LED Matrix - 12V

A "new fashioned" televisor, using an Arduino to drive the motor and display.

Moderators: Steve Anderson, Dave Moll, Andrew Davie

LED Matrix - 12V

Postby Andrew Davie » Sat Jun 10, 2017 6:11 pm

I've made a simple start on the 12V LED matrix. This time I'm doing the circuit board first - I may mockup a hand-wired one, but thought I'd get the basic layout sorted. I'm going with LEDs in strips of 3. There are 10 strips of 3, at about 3.5V each LED that makes 10.5V / 25mA per strip. For 10 strips I figure 250 mA total current. I put 47 ohm resistors there just random values, really - to balance the load. An external resistor on the main circuit board actually controls the total current to the LED matrix board. Currently 16 ohms as suited to the original 40 LED board; I'll recalculate that soon.

The board looks like it will be about 56 mm x 35 mm.

ledmatrix.jpg
ledmatrix.jpg (249.24 KiB) Viewed 3527 times


ledmatrix2.png
ledmatrix2.png (121.54 KiB) Viewed 3527 times


The attachment contains the Eagle files for the board and schematic.
I place the terminal connector on the rear-side of the board so that the LEDs are unobstructed and there's easy access for wiring/connection.
Attachments
ledmatrix.zip
(32.52 KiB) Downloaded 83 times
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1508
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Re: LED Matrix - 12V

Postby Klaas Robers » Sat Jun 10, 2017 7:12 pm

Andrew, before you proceed: breadboard one string of 3 LEDs and one resistor of 47 ohm, apply power and measure voltages and currents. You need to reserve some voltage for the switching transistor.
User avatar
Klaas Robers
Frankenstein was my uncle.
 
Posts: 1431
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Re: LED Matrix - 12V

Postby Andrew Davie » Sat Jun 10, 2017 9:21 pm

Klaas Robers wrote:Andrew, before you proceed: breadboard one string of 3 LEDs and one resistor of 47 ohm, apply power and measure voltages and currents. You need to reserve some voltage for the switching transistor.


I just revisited the for-sale page I purchased from. The LED is rated 3.2V-3.4V forward voltage, 20mA current.
So, let's say worst-case 3.4V x 3 = 10.2V (from the 12.3V supply) which leaves 2.1V. Best case 3.2V --> 9.6V, leaving 2.7V.
I need to revise my knowledge - glad I wrote much of it down.

But, assuming (say) 3.3V per LED then x3 = 9.9V. With a 12.3V supply that leaves 2.4V at the "base" of the matrix. To supply 20mA per LED string x 10 = 200mA for the whole matrix. 2.4V supplying 200mA --> resistor = 12 ohms. Given the 16 ohm resistor already installed, that wold be 150mA --> 15mA per strip. Not super bright but should be OK for testing.

So... build a single-string of 3 LEDs. Use a resistor of (say) 16 ohms. Measure the voltage drop across the strip - measure voltage at base, then voltage past LEDs... gives me the voltage across each LED (divide measured by 3). Measure the current drawn by the strip. I should find V=IR. Input V is 12. R is 16 ohms. I should ideally be 20mA. Then I modify the R value to ensure 20mA max....

OK, the above would work for a stand-alone strip connected direct to 12V. What I'm not quite sure is how to test with transistor in place; I build it, choose appropriate resistor (or use the one already in place on the board) and test. If the LEDs work, then I'm ok with voltage... if not then I'm going to have to use fewer LEDs or a higher voltage.
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1508
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Re: LED Matrix - 12V

Postby Andrew Davie » Sat Jun 10, 2017 10:50 pm

Here are some results.

First, I connected 3 LEDs in series on a breadboard.
To the positive side of this, I connected a 30 ohm resistor - I figured it was "close enough".
I connected the other side of the resistor to the VCC being fed IN to the Arduino (that's from my external PS). Value of that measured 11.4V. The power brick will measure something slightly different, of course. I just measured it - 16.9V with no load... mmmh. It clearly is marked "OUTPUT 12VDC". So much for that.
Anyway, back to the measurements on the LED.... 11.4V supplied. I measured from GND to the positive side of each of the 3 resistors. A=3.35V, B=6.75V, C=10.11V. This concurs very well with the forward voltage of the LEDs being 3.4V.
Next, I measured the current by putting the multimeter in series between the resistor and the positive of the LED that resistor was going to.
I measure 0.026A. 26mA.
So, putting it all together...

