laser blanking

[bobhaha]

NO, no no no no.

TTL can not control brightness by means of lowering or increasing frequency.

How TTL works is the driver has a + and a negative input as well as a signal input. When +5V is put in as signal, it then turns the laser on, drawing from the + and - inputs and outputting a controlled current to the diode.

Diodes are able and regularly do run into the 100+ kHz range and there is no apparent change... and this is 100,000 times a second we are talking about here! Diodes have the capacity to turn on and off at a great speed which makes the idea of soft start impossible.

Also the idea that our eyes will blur the lines into each other is incorrect. I have built a system that will turn a laser on and off with 1ms delay in between each pulse... and moving it slightly from about 20 cm you can see the separation of the lines. There is a reason why laser shows have not done this... companies spend 100k+ on some laser systems... if it was that easy they would know about it.

The only way to dim the laser is to use analog modulation, meaning it will vary the power to the diode directly. This is the only logical and sure fire way to do it.

 

[anselm]

How TTL works is the driver has a + and a negative input as well as a signal input. When +5V is put in as signal, it then turns the laser on, drawing from the + and - inputs and outputting a controlled current to the diode.
....
The only way to dim the laser is to use analog modulation,

You seem oblivious to the concept of pulse width modulation (or PWM).
Please read up on it.

Also the idea that our eyes will blur the lines into each other is incorrect. I have built a system that will turn a laser on and off with 1ms delay in between each pulse... and moving it slightly from about 20 cm you can see the separation of the lines.

You are not taking into account the obvious fact that 1ms<=> 1kHz is far too slow for proper PWM...

 

[chipdouglas]

anselm... you said earlier that you have not made one. so why is your speculation better than 2 members that have actually made it?

michael.

 

[anselm]

Only member in this thread who does have TTL experience as far as I can tell, vk2fro,
is actually not contradicting
me, in fact even backing me up by saying that the driving signal is almost the biggest issue.
Being more complex than necessary, since analog modulation isn't really more expensive.

I stand by my point. Not because I like to argue for the sake of arguing, I simply like
to get to the bottom of things.

So please, if anyone with actual TTL experience would like to enlighten me, please do.
But I'm not going to be convinced by people who only have read some things in some
threads, yet don't know about PWM and persistence of vision.

 

[ElektroFreak]

Well, having a good bit of experience with TTL and analog I'll weigh in a bit here.

Pure TTL signaling allows for only 7 colors, as has already been stated. PWM TTL is a way to mimic analog, but it only really works well with a single-color system. With PWM dimming, the output is not continuous, rather the length of each pulse is varied to simulate a change in brightness. Because of this, as the laser "dims" using PWM the pulses get shorter and shorter, with more and more off time between pulses. If we use PWM in a projector, we end up with scanned lines made up of tiny dots any time the laser is not fully on. These dots are detrimental to the quality of the scanned output. What is really ideal is a smooth reduction in output, proportional to some sort of signal voltage. This way, when the laser dims it is still on 100% of the time, and will still produce smooth, continuous lines.

When we combine colors in a projector using PWM TTL, we still only get 7 colors. The reason being is that the intensity of the beam isn't really reduced. It's just being pulsed to fool your eyes into thinking it's being reduced (through the "persistence of vision" that anselm is describing).. Because of this, while it might be possible to sort of "fool your eyes" into believing that more than 7 colors are being produced, in reality there cannot be more than 7 because of the math involved in combining TTL colors. Analog is therefore always far superior since you are actually actively blending colors to produce blended hues of varying quantities of primary colors.

 

[anselm]

Thank you very much, ElektroFreak!
I'm gonna shut up very soon now, but could you please just tell me real quick if this is BS or truth:

Also, unless the modulating frequency is a multiple of the scanning frequency,
those dots would be in a different place for every scan.

 

[ElektroFreak]

Without verifying mathematically, that does sound about right.. but most of the time persistence of vision isn't quite enough to replicate the kind of results that can be obtained using true analog modulation. I've got a couple single-color TTL systems, and the Quickshow software I use can provide PWM dimming.. the dots are almost always clearly visible in the scanned output.

 

[anselm]

Thanks a bunch for your insight.
So to resume this whole issue:
While PWM TTL theoretically could produce all kinds of colours and provide a decent
picture, in practice the modulating frequency is just not fast enough to produce a decent result.

 

[chipdouglas]

What!?
Are you sure? Because if you modulate fast enough, even if it is only ON and OFF,
the eye gets the impression of varying brightness, allowing for much more than 7 colors...

