I did redraw it so that I could see the numbers, threw in some changes to match video.

 

PWM bread boarded with no glitches. Video on you tube, 100 0057 Bernard.
Values used: C 1 = 100 uF, R 1-2 = 15k, ( R 1 not used ), R 3 = 3k9, C 2 = 1 uF, R 4 = 1 k.
Q 1-2 not used, U 3-7 to gate of Q3, N ch. FET SM, 3055V. LED,s aquablue, 5 mm, 2.4 Vf, R's 3k9.
Connected U 1 & U 2 one at a time & checked before connecting to comparator, Grounded LED cathodes to be sure that all came on, connected FET & tested - gate grounded, LED's off, gate + 12 V- all LEDs on. Connected gate to U3-7 & tested- LED's pulsing going full on & full off. Connected 100 k pot. to + & -, ct. to U 1- 5, raised pin 5 from 8 V to where LED's just stayed on at minimum brightness.

 

Your video looks great and is what I am looking for. I looked at Wendy's PWM and made the changes as best as I could per your recommendations and all I get is steady lights on. No pulsing. I tore it apart and started over 3 times. Still the same results. It seems that the FET is always on. I disconnected the comparator and FET is still on. I have changed the FET and U1-3. Since you had no glitches, I have come to the conclusion that I don't know what the heck I'm doing. I drew out the modification per your description as best as I could. Please review and correct. Thanks for your help.

 

I did redraw it so that I could see the numbers, threw in some changes to match video.View attachment 106982

Ahh great! I will breadboard this circuit!

 

What comparator are you using, your print shows LM339, should be LM393, dual, just a typo ?
No power to comparator.

 

What comparator are you using, your print shows LM339, should be LM393, dual, just a typo ?
No power to comparator.

Crap! LM339

 

Page 4, post 76 shows LM339, pick one of the 4, ground inputs to unused .

 

Page 4, post 76 shows LM339, pick one of the 4, ground inputs to unused .

when grounding unused inputs do you ground both + & -?

 

I ground all unused inputs including + & - of comparators. My guess is that there is a 90% chance that a comparator would be stable with only one input grounded, but maybe a better choice might be to ground one input & tie other high.

 

I ground all unused inputs including + & - of comparators. My guess is that there is a 90% chance that a comparator would be stable with only one input grounded, but maybe a better choice might be to ground one input & tie other high.

Ok so here is a dumb question, ground inputs to + & - of 1inputs and the rest tie to +. Do you leave the outputs of the unused alone? Sorry, still learning.

 

No connection to unused outputs; for comparator, + input tie high, _ input ground.; gates , inputs high or low, or mixed, just somewhere.
When you have a working circuit & want to change appearance might try Stream #20. LED brightness is non linear vs current , a small change in current @ full brightness is not noticed, but same change @ dim is quite noticeable, which matches the RC discharge quite well, but RC charge is opposite which is not pleasing in throbbing LED's. A step in improvement is to use constant current charging of C by use of a LM334. By connecting a R between pins 1 & 3, current can be adjusted from 0 to 10 mA. For a 2 sec rise R would be around 100 ohms so a 500 ohm to 2k pot can be used to select timing. R1, 15k, was selected for a fall of 2 sec.
Adjusting U1-5 V helps some in keeping some current flowing in LED's at minimum brightness, but lowering input V to U2-8 by adding 2 series connected diodes between +12V & pin 8 seems better.

 

Ok, I didn't do the circuit from post #291 because I don't have the LM334 but I have been working on the original PWM with your updates and did the pot to 555 pin 5. For some reason I cannot use 12v and had to lower the supply to about 8v to get a little flicker as you will see in the 1st video.

This is the schematic that I used.

The 2nd video is without the pot to 555 pin 5. I just can't seem to get it like your video breadboard.

 

Schematic in post # 292 looks good. Do not make modifications until basic circuit is working, U1-5 no connection, save pot for testing.
Ground U3-2, output, and see if LEDs are off.
Temp. disconnect U3-4, - input, connect pot wiper to U3-4, & adjust to 6 V. LEDs should blink on & off at slow clock rate, about 2 sec. on, 2 sec. off. Look at Stream # 19, substitute a straight 6 V for fast clock, whenever red line is above 6V, LEDs are on.
I''ll be back later.

