Open Source Scanfail Unit (phase 1)

[Miss A.]

Karl is talking in this thread about 'more input guys', but how can I convice him that discrimination here is not allowed?
Well, here is my input:



How it works:
X is the input. This signal is delayed for 5 milli seconds. (safe enough for your eyes). The two X-signals are compared. If one of the signals is still at the same level after 5 ms, the opto coulper will shut down the laser module.
For Y the same story. The NOR gate will combine the two signals to the opto. In fact you have to set three opto's parallel to shut down the three colors.

Greetings,
Audrey

[Banthai]

oops sorry Audrey,

we are just not used to having the female input on this forum ....
"Girls with lasers" is something that we never expected,

but it would make a great title for a new publication

I think im digging myself into a bigger hole here

ive done an edit to the post

but if we had a basic circuit perhaps others will help us out with this thread ( come on Guys n Gals ! ) more input the better

but seriously ....

Audrey .... your circuit diagram is a great start .....

is the 5 ms delay just implimented using an R/C network on one side of the comparator or are you thinking of doing it another way ?


all the best ... karl

[Miss A.]

Hi Karl,

I'm not sure if an R/C network is accurate (and stable) enough.

In the world of electric quitars they often use a so called Bucket Brigade (not a joke!) module that creates distortion effects.
It's mainly a shift register chip and a clock:
The analog signal comes in, it will move to the next place in the memory and will come out after the preset time. Kind of a memory chip.
Please search for the TDA1022 with Google or read this: http://www.geocities.com/scaprile/fxa.html

[stanwax]

Ahh the TDA 1022 - haven't messed with one of them in a long time (read 20 years!!!).
Audrey I think the concept is good but I doubt the 1022 is gonna be right for the job - my reasons - you need a split 15V supply for them a dual phase clock and means of cleaning the clock frquency out of the output. I think this could end up with the thing being over complicated - like Bill says every component is one more to fail plus if you can still buy them I doubt they will last for ever as its so much simpler these days to do the same and more with a digital delay. I think simple RC would suffice as we dont need precision and even though RC would be frequency sensitive all we need from your idea is to measure the value at two different times in effect. As long as the time did not become too long then we would be OK.

Rob

[Miss A.]

Mmm, yes I'm open for every idea, but you still have to compare the input value with the value of 5 ms before. I think this can only be done with a shift register or momory chip or someting...
Can an R/C network create a delay in signal? I thought R/C can only be a filter.
Here we are talking about DC voltages, not AC with a frequency.

(Other Bucket Brigade chips are TDA2108, MN3207 and MN3102.)


.

[stanwax]

Though I still think its making the circuit more complicated than we need to make your idea work Audrey maybe a better idea is to use sample and hold instead of the delay. If you sample every xxxmS then you will get the same effect as having a xxxmS delay.

Rob

[Miss A.]

Yes, it's true that in that case sample-and-hold is the same as a delay time. But I think we are both on the wrong track.
Within the 5ms, the X-signal could change and be back at the same voltage level as at 0ms.

Example:
0ms > 3.36V
1ms > 3.36V
2ms > 3.36V
3ms > 4.00V
4ms > 4.00V
5ms > 3.36V

In this case the signal is at 0ms the same as at 5ms.
Our safecircuit would than shut off the laser, which is wrong, because the signal has changed witin that 5ms.
So, we can not use sample-and-hold or a delay function at all

Our circuit has to be reset every time the signal has changed.
In the example above it has to 'count' from the beginning at 3ms, because here the signal has changed from 3.36V to 4.00V.

Please tell me if I'm wrong about this.
I'm having my period this week, so I can not think very clearly right now.

.

[stanwax]

It did cross my mind that this may be a problem - which is why its probalbly best to just compare the ilda signal with the actual scanner feedback, that way you are testing it with real values in real time.

Rob

[mixedgas]

bucket brigades are %$#@ hard to get in the US,the local chip sales guys list them as "obsolete" what Audrey wants is a "all pass filter"
we use them in amateur radio to generate phasing method single sideband , they consist of a string of OPAMPS and a few Rs and Cs. The best you can get out of a single opamp is ~270 degrees of shift , for a flat 90' of shift from ~100 hz to 3 khz, we usually use 3 opamps.

It would be easier to use a atmel or pic with its built in a/d and a dac to introduce the delay, but that brings us to our NO micros/No dsp rule.

OK, Audrey, also look for a 16:1 cmos multiplexer chip, and use a string of 16 caps. Say CD4067 ? That should get you 16 clocks worth of delay and if each clock is a point, well you get the idea. So how about a switched capacitor filter (something like a MF10 or MF100) ran as all pass. many folks make them ie Maxim, Linear Technology, National Semiconductor,

Audrey, should I ever start up a full size laser company again, Your Hired as chief engineer !

Steve Roberts

[dave]

Yes, it's true that in that case sample-and-hold is the same as a delay time. But I think we are both on the wrong track.
Within the 5ms, the X-signal could change and be back at the same voltage level as at 0ms.

Example:
0ms > 3.36V
1ms > 3.36V
2ms > 3.36V
3ms > 4.00V
4ms > 4.00V
5ms > 3.36V

In this case the signal is at 0ms the same as at 5ms.
Our safecircuit would than shut off the laser, which is wrong, because the signal has changed witin that 5ms.
So, we can not use sample-and-hold or a delay function at all

Our circuit has to be reset every time the signal has changed.
In the example above it has to 'count' from the beginning at 3ms, because here the signal has changed from 3.36V to 4.00V.

Please tell me if I'm wrong about this.
I'm having my period this week, so I can not think very clearly right now.

.

Little too much info there..