LED prop project for a Dr Who double

Discussion in 'The Projects Forum' started by Fenris, Jul 14, 2009.

  1. Fenris

    Thread Starter Active Member

    Oct 21, 2007
    288
    2
    Hi all

    OK heres the thing. I have a mate who does gigs as the current Dr Who. He needed a Prop and we had arranged for me to do the circuits and a third party to do the replica prop case for my kit to go in. Unfortunately the third party is unable to complete their part so it has fallen to me to do both parts!

    A compromise has been reached due to the limited time span to etch the PCB's and build the thing. The Case is a part of a hamster cage :D Basically a transparent plastic ball with for 50mm dia' ports at each of the cardinal points round the equator and it splits in 2 which makes it easy to fit everything!

    Orac was a term used for how he wants it to look, Google Blake's 7 for that one, and to that end I have drawn up Bill Marsden's from 4 - 20 circuit as a PCB. The only other spec is that all the LED's have to be BLUE. This runs with the theme of Dr who gadgetry in the new series.

    To this end the resistor values for the circuit needed changing to suit an all blue set up. Bill Kindly assisted, held my hand is more like it, me with how to work this out as, I have mentioned many times, my theory sucks!

    A resistor of the value 150 or 180Ω has been calculated. This takes into account the Vf of the 555 chip and of the blue LED's I have chosen. The 2 values represent the 'typical' and 'max' Vf of the LED as per the spec sheet.

    So that design is pretty well sorted, except I found out this morning when I was checking out the dimensions of the ball that I had added 10mm to the circular PCB I'd done!!!!! So As it is so remarkably tight as is now I intend to do a 2 part PCB one with the 8 555 timer circuits and their minor extra components and a second that will hold the 40 LED's and their resistors. I should mention I am doing 2 of the circuits side by side. Other than that I think it's pretty much sorted.

    Now That circuit will show in the top half of the ball. Another circuit will show through the bottom half. This one is a pair of 4017 chips driven by a 555 timer circuit which will have a pot to change the speed. The 4017's will be joined such that the second 4017 will advance 1 LED for every 10 of the first. A 2 part design is being used again. The first will contain the 3 chips and the 1-10 count LED's with their driver transistors and resistors.

    A second PCB will stack on top and this will contain the 10-100 count LED's their transistors and resistors. This design has been pulled from 2 sources. One that shows how to connect for the binary count and the other 4017 circuit showed how to connect driver transistors to the 4017. So I am using the binary type circuit with the driver transistors to be able to handle the blue LED's. I have calculated that the nearest standard value resistor is 220R to act as the current resistor. The transistors will be 2N3904 NPN's.

    I am assuming I have covered the bases with this one. But could someone just check over the pic which shows my intentions and give me a yay or nay.

    I should add that the voltage will be 9V.
    The spec of the LED's is-
    Vf 3.6V typical 4.2Max
    If 25mA

    I have just found that the above LED's are going to cost an arm and a leg. Actually I blacked out when I worked it out.

    So I have found a bulk buy of 100 LEDs for 13 pounds sterling. They are rated at
    Vf 2.8min 3.3typical 3.8max
    If 30mA

    For the from 4 20 circuit the current resistor will have to be 150R and for the 4017 counter circuit the nearest standard value to 190R is required. Am I still on track with this?

    regards

    Fenris
     
    Last edited: Jul 14, 2009
  2. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    You lost me on several points, I just got home from a hard day at work, so it's not surprising.

    Something to consider, you don't have to drive the LEDs to their max, they will still be very bright at half the current, so don't fixate too much on driving them hard. You will also drive the 555's a lot less. With the From Four, Twenty, half the LEDs will be on at any one time, it is inherent in the design, but at times a single 555 can carry almost the full load.