3x3.4V = 10.2V, leaving 1.2V from the 11.4V supplied. The resistor is taking a bit, and I measured across that an actual "remainder" of 1.05V.

1.2V / 0.03 (30 ohm resistor) --> 40mA current
1.05V / 0.03 --> 35mA current

Not quite consistent with my measurement of 26mA.

Now I'm not quite sure what to say. With a 30 ohm resistor I'm allowing a bit too much current to the LEDs. They seem happy. Bright, but happy. If I wanted 20mA, then I'd change the resistor to 1.2/0.02 = 60 ohms. The resistor on the LED *driver* board is something like 16 ohms at the moment. Rather than pulling that to bits, I could set the "balancing" resistors to cope with the change. So, I just tested - switched to a 56 ohm resistor (instead of the 30) and the measured current is now 21mA. LEDs still very bright.

If I *keep* the existing 16 ohm resistor, then I'd want to make the "balancing resistors" about 60-16 ohms each --> 47 omhs, say.
That way I can use the existing board that feeds 22+V to the current LED matrix with the 16 ohm resistor. Just feed 12V to the board instead, and connect a new "12V capable" matrix to it, with 47 ohm balancing resistors. And there's roughly 1V "left over".

How does that sound?


led2.jpg
led2.jpg (233.68 KiB) Viewed 3515 times
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1508
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Re: LED Matrix - 12V

Postby Andrew Davie » Sat Jun 10, 2017 11:23 pm

Mmh. I just "ran" the televisor with the LEDs still connected. They dimmed significantly when the motor started running the disk up to speed, and came back to really bright once the disk was basically at speed and the motor wasn't under load. I guess my 12V supply isn't supplying enough current...?
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1508
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Re: LED Matrix - 12V

Postby Andrew Davie » Sun Jun 11, 2017 1:34 am

Have ordered some 12V 2A power bricks from China - so that will be my target - run in under 2A, and of course assuming they can actually do 2A which I doubt.

Meanwhile I've been re-reading the 2007 threads about LED matrix design. And, with reference to this post it seems to me that with 1V "left over" after the LEDs, then the resistor on the emitter of the TIP122 (not a TIP31 anymore) would have to be be something like 5 ohms. I got that by calculating the array (10 strings of 3 @ 20mA) would require 200mA, and there was 1V at the "base" of the array, so 1/0.2 --> 5 ohms. That's low - and doesn't give me much leeway for balancing resistors.

Maybe it's better to bite the bullet and have 20 "strings" of 2 LEDs. Then I'd have 400 mA but would have 11.4V-(2*3.4V) = 4.6V and that would make the limiting resistor 4.6/0.4 --> 11.5 ohms which gives me (only slightly) more leeway. Bummer, though. Just out of interest - 30 LEDs in parallel = 0.2 mA /string and 600 mA total. The voltage would be 11.4-3.4 = 8V and resistor 8/0.6 --> ~13 ohms not worth the effort.
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1508
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Re: LED Matrix - 12V

Postby Andrew Davie » Sun Jun 11, 2017 1:45 am

I'm currently thinking of alternatives. How about I have a higher voltage on my input supply. Say, 15V. I run the motor direct off that, and the LED matrix too. I get a DC-converter to convert down to 5V for powering the Arduino. 15V would give me (15-3*3.4)=4.8V left at 20 mA --> 24 ohms which seems much more workable for current limiting resistor + balancing resistors. Assumption is I can run the motor at 15V and that I can get a good DC converter.
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1508
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Re: LED Matrix - 12V

Postby Andrew Davie » Sun Jun 11, 2017 2:04 am

Andrew Davie wrote:I'm currently thinking of alternatives. How about I have a higher voltage on my input supply. Say, 15V. I run the motor direct off that, and the LED matrix too. I get a DC-converter to convert down to 5V for powering the Arduino. 15V would give me (15-3*3.4)=4.8V left at 20 mA --> 24 ohms which seems much more workable for current limiting resistor + balancing resistors. Assumption is I can run the motor at 15V and that I can get a good DC converter.


