1. Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.
    Dismiss Notice

DS1302 (Parallax SX52) 16x16 LED matrix scrolling clock - additional circuitry??

Discussion in 'Microcontrollers' started by tdg8934, Apr 11, 2007.

  1. Mike - K8LH

    Mike - K8LH Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    3,637
    Likes:
    109
    Location:
    Michigan, USA
    Wilf updated that other forum thread to say his 1/32nd duty cycle design won't work because the 500-ma peak ULN2803 pins won't handle 2.5 amps peak. Now he's recommending his earlier 1/16th duty cycle design which scans the 16 anodes instead of the 32 cathodes and we're back to 2.5 amp P-FET anode drivers.

    May I ask what you paid for that 20-amp PSU?

    The more I think about it the more it seems that you may need a higher supply voltage for your 1/16th duty cycle display. For example, if we could somehow measure the peak voltage across one of those 10 ohm current limiting resistors and it was 50.0-mv then we'd know we have 5.0-ma peak current going through the LED. Now to get more current through the LED don't we need more voltage across the LED? But a 50.0 ma peak current going through that same LED and resistor would result in less voltage to the LED because the voltage drop across the resistor would increase from 50.0 mv to 500.0 mv.

    I think I need to apologize for mistakes and my misleading design suggestions. I remembered enough from my earlier experiments to recommend 1/8th or higher duty cycles but I don't think I realized why. Now it seems (to me) that you might need a higher supply voltage when using extremely low duty cycles.

    What do you think?

    BTW, I went back to one of my earlier posts and added a qualifier to that 1/16th duty cycle design which states that a higher supply voltage may be required on extremely low duty cycle designs. I just want to make sure that I don't mess up anyone else who may stumble across this thread and that design.
     
    Last edited: Apr 16, 2007
  2. tdg8934

    tdg8934 New Member

    Joined:
    Oct 15, 2006
    Messages:
    147
    Likes:
    0
    Location:
    Somerset, KY
    Mike,

    I replied to Wilfs new design this morning about the use of the 2.4 amp Fairchild FDN304PZ P channel MosFET since he did not state if the PNP 3A transistor could be used instead or not.

    I got a great deal on Ebay for this $155 new 5vdc 20A power supply (looks pretty big - but wont know till I get it) for a winning bid (and ONLY bid) of $10 + $12 shipping from a company outside of San Diego, CA (Chula Vista, CA). I was in the Navy for 10 years (I left in 1992) so I know San Diego well.

    Are you thinking that I will need a higher voltage power supply now than the 5vdc? If possible I would like to stick with that since all of the other components will need 5vdc (ULN2803, SX52, etc..). Let me know. Agh... If the 5vdc 20 Amp supply has 20 Amps available why can't it be used? It sounds like he explains why it CAN be used. I'm not clear on this.

    Remember I have 3 LED Matrixes. The first is a 7x20 that uses the same 10mA LEDs as my 2nd 16x16 LED Matrix. This 7x20 uses PN200 higher current (500mA) PNP transistors and a 10 ohm resistor to the each of the 7 rows (but across the 20 columns - that are then fed into {3} ULN 2003s). I have always used my 1 AMP 7.5 vdc p/s into a 1 AMP regulated 5vdc development board for this and got the same "ok" display brightness as my 2nd 16x16 LED matrix (which has 2N3906 PNP standard current transistors for the 16 rows and {2} ULN2803s for the 16 columns). The 3rd LED Matrix is under construction. This is the (4) 8x8 LED modules (which are bi-colored so it actually is (4) 8x16 - (2) 16 columns in red and (2) 16 columns in green LEDs wired to (4) ULN2803s and the anodes are the rows. The 32 columns are currently wired to These rows will need to go to either P-FETs or some high current PNP transistors using the new 5vdc 20 Amp power supply to arrive this week hopefully.

    If you would like I could replace all of the 100 ohm resistors in the 16x16 LED matrix with 10 ohm resistors (from the 100 ohm / 1K 2N3906 circuits) which would make it a little brighter. This could be then measured if you would think it would help. OR I could measure the voltage across the original 7x20 display that has the 10 ohm resistors, 1K and the PN200 high current resistors. I'm not sure what you are getting at here. Please clarify what needs to be done.
     
    Last edited: Apr 16, 2007
  3. Mike - K8LH

    Mike - K8LH Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    3,637
    Likes:
    109
    Location:
    Michigan, USA
    I read your post on the other forum. That MOSFET Rds spec is 'drain source resistance' measured in milliohms. You'll be using TTL signal levels on the gate which is Vgs = 4.5v so your Rds should be something like 52 milliohms which is well below the 100 milliohms Wilf was talking about. You should have very little voltage drop across the P-FET when it's 'on'.