    Millwood showed a variation of the concept on another thread (Good Old Fading LEDs) that might apply, if the frequencies are fast (say 30-60Hz) and very close to each other (within a Hz) you could get some interesting phasing display's, think plasma (not sure, but I think that is how it would look). If you put pots where the feed back resistors go for the 555 oscillators it would allow you to play with the concept. Do you have a freq counter? Many DVMs have them built in, good enough for this application.

    This might be something to use for a future effect. I've always found LED circuits fun, instant gratification if you will.

    You going to make multiples of those PCB boards?
     
  3. Fenris

    Thread Starter Active Member

    Oct 21, 2007
    288
    2
    Hi Bill

    Sorry about that. I think I'm starting to panic a bit. It's not a lot to do really but it's the working out the maths and my inability to process the info myself that slows me down once I get all the theory sorted I have to order the parts and make the PCB's. Building the circuits will be a breeze.

    As it stands I intend to build your 4 20 circuit with the Blue LED's twice. So i'll have 2 lots of 20 LED's doing their thing. It is now going to be a 2 piece PCB for that part. One for the chips the other for the LED's. The LED's I was going to get are way to expensive so I have chosen another variety with slightly different specs. I'm my appear fixated because I lack the understanding which, I guess, is why I cling to specifics I can't see/understand what lee way I may have.

    The new circuit is in addition to the first and again will be in 2 parts. The IC's and the LED's that will 'chase' continually. On the other PCB of this circuit the LED's will also continually 'chase' but at a 10th of the rate of the main circuit.

    I have considered changing the relevant resistors to pots but I fear I lack time to experiment too freely. If any one was interested I would make available the PCB's if thats what you mean or post the patterns here. Subject to me making it work of course :D. I would also probably make one myself as it is as you say 'an instant gratification' type thing.

    I know the 2N3904 can carry 200mA so there is massive clearance to drive the LED's. So in point of fact in this one tiny case I may be aware of how broad a tolerance the circuit has. :D

    The colour picture- on the left is the double 4 20 of your design. I made a mistake and the PCB is to big for the case it is to go in which is way it will be changed into a 2 PCB circuit as it's the only way I will be able to fit the lot in.

    The 2 PCB's on the right are the 4017 cascaded circuits. The larger PCB had the chips and the 10 Chaser LED's the smaller PCB will go on top and has the LED's that go at a 10th of the speed of the main board.

    regards

    Fenris
     
  4. millwood

    Guest

    it is hard to see from the pcbs how many leds each transistor is driving but if each transistor is driving two LEDs, the resistor value would be roughly = (9v-1v-3.3v*2)/30ma=47ohm.

    or if you are driving one LED: = (9v-1v-3.3v*1)/30ma=160ohm

    the trick here is to make sure that if your LEDs come in at the minimum forward voltage drop, it doesn't see excessive current: (9v-1v-2.8*2)/47=51ma, or about 70% higher than its rated current. I typically make sure that it doesn't more than 1/3, or you risk burning off the LEDs fast.

    so driving two LEDs in serial wouldn't be a great idea here.

    for one LED, the current would be = (9v-1v-2.8v)/160=33ma, or 10% higher than its rated current of 30ma -> it works.

    the 1v I subtraced is to account for Vce(sat) in the transistor, rail voltage fluctuations, etc.

    that kind of current rating is typically useless, particularly for linear applications. you will have to go through thermal derating to see if you are safe - you are likely are, but not for the current ratings.
     
  5. Fenris

    Thread Starter Active Member

    Oct 21, 2007
    288
    2
    Hi Millwood

    I think I'll go bang my head on a brick wall! As you can tell I'm as thick as they come!! OK I get the 1V in the sum (voltage across the transistor) but I looked for that and couldn't see where to get it so I expected a correction to that. Something may have sparked in the final brain cell :D. For the 4017 circuit each transistor drives 1 LED. The first 4017 drives 10 LED's and @ carry's to the CLK of the second 4017 which drives the other 10 LED's which advance at a 10th of the rate of the first.