Well, technically I have a power brick that, though it says "OUTPUT 12VDC" is measured at 15.9V - so I could go with that.
And also the Arduino is supposedly capable of inputs up to 20V. So I could give it a go, I suppose.

15.9V-3*3.4=5.7V/0.2=28.5ohms -- that would work, I could have 16 ohm balancing resistors and 12 ohms limiting.
The Arduino would be uncomfortable I suppose, but within absolute range of voltage input (20V)
The motor would have to handle 16V, though. Seems to be plenty of motors that can handle that.
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1508
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Re: LED Matrix - 12V

Postby Steve Anderson » Sun Jun 11, 2017 5:02 pm

With all these 'power-bricks' check if they're regulated, mostly they're not, so a 12V/2A version may be close to 12V with a 2A load, but at lighter currents (or none at all) it could be close to 18V or more. Also at full load the ripple could be significant, it depends on the size of the internal reservoir capacitor..which of course you can't see.

If I had known you were going to place an order I would have suggested getting 18V versions and doing the regulation yourself. The power-bricks are mainly to get around the certification requirements of the museum...if that's why you're doing this. It's primarily the internal transformer and the 220V wiring (where present) that they're mainly interested in, anything downstream doesn't really have any bearing as long as it's all low-voltage, say below 50V.

An alternative is to use an old laptop power supply/charger. These usually provide around 19V at several amps. They're efficient, small and dissipate little heat, and of course, certified. Then regulate the 19V down to 12V yourself. The only downside is they are 99.9% switched-mode devices so they may have a small amount of noise on the output, but all being well the downstram 12V (or whatever) regulators should mop up most of that.

Steve A.

An afterthought...I also hang on to the old-style mobile phone chargers, I have several old Nokia versions that put out 5.7V at 500mA or so. Maybe not so good for analogue circuitry but with a low-dropout 5V regulator perfect for most logic-based devices. e.g. an LP2950-5.0 delivers 100mA in a TO92 transistor 3-lead package. 3.3V systems even easier an far less heat dissipated in the 3.3V post-regulator. (A LP2950-3.3). There are others too, and quite a few that will deliver higher currents if you should need them.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 3800
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: LED Matrix - 12V

Postby Klaas Robers » Sun Jun 11, 2017 7:04 pm

I agree with Steve: use a higher voltage for the LED cluster supply. If you look at your diagram in the link that you gave some posts up, you see that the voltage at the bottom of the LED cluster, also the collector of the video transistor, is quoted as 6 volt as a minimum. Now follow me step by step.

- The voltage over one of your LEDs is 3.5 volt at the max. current of 20 mA.
- If you place (just) 2 LEDs in series, the voltage over the both LEDs will be 7 volt.
- Add 1 volt for the current balancing resistor (47 ohm for 20 mA, thats correct),
- that will give 8 volts for the cluster chain.
- Place that on top of the 6 volts of the collector,
- and you need 14 volts minimum for the LED power supply, not 12 volts.!

So: for a supply of 12 volts, you can't circuit 2 LEDs in series, and have to place all LEDs in parallel.
- That gives, for 30 LEDs, a total current of 600 mA.
- 30 LEDs, and 30 balancing resistors of 47 ohm.

And: for a power supply of 15 volts, you can circuit 2 LEDs in series for each branch.
- That gives 15 parallel branches of 2 LEDs in series,
- with in each branch one balancing resistor of 47 ohm,
- a total current of 300 mA.
- 30 LEDs, and 15 balancing resistors of 47 ohm.

But is you use a laptop power supply of 19 volt, abundantly available from old laptops,
- You can place 10 parallel branches of 3 LEDs in series,
- with one balancing resistor in each branch,
- a total current of 200 mA
- 30 LEDs and 10 balancing resistors of 47 ohm.