    As for your first two matrices -- it sounds like the 7x20 is scanning columns of 7 LEDs because the current limiting resistors are on the rows. That's another extremely low 1/20th or 5% duty cycle which may explain the just 'ok' brightness level. Or, it may just be that the matrix is being driven directly by microcontroller pins which cannot sink or source the 'peak' current required for that low duty cycle. Are there any transistor drivers on that 7x20 matrix?

    Honestly, I'm beginning think you need a higher supply voltage on the anodes of these extremely low duty cycle matrices.

    Anyone else here on the Forum care to pass along some your wisdom and experience, please?
     
    Last edited: Apr 16, 2007
  4. dave

    Dave New Member

    Joined:
    Jan 12, 1997
    Messages:
    -
    Likes:
    0


     
  5. tdg8934

    tdg8934 New Member

    Joined:
    Oct 15, 2006
    Messages:
    147
    Likes:
    0
    Location:
    Somerset, KY

    Mike,

    So it sounds like (at least from your end - I will wait to hear back from Wilf about using the 3A PNP transistors or not - that I should use the Fairchild 2.4 A FDN304PZ part {16 of them})

    The first 7x20 has 7 rows of 20 LEDs (left to right) and 20 columns of 7 LEDs (up and down). I get confused as maybe I am getting rows and columns mixed up as I understand it. The 20 columns (as I see it) are connected to (3) ULN2003s and are then driven by some 4017 and 4081 chips (so no interupt is needed in software). The scanning is done in 4017/4081/ULN2003 hardware (not in a software interupt). The 7 rows are whats brought into the SX28 microcontroller as RB or RC or something. This sends out 1 byte of character data that the 4017/4081 scrolls from left to right. I have attached the program and drawings for you to view.

    If you do beleive I need a higher voltage supply with higher current (e.g. 5A or 10A) then what do you recommend the voltage should be so that I can start looking on Ebay again (e.g. 12 VDC ?).
     

    Attached Files:

  6. Mike - K8LH

    Mike - K8LH Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    3,637
    Likes:
    109
    Location:
    Michigan, USA
    Ok, that 7x20 design has drivers on columns and rows. You say this matrix has just 'ok' brightness?

    What new supply voltage? Gosh, I don't know. Again, my projects had higher duty cycles and so they worked fine with 5 volt supplies.

    Aren't there some good variable voltage 3 amp or 5 amp regulators around that could be fed with 12 volts and provide a variable 5 to 9 volt output for the display?
     
  7. tdg8934

    tdg8934 New Member

    Joined:
    Oct 15, 2006
    Messages:
    147
    Likes:
    0
    Location:
    Somerset, KY
    Mike,

    When I saw it has ok brightness, it refers to that in a normal lit room the message is barely readible but if the lights are off, it looks pretty good. I have a 7 LED arm swinging "floating display" clock that is bright in light or darkness AS WELL AS the MK123 scrolling clock I mentioned earlier that looks good in normal light and low light settings. This is what made me think it used new high rated MCD LEDs.

    All of the supply voltages have been powered from my SX protoboards power which is from a 7.5 vdc p/s 1A into the boards 5vdc regulator.

    If we go with this approach of a new higher voltage supply, then the logic chips should be powered from the SX protoboard 5vdc power and ONLY the anode FETS would receive Vcc (12v, 13.8v, etc.. at 3 A, 5 A etc...) and the ground would be common to both. However, I also think as Wilf suggested, the use of 2200 uF decoupling caps and 0.1 uF too by each chips (Vcc / Vdd / Vss) and Gnd.

    You also state that this new power supply would provide a variable 5 - 9 v for the display. I thought that you said that 5 v is not enough. A bit confusing.
     
  8. Mike - K8LH

    Mike - K8LH Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    3,637
    Likes:
    109
    Location:
    Michigan, USA
    Sorry for the confusion which seems to be cumulative (grin).

    When I was wondering if there were any inexpensive power 'bricks' on ebay I should have said Laptop power bricks which are usually in the 12 to 16 volt range and 3 to 6 amps or so. I use those type 'bricks' along with linear regulators and filters on my project boards. I was surprised when you said you had purchased a 5v 20a supply.

    Yes, Wilf's suggestions for filter and spike capacitors are "standard practice".