    As for the transistor being fit for purpose DAMN! Can I go out on a limb and suggest a 2N2222? Yes I am guessing......... or a BC547? This is the bit I have problems with. Correlating data sheet info such that I can go 'oh yes that will do the job'. I may as well read a blank sheet.

    I seriously admire you guys who can do this stuff I'm also green with envy. Could anyone tell me what NPN transistor, base resistor I need and resistors for the LED please.

    Regards

    Fenris (feeling slightly embarressed)
     
  6. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    How much is your LED costs? I paid 29¢ each for mine.

    If you have more than one PCB maybe you can use a different color to test it out? Green, perhaps?

    I know you're not in the States, so this is for reference.

    http://www.bgmicro.com/index.asp?PageAction=VIEWCATS&Category=115&Page=4

    My thought is you might be able to find something similar for reference.
     
  7. Fenris

    Thread Starter Active Member

    Oct 21, 2007
    288
    2
    Hi Bill

    Thanks for the link Bill I continued to search most of the day for a better price and reasonable spec.

    I have found these at rapidonline. £13.23 for 100 works out at 13p each ish. The spec is

    Vf 2.8min - 3.3typ - 3.8max
    30mA

    They are a diffuse blue with a 60Degree angle.

    I have redrawn the 4 20 PCB and managed to get it on a single PCB. I stood the resistors up :D The PCB is around 102mm in diameter.

    Have you any words of advice for what transistor I would need to use and what base resistor I would need to suit the second circuit? I got the idea for the transistor driver from your 4017 light chaser circuit which has 3 4017's driving a lot of red LED's via, what you describe, 2N2222A or similar transistors. :) I fancy building that one sometime!

    Here is a picture of the PCB's. The one on the left has been redone and subject to checking the layout are ready to print of for a bit of toner transfer fun. Please note that the circuits on the right are missing a lot of traces. This is because I am going to flywire from the output pins of the 4017's to the base resistors of each transistor. This will be done in a spiral loom to add to the effect of the unit.

    I have got a bag of 100 red LED's I could use but as I have found the blue ones a lot cheaper now I can save them for something else:)

    I am also finding that 3mm Blue LED's seem to be cheaper as well. I am using 5mm though.

    regards

    Fenris
     
    Last edited: Jul 14, 2009
  8. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    LEDs will keep dropping in prices, as new technology comes out. I literally have thousands in my closet, on reels. LEDs are one of those technologies that are almost impossible to predict. We have hints, but the breakthoughs keep coming out of nowhere. Quantum dots are the latest buzzword, seems some scientist received results he didn't expect when he excited a sample with a solid state laser, they glowed pearl white, extremely bright.

    OK, the 4017 has a max spec of 6.8ma (I think). I have driven a slightly different variation with 10ma for literally decades with no problem, but as I said in the article, I wouldn't recommend it professionally. I just went through the math again, I did screw up. The base resistors need to be 1.2KΩ to 1.5KΩ. The 470Ω will allow 17.7ma, which is not good, very bad. It appears I got those confused with the LED resistors, 17.7ma being approximently 20ma, a good value for LED currents. The 220Ω is a little low, but OK, it will allow 24 ma through the LED.

    Dagnabit, thanks for making me do the math again, I have some redrawing to do. In the words of Homer Simpson, DOH!

    Sorry.

    About the 2N2222, it is speced for a metal case style to carry 0.6A max. I would limit it to 300ma (.3A). If it is fully on it won't get hot, nor warm. The 220Ω resistors will get a little hot, but not very, as they will have 0.15W each. Since they are on for only a little while, it is actually a lot less than that.

    I would feel comfortable replacing the 2N2222 with almost any transistor out there, as long as it is listed as a small signal and switching transistor.

    The BC547 is considerably wimpier than a 2N2222A (this is a mil spec part), but it would work nicely for this application. It's max Ic is only 100ma, compared to the 2N2222 600ma.
     