 

Schematic in post # 292 looks good. Do not make modifications until basic circuit is working, U1-5 no connection, save pot for testing.
Ground U3-2, output, and see if LEDs are off.
Temp. disconnect U3-4, - input, connect pot wiper to U3-4, & adjust to 6 V. LEDs should blink on & off at slow clock rate, about 2 sec. on, 2 sec. off. Look at Stream # 19, substitute a straight 6 V for fast clock, whenever red line is above 6V, LEDs are on.
I''ll be back later.

yes after grounding theu3-2 or gate of the FET, all the LED's go off. I added the pot wiper to u3-4 and i can change the pulse rate. looking better. the supply v is still about 10volts and when i increase to 12 all the led's stop pulsing and stay on with no control

 

Progress ?
Varying the pot should not change the pulse rate, just % time that LEDs are on, duty cycle; with pot below 4V (VCC 12 V ) LEDs should be on solid, 100 % duty cycle. Raise pot slowly, LED on time should shrink; pot above 8 V, LEDs off = 0 % duty cycle.
Re connect every thing. Disconnect lead from U3-5 ,+, connect pot to U3-5; pot low, LEDs should be off; slowly raise pot , LEDs should get brighter & above 8 V should be full on. No? Can you listen with hi Z earphones or such to U2 ( labeled U1 ) -3, should be loud 200 Hz +/_ . No ? bad or reversed C2, bad 555 ?. bad connection.

 

Progress ?
Varying the pot should not change the pulse rate, just % time that LEDs are on, duty cycle; with pot below 4V (VCC 12 V ) LEDs should be on solid, 100 % duty cycle. Raise pot slowly, LED on time should shrink; pot above 8 V, LEDs off = 0 % duty cycle.
Re connect every thing. Disconnect lead from U3-5 ,+, connect pot to U3-5; pot low, LEDs should be off; slowly raise pot , LEDs should get brighter & above 8 V should be full on. No? Can you listen with hi Z earphones or such to U2 ( labeled U1 ) -3, should be loud 200 Hz +/_ . No ? bad or reversed C2, bad 555 ?. bad connection.

ok, I am going to re-breadboard again from the beginning per schematic from post 292 and try to get it to do the fade on & off like yours. Once I get that then I will do the update. When I watched your video, and I saw your schematic in the background, correct me if i'm wrong, you said the NPN & the PNP is not needed?

 

This is open for comments, but charging gate C of say 20 pF thru a R of 1k takes about .02 u sec. which is a small fraction of fast clock of .04 sec. At high clock speeds in M cycle range & higher Cs , low impedance drivers are necessary. Here the 2N3904 & 2N3906 are not needed.
I would check operation of each of the 4 parts of the circuit before connecting them together. As U2 is too fast to see, use a 100 uF for testing, then replace with 1 uF . If U2 is operating, pin 3 should measure about 5 V, average of 11V & ground. Wish you success. Do not bother with mods yet, might be better ways to shape the output.

 

I don't know what to say. I have rechecked the connections and polarity and copied the schematic and I still get the same results. If the schematic is exactly what you built in your video then I don't know what is going on. I have replaced all the components 4 times and still the same result.It's like the FET is stuck on. The only thing different compared to your schematic is the LM339 is used and not the LM393 which shouldn't make a difference. I got so frustrated that I went back to Wendy's original PWM fig D and all I get is pulsing led's. I was so excited seeing your video of the fade in & out and the traveling marquee. It's almost 6am and I haven't even slept. Anyway I am going to sleep for a few hours and try again

 

I was going to suggest that before you tear your hair out that you send the populated breadboard to me to fix or maybe better maybe I should stuff a circuit board with Stream #21 in exchange for some of your surplus parts ?? Video on you tube, 100 0058 Bernard.

Video is gone??

 

hey i'm all for whatever works. You can send me or i can send you. what parts are you looking for? I might have something you need.