LM3914 change signal input from an LDR from Spike to Liniar curve

sparky 1

Joined Nov 3, 2018
756
Using the microprocessor to convert a non linear analog output is the other approach.
In either case the availability of good documentation makes adjustment stage have less trial and error.
When the results satisfy the goals then the choice is easier to make, it includes leveraging your time.
 

Thread Starter

q12x

Joined Sep 25, 2015
1,657
Using the microprocessor to convert a non linear analog output is the other approach.
In either case the availability of good documentation makes adjustment stage have less trial and error.
When the results satisfy the goals then the choice is easier to make, it includes leveraging your time.
What I already mentioned:
1597783461675.png
I still considering it, i didnt forget it. But now we are concentrating on the discrete options as much as possible.
This is a long road of experimentation and research and it will take awhile, we are aware. I think it will be an interesting solution in the end.
 
Last edited:

ronsimpson

Joined Oct 7, 2019
2,988
Pin 4 Rlo, try adding one diodes there. This will jack up the "zero" by 0.65V.
1597785394187.png
I am trying to hold the 10LED point constant but being the 2LED points together.
1597785667358.png
 

Thread Starter

q12x

Joined Sep 25, 2015
1,657
Pin 4 Rlo, try adding one diodes there. This will jack up the "zero" by 0.65V.
View attachment 215074
I am trying to hold the 10LED point constant but being the 2LED points together.
View attachment 215075
And the winner is ... YOU mister ronsimpson
I truly didn't expect it. OAU. What a breakthrough ! You are truly fantastic.
How did you think of it?
It is working and very good I may add. I mean... there are some super tiny super little hiccups but in 98% of time, is doing its job, very well.
You are something. Please, can you explain your line of thought? From the beginning?
I probably have to go into reading a bit more closely about 3914 specifications to find what you (probably) did in there.
Very unexpected ! Super. I will actually make a movie, put it on my youtube account and share it here, as a thanks to all who contributed to this project. A before and after. But it will take some time to make it.
I am not forgetting mister sghioto circuit, and I will test that too, with the clear red led. I really liked the idea and it made me super curious. I'll do it for science! :]
Ohoa... what a breakthrough, mister ronsimpson ! You are truly amazing. I have no words to thank you enough.
Please, explain your line of thought from the beginning until the end. I'm so curious.
 

Thread Starter

q12x

Joined Sep 25, 2015
1,657
aaaaaaa.... i jumped to happiness too soon.
Your diode on pin 4, definetly boosted UP the sensibility. Now I can wave my hand from a longer distance than before. And almost all the leds are getting in line, almost. 01 and 10 are the laziest, and sometime 02 LEDS. It's an improvement in sensibility but still, the same problems, but a bit not that obvious this time, thats why i jumped so quickly. Sorry for that. I feel sorry. But it is a breakthrough nevertheless. And i'm still happy. Its really a good thing.
 

Thread Starter

q12x

Joined Sep 25, 2015
1,657
Here is the movie I promised : I hope it will help you a bit to see the entire composition.
Sorry for the bad rezolution, for some reason my phone camera got stupid, and me with her.
Put a pause in the first 3 seconds of the movie and read what is there on the table to distinguish where components sit.
Right now, in the room light (that is not very bright) is working pretty good. But there are days when the light in the room get very bright when the sun is up. My LDR will get light from a blue led from about 15cm(7inch) distance from top to down. That is the original project i want to attach it to. To give you a context. Until then, Ill calibrate it to this room light I show in the movie.
You will notice how close ill have to get in the first part to turn off the 2 leds, especially the first. And then in the second part, how large the distance is to turn off the 1st led. Very good improvement. Thats why I jumped from chair in ecstasy. :]
To summarize it, Leds 1,2 and 10 are the uneducated black sheep for this circuit.
Ah, actually ill put you the image here :
(click on image to enlarge)
Screenshot_3.jpg
 

Thread Starter

q12x

Joined Sep 25, 2015
1,657
Could some of the light from the LEDs be interfering?
SG
Good question. I dont think so. I thought on it from the start. I dimmed the light of the leds so much, that even in darkness are unoticeble. Those are super bright leds, and you can barely see them lit. I admit is white on white, but trust me, it will glare the image when is fully lit. Also leds high are underneath the LDR high. And LDR is pointed away from them. Nothing to reflect light from leds to ldr either. But is not the cause for the black sheep ones. I will test your sugestion with a wall between them, very soon.
There is a 100k rezistor to dim all leds. I didnt include it in the original circuit. I will soon. My mistake. But I was thinking this 100k here is for tests and when circuit is fully developed, I will lit them fully, without any R on them.
Screenshot_4.jpg
 
Last edited:

Thread Starter

q12x

Joined Sep 25, 2015
1,657
Ok, I made another very short video. I find out I had 2 applications for video/photo, one very bad that I just remove it. Now everything looks better.
Things to notice in this new video:
- it is daylight and is much more brighter than night ceiling light like 2 times brighter.
- I put 2 diodes in series as you sugested ronsimpson and it get even more sensitive and more linear i may risk to say. It's not that obvious from the movie, but i feel it. Or I am too sensitive now.
A) The thing is that with wire on pin4, will give very high error on led1 and 10. -10 being hiper sensitive responding too quickly and 1 almost immune and irresponsive.
B) With 1 diode on pin4 things got very good in the night light, but now in daylight, it was way too sensitive to light and barely seeing my shadow. I had to do something to reduce its sensitivity.
C) putting 2 diodes in series, lowered the sensitivity even more, though is quite sensitive as it is now, i had to put 2 shadows to see the effect, but the linearity is way more better than before. By linearity i mean those 2 black sheeps, led1 and led 10, are closer in responsiveness to the rest and following them as expected. The daylight is influencing quite much and terribly and I would not completely trust this particular experiment, though is impressively close to that straight line , like the POT line. Very close. I must check this formula at the night light too.
Here is the video (0.25s) :
 
Last edited:

Thread Starter

q12x

Joined Sep 25, 2015
1,657
I just put in 3 diodes in series mister @ronsimpson , and it is GOOD, VERY GOOD. Now we can jump in ecstasy. VERY big difference. I know a diode is eating 0.7V so I eat 2.1V from pin 4 with those 3 diodes in series.
I am not speaking very often in english so i have to search for my words. I would wanted to tell more details. I hope i touch the essentials. I must put it to more other tests, like other brightness of other sources. It would be nice if I can vary the sensibility from that pin4, like i just did adding all those diodes. In the end it must have a blue led as single source light. I tested it already and is like bum, on/off all at once, but i put a paper over ldr to difuse it a bit and it did the trick good enough. I hope i can find a difuse plastic somehwere, or make one with a sandpaper maybe.
I have an excelent result so far, with moderate bright light from outside and also brighter from the lamp + daylight. I'll have to test it in the night light also. But im more interested in the bright light since it will have one for the final application.
I'll also have to update the original schematic too.
Mister @ronsimpson and mister @sghioto all my respects to you. Thank you so much for your invaluable effort and help. You both deserve it and you are both the winners. :] Haha.
Here it is:
 
Last edited:

Thread Starter

q12x

Joined Sep 25, 2015
1,657
Yes, is kind of darkening so I open the room light and It is best working with 2diodes comparatively with 3 diodes used in daylight. That is all, I suppose. I'll have to integrate it into my other circuit and make it work with a superbright led . I'll update you further. I am wondering if is possible to fine adjust the pin4 voltage leak?
 
Last edited:
Top