The advantage of a lower total current is that the transistor heats up less,
- for 600 mA it dissipates about 3 watt,
- for 300 mA it is about 2 watt,
- for 200 mA it is about 1 watt.
- In all cases the transistor needs some from of heat sink.

And yes, you might stabilize your own 12 volt from the 19 volt ( 7812 ) and then 5 volt from the 12 volt (7805). Then you still have only one power supply brick for the museum.
User avatar
Klaas Robers
Frankenstein was my uncle.
 
Posts: 1431
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Re: LED Matrix - 12V

Postby Steve Anderson » Sun Jun 11, 2017 7:54 pm

Actually I wasn't quite thinking straight, but Klaas's idea of using the 19V supply from an old laptop PC supply/charger on the LED drive makes a lot of sense. It's still a regulated supply and if there is any noise it shouldn't upset anything. This can be further regulated down by 78xx's to whatever you need for the rest of the circuits. Are you using PWM to drive the LEDs, or an analogue arrangement? If using PWM then a switched constant-current source for each 'string' is ideal. I'll elaborate if PWM is your chosen mode.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 3800
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: LED Matrix - 12V

Postby Andrew Davie » Sun Jun 11, 2017 10:08 pm

I tried a little experiment today. I wired up three strings of three LEDs, without balancing resistors. I just had the original 56 ohm resistor in series with the array. I wanted to see/understand the effect of the LEDs without balancing resistors. Basically, none visible. They all worked fine and looked equally as bright. It makes me wonder just how necessary these balancing resistors ARE, because the resistance across each string of three LEDs seems to be pretty constant.

There are a few things I'm not "getting", once again. But thanks Klaas and Steve for your patience and assistance.

Klaas Robers wrote:I agree with Steve: use a higher voltage for the LED cluster supply. If you look at your diagram in the link that you gave some posts up, you see that the voltage at the bottom of the LED cluster, also the collector of the video transistor, is quoted as 6 volt as a minimum. Now follow me step by step.

- The voltage over one of your LEDs is 3.5 volt at the max. current of 20 mA.
- If you place (just) 2 LEDs in series, the voltage over the both LEDs will be 7 volt.
- Add 1 volt for the current balancing resistor (47 ohm for 20 mA, thats correct),
- that will give 8 volts for the cluster chain.
- Place that on top of the 6 volts of the collector,
- and you need 14 volts minimum for the LED power supply, not 12 volts.!


Firstly, in that original diagram the 6V was calculated with the assumption that the LEDs used 2V each, and three in series were 6V. With a 12V supply, then 6V remained, and that's what's written at the base of the LED matrix... "6V 125mA" (because, 5 strings each 25 mA).

In my current case, the LED voltage is measured at 3.36 (let's say 3.4V) and therefore 3x = 10.2V running at 20mA each string. 10 strings = 200 mA. The "remainder" would be 1.8V and for 200mA that's a resistor of 1.8/0.2 = 9 ohms. In previous calculations the balancing resistors were as far as I'm aware *ignored* for the purposes of calculating the limiting resistor value. However, that looks wrong to me. The total resistance in this current case should be 9 ohms. Note also that I'm not catering for balancing resistors, because at the moment I can't understand why they are required - given the LED strings are pretty much consistent in resistance - in my view possibly at least as consistent as the difference between resistors.

So my first point of not understanding is "place that on top of the 6 volts of the collector". I don't understand this at all. The "collector voltage" in came from the calculation, and in this case it's 1.8V. I'm confused as to how I'm supposed to calculate a collector voltage which isn't anything other than the "remainder" after LED voltage drops (and balancing resistors if present) are taken into account. I had a look at the TIP122 data sheet, and couldn't really see what I should be looking at/for. I was trying to find something like "what voltage difference is required for operation" but other than Vbe(on)=2.5V as a possible, nothing seemed like a likely candidate.

Currently I'm stuck on this: LED string requires 10.2V @ 20mA. That's less than 12V. So, what's the extra voltage requirements for the transistor and the limiting resistor?

Klaas Robers wrote:But is you use a laptop power supply of 19 volt, abundantly available from old laptops,
- You can place 10 parallel branches of 3 LEDs in series,
- with one balancing resistor in each branch,
- a total current of 200 mA
- 30 LEDs and 10 balancing resistors of 47 ohm.