Or am I missing something obvious here....? :thinking:

Thanks a bunch for your insight.
So to resume this whole issue:
While PWM TTL theoretically could produce all kinds of colours and provide a decent
picture, in practice the modulating frequency is just not fast enough to produce a decent result.

since we are clearing things up... your original post was stating that ttl can do all the colors. pwm wasn't mentioned till later.
michael.

 

[Cyparagon]

1. The dots might be in a different place every scan, but they'll also be different lengths and you'll invariably get some areas that are brighter than others. Fairly often (especially with complex graphics and/or slower scanners) the framerate drops below 30, and then it will be very noticeable.

2. There will be some cases where the modulation IS a multiple of the scanned points, since the number of scanned points varies a LOT throughout a show.

3. There is no show software I'm aware of that can fade colors via PWM, and probably for the reasons already stated.

 

[planters]

A question re. blanking/shuttering. If an array of several diodes is being driven by a lab style power supply and you wish to turn the lot on and off under the control of a 0-2.5V signal such as comes out of Quickshow, but you can't modulate the primary supply, mechanical shutters are too slow and polarization rotators are too expensive is it practical to interpose a relay between the power and the diodes without introducing damaging voltage spikes? Could it handle amps and how fast could it be {expensive}?

 

[Cyparagon]

is it practical to interpose a relay between the power and the diodes without introducing damaging voltage spikes?

Absolutely not. When a constant current driver senses there is no current draw, it raises the voltage to try to push the current through. When the relay closes again, your diodes are toast.

Additionally, relays are even slower than mechanical blanking.

 

[anselm]

You might instead consider a transistor which "redirects" the driver's current over a bleed resistor when you apply the blanking signal to it's gate...

 

[Bionic-Badger]

Thanks a bunch for your insight.
So to resume this whole issue:
While PWM TTL theoretically could produce all kinds of colours and provide a decent
picture, in practice the modulating frequency is just not fast enough to produce a decent result.

Also remember that when you use PWM for anything, it relies on the signal passing through "low-pass" filters that smooth out the signal into an "effective" analog signal. It works great on devices like motors, and a PWMed signal will have a measured lower voltage on a multimeter from the slow ADC conversion. However, if the device you're running PWM to has a very wide frequency response and/or the resolution ends up affecting the PWM modulation wavelength (like spreading out a whole PWM cycle over 5 meters of space), you may not get the effect you want, causing noticeable pulsing. It might be better, if possible, to pass the PWM through a low-pass filter to convert it into an analog signal, though that won't fly with TTL.

 

[NME]

Seems like anselm's gettin a bit of a hard time I'm along his exact train of thought though. PWM seems more than plausable to me, however pulse speed would be critical.

I think it fairly unlikely that the pulse frequency would ever be a multiple of the scanning frequency for an extended period of time if at all, considering the infinite number of frequencies out there. And even if it were, if the pulse frequency was a MUCH larger multiple of the scanning frequency, it would still be practically invisible due to the closeness of the points (Ill admit, for a larger image this proportionality would have to increase dramatically).

Elektrofreak's Post #21 hits the nail on the head. It is however all about scan speed vs ttl speed. (And mathematically the closeness between the scan frequency and (ttl frequency)/N where N is a whole number giving a resultant frequency as close to the scan frequency as possible. The difference between them needing to be greater than one can visually perceive) The degree of closeness becomes less and less critical as the PWM increases relative to the scan frequency, or as both speeds increase.

1. The dots might be in a different place every scan, but they'll also be different lengths and you'll invariably get some areas that are brighter than others.

^ This seems irrelevant to me as the same occurs with an unblanked constant beam: If you have a dense area of points it will be brighter than a sparse area, even for a CW beam.

Also remember that when you use PWM for anything, it relies on the signal passing through "low-pass" filters that smooth out the signal into an "effective" analog signal. It works great on devices like motors, and a PWMed signal will have a measured lower voltage on a multimeter from the slow ADC conversion. However, if the device you're running PWM to has a very wide frequency response and/or the resolution ends up affecting the PWM modulation wavelength (like spreading out a whole PWM cycle over 5 meters of space), you may not get the effect you want, causing noticeable pulsing. It might be better, if possible, to pass the PWM through a low-pass filter to convert it into an analog signal, though that won't fly with TTL.

Ya mate, that IS analogue , albeit PWM controlled. They're on about straight-up PWM using your human visual response to do the 'low-pass filtering' for you. Yours is probably a better idea, avoiding all aforementioned problems.. Doesnt seem to hard to contruct analogue when you put it like that...

 

[chipdouglas]

bottom line here is that ttl is 7 and anolog is all... and next point is to do things the right way the first time. especially when you are constructing it from scratch and there isn't really a price differece between a ttl driver and an anolog driver.

michael.