    Last edited: Jul 15, 2009
  9. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You might consider using 2N7000 MOSFET transistors instead of bjt's like 2n3904s. The 2N7000s are available at Farnell Uk for around 7.6p each in a TO-92 package:
    http://uk.farnell.com/fairchild-semiconductor/2n7000-d26z/mosfet-n-to-92/dp/1467958

    They can sink up to 200mA current, have an Rds(on) of around 5 Ohms, and it would require very little power to charge/discharge the MOSFET gate compared to driving the base of a conventional bjt. This would help to prolong the life of the battery.

    Speaking of batteries, what are you planning on using to power the ball, and how long does it need to operate?
     
    Last edited: Jul 15, 2009
  10. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    9V battery, so power is an issue. Keep the base resistors for the gates? The package sounds a lot different, I've never used them myself. Could you link it please?

    I liked the Darlek voice, so I'm waiting to see where this goes.
     
  11. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Link added.
    The TO-92 package is the same as for PN2222, 2N3904, etc.
    It's best to use a resistor on the gate to avoid hf oscillations; anywhere from 10 to 220 Ohms would work. Once the gate is charged or discharged, there won't be any current flow. From the standpoint of using 4017s for driving a gate, that is a good thing. Gate charge is about 1.5nC, which is very low.
    [eta]
    To ensure staying within the absolute maximum ratings of the 4017, use 1k resistors from the 4017 output to the gate. It'll still switch in less than 0.2uS, even with a 4017 rise/fall time of 50nS.
     
    Last edited: Jul 15, 2009
  12. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    What I'm seeing, this part might be a drop in replacement, correct? Might need some twisting to get the leads in place, but there is even a chance it will fit as is.
     
  13. Fenris

    Thread Starter Active Member

    Oct 21, 2007
    288
    2
    Hi guys

    Thanks for all the input! I am glad that my misdirected wanderings have been of help Bill ;) The Power is indeed 9V. As such I am thinking 3 Battery's, one for each circuit. I am going to be cutting out some card board disks the same size as the PCB's to see how they fit. I am hoping that they will effectively sandwich the battery's (with insulation added!) As for run time.......... Well all day probably but It can be turned off at intervals I am sure.

    There will be an on off switch of course and also the pot for the 555 timer of the 4017 circuit these will be fitted as a 'tactile' part of the overall thing. I also have a halogen car bulb that will be doing duty as a 'valve' double. No it is not going to be powered :D

    The Mosfets sound an interesting Idea for saving on the battery power and they are a straight swap it seems :) I have revised the circuit diagram for it. Please note Transistor Symbol has been annotated with Mosfet pin out.

    EDIT; just reread Sgt post. I will use 1K resistors for the gate pin.

    Any other comments would, as always, be gratefully received. I think I will be ordering the parts next week to build all the circuits. The guy needs the finished item for the weekend of the 8th-9th of next month :eek:

    Just caught your post Bill :D It is a drop in replacement :D

    Added piccy of the ball which will contain the circuits. You can see the 3 cardboard PCB's. The bottom one, the LED's will all face down, will be the dual from 4 20 from Bills ebook and the upper 2 make up the 4017 circuit.
    I should mention that the ball is upside down so the 4 20 circuit is the top one with the LED's on top and the 4017 based circuit will be upside down but the LED's will be bent to point outwards.
    regards

    Fenris
     
    Last edited: Jul 15, 2009
  14. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Here is what happens to 9v alkaline batteries under load:
    http://www.powerstream.com/9V-Alkaline-tests.htm

    So, for each battery:
    1) How many LEDs will be lit at any one time?
    2) How much current do you really need to supply to each LED?

    If you need for this thing to run all day, you will have to limit total circuit current considerably. 9v "transistor" batteries will become exhausted very quickly even under light loads. With a 40mA constant draw, a fresh battery might last 5 hours or so before it's output drops to 7v.

