led multiplexing dim or ??

[MrDEB]

been getting some input from audioguru about multiplexing a 10 x 10 led matrix.
The issue of leds being to dim was discussed at length.
After finding this circuit I wonder if dim leds is an issue??
the planned circuit is attached (only one led of planned display is attached.
found another circuit that uses 2n3904 at both ends of the leds with one end having a 10K resistor connected to the base and a 300 ohm resistor connected to the collector.
Will try and locate circuit
Any advice??



https://www.best-microcontroller-projects.com/led-dot-matrix-display.html

 

[blueroomelectronics]

Not sure I understand how a pair of 4017 Johnson counters display anything useful. The original circut uses a PIC with a parallel output. The 4017 works fine as a column driver I use a Johnson counter CD4022 in my Dragonfly schematic to drive seven segment LEDs.
The schematic on the left each LED get 23mA which is fine, should be nice and bright.
Since you're building a spectrum analyzer why not just use ten LM3916's no muxing nessary.

 

[audioguru]

The link has a PIC with a 0.2V voltage loss when its output currents are a few mA.
The ULN2003 darlington driver has a loss of about 0.7V.
Red LEDs have a loss of 1.8V.
So the total losses are 2.7V and the 470 ohm current-limiting resistors create a current of 4.9mA which is not bright.

I can't see why your circuit is wired so oddly.

The display is multiplexed among 8 columns so each lighted LED has an average current of 4.9mA/8= 0.6mA which is very dim.

Maybe your LEDs look OK when they are very dim.

Bill Bowden's circuit on the left has pulses of 21mA in its LEDs. The 21mA is divided into 5 columns so each lighted LED has an average current of 4mA which is not bright and is not dim.

 

[blueroomelectronics]

Why do the ExpressSCH schematics look terrible? Is there no snap function?

 

[MrDEB]

I havn't fine tuned the schematic

I kinda tried some other pc board prgs but big learning curve.
This may sound dumb but why are we dividing the current by number of rows?
I thought each row comes on seperatly? The lm3916 controls the colums and the 4017 controls the rows or vice versa.
I am kinda lost or ??
AFTER LOOKING at the multi plexed display using a pic I see no difference between a pic and an lm3916 as far as sinking current?? Of coarse putting a transistor at both ends.
Lost in Salmon, Idaho

 

[audioguru]

An LED that has continuous 20mA current looks bright.
Multiplexing turns on one column at a time. So if there are 10 columns then an LED in one column is turned on for only 1/10th of the total time.

Your vision is slow and sees only the average amount of current which is 1/10th the continuous amount so the LED looks very dim.

 

[ericgibbs]

Why do the ExpressSCH schematics look terrible? Is there no snap function?

hi Bill,
It looks like he has the Snap ON, he just isn't taking care in placing the end of the wires.. that why its looks a mess.

MrDEB,
Try to arrange the circuit drawing layout so that it 'flows' from left to right across the paper, it makes it a lot easier to read, so you will get better forum feedback...
Also switch the 'grid' OFF, using the Options menu before you screen capture.

I know these points might sound picky, but a better presentation will get a better response..

If you build the 2 * 4017 [x,y] LED scanner, how do you plan to change the 'pattern' being displayed.?

 

[MrDEB]

The lm3916 should take care of it

This circuit is using the design of the Oriental circuit as its being called.
https://www.electro-tech-online.com/attachments/audio-spectrum-analyzer-en-pdf.15065/
On circuit drawings you mentioned SNAP ON. should I have it on or off? I need to check which I have set.

 

[blueroomelectronics]

The PDF you posted looks pretty good, why not just build it exactly as in the PDF? Why bother trying to redesign it?
If you want to get better with any bit o software, practice, practice, practice. Try drawing something simple. Make it look like other schematics on the site. Try to draw something simple.

 

[MrDEB]

The leds may be to dim

we are building a stereo EQ unit. So 2 displays are being built.
The leds being onm only 1/10 of a second would possiably be dim.
Audioguru is recommending hitting the leds with what ever the max pulse current is. I believe in this case the pulse current is 50ma (need to recheck the data sheet)
What we may do, seeing how we have several different colored smd leds , we may insert a pot in series with each row of the matrix. then adjust the brightness accordinally. Still have a resistor in series so as not to pop any leds.
Wanting to use same circuit for the audio out to an amp but with all the switching of the matrix, it may introduce tooo much noise. A TILT circuit may be implemented. Not sure.
This is a project that one of my Boy Scouts wants to build. I am just along for the ride so to speak.

 

[audioguru]

The LEDs must be multiplexed much faster than 1 second for all 10 columns.
You don't want each lighted LED to flicker so multiplex them at 100Hz or faster.

 

[MrDEB]

Thanks now just need to compute

proper transistor biasing etc.
Sure wish the unit could be used to pass the audio from input through the filters/multiplexing display then out.
would the quad bi lateral switches be quieter that the transistors used in the multiplexing. or ??
just wondering.???

 

[blueroomelectronics]

I still don't understand why you just don't build a non multiplexed version. Or at least the version in the PDF you posted.

 

[audioguru]

Hi MrDEB,
I notice a 1uF capacitor to ground at the output of the filter opamp. Then the opamp won't work, it will oscillate instead.

The capacitor is probably part of a peak detector circuit at the output of each filter opamp. The 1uF capacitor should be at the collector of the switching transistor.

 

[MrDEB]

Oups!!!!

Took another look at the oriental circuit--boy are you sharp.
I put the cap on wrong sife of diode.

Maybe I will get this thing built (Chris I mean) I need to order some parts so we can finish the pocket unit.
I looked at the circuit board he has assembled and Think he installed the wrong caps (TILT). I see mostly cermic ones instead of poly caps.
As I said before we may just input the audio in parrell (the eq display and a TILT unit but use dip ic's instead of the surface mount.
got to go.
THANKS for all your input

 

[audioguru]

Ceramic capacitors might have a tolerance of +80%, -50%. You need 5% poly ones.

What is a TILT unit?

 

[Ubergeek63]

Ceramic capacitors might have a tolerance of +80%, -50%. You need 5% poly ones.

What is a TILT unit?

I think Y5Us are generally 10% but vary that much over temperature and again over voltage.

Anything starting with an X in the material designation is much better, usually only 10%. COGs or NPOs, on the other hand, are very stable.

 

[MrDEB]

A tilt unit alters the wave form

instead of cutting it into sections
**broken link removed**
a different more modern approach than different frequency bands, as I understand it.

 

[audioguru]

I don't like "TILT".
I think you should not try to fix how you hear.
If my ear canal resonates then leave it like that because it is normal.
If the bass is not strong when the level is low then that is how things sound so don't "fix" it.

TILT is not modern, it was used in the 70's. May it rest in pieces.

I worked for a Pro Sound manufacturer. They devepoped a speaker with a diffuser for high frequencies but the diffuser messed up its frequency response. I made an equalizer for it that made it sound fantastic. Only 4 frequencies needed cuts and boosts.

My home stereo doesn't have equalization except for a +10dB boost at 35Hz so my little 8" woofers sound and feel like huge sub-woofers.

 

[MrDEB]

So I should forget the TILT?

just go with the 10 band equalizer?
any thoughts on running the audio as well as the display through same circuit ?
I even contemplated using 20 lm3916 or 3914's?
but then one must still add filters to each 39xx
I found where I can procure 3914's for $1.25 each