscopes.

The scopeshot above isn't really representative of the circuit we are discussing, it was just the first one I found in my data that showed approximately the same kind of thing. The scope is buried under a pile of stuff at the moment; perhaps I'll  dig it out and do an actual shot of the actual device, later on.

The DSO is an old Link scope made in 1997 that uses the PC (bidirectional parallel port interface!) for display and control. It is the Link DSO-2100M, generously donated by a friend from another forum. I am using the latest available version of the software; I don't really think you are going to be able to tell me anything about this scope that I don't already know, but I'll be interested in your comments regardless. Thanks!

[Redoanullah]

Has anyone successfully replicated laser saber's v4.1 ??
  I want to know in detail about it!
I tried the SJR looper with 22# SWG -100 turn as primary, 300# SWG -200 turn as secondary and used 2N2222, 2N3904, MJE3055 transistor.
.
I failed.
It doesn't even worked a bit!!
.
I think there is something in it's switching that matters!!
.
What's that capacitor doing across that schottky diode?