    You should use CMOS 555 timers instead of BJT 555 timers. You should have 0.1uF caps across each IC's Vdd/earth pins.
     
  15. Fenris

    Thread Starter Active Member

    Oct 21, 2007
    288
    2
    Hi Sgt

    Thanks for the link. I certainly get the gist of the problem........ Theres always something out to thwart you :)

    Bills circuit, the 4 20, consists of 20 LED's only half will be on at any one time. I have 2 of these circuits on one PCB and I was reckoning on each having it's own battery.

    The 4017 circuit - The first 4017 will run continuously so only one LED will be on at a time at whatever speed the pot is set at. The second 4017 that is driven by the 'carry' pin of the first will have one LED on at any time as well but it will be on 10X longer before it moves on to the next LED. This circuit will also have it's own battery.

    If you can give me the specifics of lessening the load I'll be happy to follow them. The cap across the CMOS 555 chips power pins would be a non electrolytic type?

    regards

    Fenris
     
  16. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    OK, so 10 LEDs out of 20. If you're planning on running the LEDs at 20mA each, that's 200mA total. Expect a run-time of under 1 hour.

    OK, so two LEDs. If they're running at 25mA each, that's 50mA. You might actually get to 8 hours on this, using a fresh alkaline battery.

    If you can give me the specifics of lessening the load I'll be happy to follow them.[/QUOTE]
    Use less power, add more battery capacity, or both. You may have to resort to running a boost-type circuit. Bill, remember that EDN design for the inductive boost for the LEDs? I really don't have time to do it now.
    Right. You also need 0.1uF caps on the 4017 counter Vdd/ground pins.
     
  17. Fenris

    Thread Starter Active Member

    Oct 21, 2007
    288
    2
    Crikey! Less than an hour :eek: OK then. Drive the LED's at less current then? So adjusting the LED's current resistor in an upwards direction would help? As in the LED's being slightly less bright. Or is the resistor just soaking up the difference?

    These are the LEDs I intend to use.

    http://www.rapidonline.com/Electron...cs/5mm-LEDs/Standard-5mm-LED/61374/kw/55-1754

    I had vaguely thought of Caps across the power points. I will add these to the circuits!

    regards

    Fenris
     
  18. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Well, Bill's 4-20 circuit is going to have to stay with BJT 555's, unless you want to have to add N-ch and P-ch MOSFET drivers to CMOS versions; as the CMOS 555's won't handle the current required to drive the LEDs directly.

    Of course, your battery won't handle the current for long either... the LEDs will still light, but after a while they will be quite dim. At 60° diffused, 45 mcd to 68 mcd with 20mA current, they aren't going to be very bright to begin with.
     
  19. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    Lithium is expensive, but should last longer. I would also think about dropping the current to the LEDs. I've got to go to work, but I have some other questions/ideas.
     
  20. Fenris

    Thread Starter Active Member

    Oct 21, 2007
    288
    2
    Hi Guys

    Well A budget has been drafted..............£20 or less............ So it's going to be the 4017 Circuit only which is mainly down to it's battery life advantage :(

    However I am going to be building Bills 4 20 Circuit for myself using the red LED's I have :) . I rather like the design and it's flashing LED's!

    The core of the circuit will be 4 10 hole x 7 rows of vero board which will each have one of the 555 Circuits on board. I will also implement the variable resistor option whereby I can play with the timing.

    Using single core mains wire I will solder a piece at each corner of the top PCB, slip a piece of sheathing back on to act as a spacer and then slide the second 555 PCB in place and so on as the first. The cunning part of this plan is that the wire 'pillars' are lined up on the positive and negative rails of each circuit :)

    Initial tests will see the LED's and their current resistors set up on a bread board so I can see how they act. Once thats done I can set them up on a final PCB.

    I will continue to post on both items as I progress.

    Regards

    Fenris
     
Loading...