Not against using a 19V power supply at all. Just want to make sure that we're talking the same language and I fill in the gaps in my understanding. With the above, you are using 3.5V so a string of 3 LEDs is 10.5V and the current is 20mA/string = 200mA. With 19V that leaves 8.5V left over an 8.5/0.2 = 42.5 ohms, not taking into account any balancing resistor. OK, 47 is a good "next one up" candidate, but I'm still hazy as to how/if the balancing resistor comes into this calculation.


Klaas Robers wrote:The advantage of a lower total current is that the transistor heats up less,
- for 600 mA it dissipates about 3 watt,
- for 300 mA it is about 2 watt,
- for 200 mA it is about 1 watt.
- In all cases the transistor needs some from of heat sink.


With the current array running at 22V or thereabouts it's 8 strings of 5, about 30mA = 240 mA and the transistor is barely warm. It doesn't have a heatsink. I'll keep the above in mind, but not quite understanding why a heat sink will be necessary if it's already cool enough with a display drawing more current.

Very frustrating that I am having trouble with this basic stuff. Appreciate your comments and honestly not trying to be stubborn and just trying to understand :)
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1508
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Re: LED Matrix - 12V

Postby Andrew Davie » Sun Jun 11, 2017 10:10 pm

Steve Anderson wrote:Are you using PWM to drive the LEDs, or an analogue arrangement? If using PWM then a switched constant-current source for each 'string' is ideal. I'll elaborate if PWM is your chosen mode..


I am using PWM to drive the LEDs. It seems to be functioning very nicely at the moment, so I'm interested to see what you're suggesting and how it's an improvement. I like that I am using very few components at the moment, and the basic simplicity of the design.
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1508
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Re: LED Matrix - 12V

Postby Steve Anderson » Sun Jun 11, 2017 11:01 pm

Well. if you'r happy with the current arrangement, leave it be. My suggestion would probably have been more complex with no real advantage. It's just the way we all do things in a different manner. The end result would have been the same.

So returning to your current arrangement as I understand it, you have three LEDs per 'string' which should drop 10.5V at the current you wish to drive them. On the assumption you're using a transistor (NPN) or a MOSFET (N-channel) to buffer the micro's output that means you need a resistor of 8.5V/25mA = 340R per chain/'string' for a 19V supply, 330R is near enough. You do not need any other resistors at all.

As you're only switching a small current, 200mA or so, you could use almost any NPN tranny to do the job. A 2N2222, a BC547, endless.

If you're sticking with a 12V supply, in theory the resistor per chain comes down to 1.5V (less a liitle bit for the transistor/MOSFET saturation voltage) divided by 25mA = 60R, say 56R in practice, though I don't recommend this, any small deviation in supply volts could overdrive the LEDs, plus any variation in the LEDs forward voltage could have a similar effect.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 3800
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: LED Matrix - 12V

Postby Andrew Davie » Sun Jun 11, 2017 11:52 pm

I decided to wire up a matrix and see if it worked. I have a 10 strip matrix, each strip 3 LEDs. The strips are wired in parallel. There are no balancing resistors. There is a single limiting resistor of 10 ohms. The input voltage is 11.4V - a bit lower than what I expect my final version to be. So the video shows all LEDs bright and shiny, and to the eye there's no apparent difference. Hard to confirm via video - but they're bright and steady. I measured the "remaining" voltage at the power-in, and that was a small but consistent tad over 50 mV. The LEDs dimmed when I started up the motor - indicative I think that there's not enough current for both, with my current power input.

But, basically, this is the matrix as I thought it should work - without the transistor of course, but shining brightly with 12V input. I'm throwing this up to make clear what I'm suggesting, and to help others explain to me what I'm missing with my understanding. So, have at it!


youtu.be/a24D0-xzvd8
User avatar
Andrew Davie
"Gomez!", "Oh Morticia."
 
Posts: 1508
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Next

Return to Andrew Davie's Arduino Televisor

Who is online

Users browsing this forum: No registered users and 1 guest

cron