    Yes, all IC's should be connected to 5v VDD and the anode drivers to the higher voltage supply.

    I mentioned the 5v to 9v range because that's the range I seem to recall those variable linear regulators covered.

    Now here's where we need some expert advice. It's my understanding that you can't feed the gate of a P-channel MOSFET directly from a TTL level microcontroller pin while the Source pin is at some higher voltage level. Much like the VPP switching circuit in a microcontroller programmer (below) that switches the 12v VPP voltage, you need another transistor. Is this right guys? Is this because you don't want to see that higher voltage (through the pull-up resistor) appear on the microcontroller pin unless that pin happens to be an open collector or open drain output?

    [​IMG]
     

    Attached Files:

    Last edited: Apr 17, 2007
  9. tdg8934

    tdg8934 New Member

    Joined:
    Oct 15, 2006
    Messages:
    147
    Likes:
    0
    Location:
    Somerset, KY
    I have been looking on Ebay at bit this morning to find a good deal like the 5vdc 20A supply I bought last night. All of the variable supplies are just too much money. Is there anyway we can use the current 5vdc 20A? If not I will continue to look. What is the recommended good current rating (e.g. 3 A + or 5 A + or 10 A +, etc..). Most of the higher current p/s seem to be for CB radio's and communication equipment. These are at 12 vdc or 13.8 vdc.

    Is this what I should be looking for (IF we absolutely can not use my 5vdc 20A new supply)?



    I just posted at the same time you did. ARE Q3 and Q4 the same P channel FET (2907A)? The arrows are different in the symbol.
     
    Last edited: Apr 16, 2007
  10. Mike - K8LH

    Mike - K8LH Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    3,637
    Likes:
    109
    Location:
    Michigan, USA
    No, please don't buy a big honkin' CB power supply or one of those expensive variable supplies (grin). I'd still try to find a good cheap 12-15v 3-5a Laptop power 'brick'.

    I just ordered samples of the National LM350 adjustable 3-amp linear regulator in a TO-220 package. I'll throw it and a regular old 5v 1.5a TO-220 regulator plus filters, etc., on a radio shack protoboard to play with next weekend. You might consider doing the same.

    <added>

    No, Q3 is a 2N7000 N-channel MOSFET and Q4 on this particular programmer is a BS250P P-channel MOSFET. The little arrow on that center line in the symbol tells you if it's N-channel or P-channel. Both are relatively low current and come in TO-92 packages. I used this circuit on my first PIC programmer design (below).
     

    Attached Files:

    Last edited: Apr 16, 2007
  11. tdg8934

    tdg8934 New Member

    Joined:
    Oct 15, 2006
    Messages:
    147
    Likes:
    0
    Location:
    Somerset, KY
    OK I will look for a 12v P/S like you sugested. Are you stating 3A as a minimum because the total current of the board would be 2.56 amps therefore the 3A power supply would meet the current requirements?

    You also mention getting a LM350 voltage regulator. Is this just to make sure the "brick" supply is regulated to a certain desired voltage? I'm not sure in it's purpose.

    If I get a 5 A power supply then I would also need to get a 5A rated adjustable regulator too then?
     
  12. Mike - K8LH

    Mike - K8LH Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    3,637
    Likes:
    109
    Location:
    Michigan, USA
    Yes, I'm thinking a 3 amp 'brick' would be good for a display with 10-ma "average" current for each LED (2.5 amps total "peak" current). That should be magnitudes brighter than what you're getting now. And if you should happen to find a 5 or 6 amp supply the extra capacity won't hurt.

    The LM350 adjustable linear regulator would be used to power the display portion of the project with the 12 to 15 volt power 'brick' as the input to the regulator.

    If you get a 5 amp supply and you want a 5 amp variable regulator then perhaps you should look at the adjustable version of the National LM1084 series (or other manufacturers) regulators.
     
  13. tdg8934

    tdg8934 New Member

    Joined:
    Oct 15, 2006
    Messages:
    147
    Likes:
    0
    Location:
    Somerset, KY
    OK. I understand a bit more now.

    Should I use the SX52 protoboard 5vdc 1 A power for the SX52, all the logic chips, etc..? Does it make sense to also get a 5vdc regulator like you mention to power the logic chips? Wilf made it sound like in him last posting that the 5vdc 20A should not be used for the logic chips - that is why I am asking (but then this is much less than 20 A {being 3 A or 5 A})

    I like the idea of getting as much current as possible - just in case I want to expand on this design in the future and need more current. I did have a hard time on Google locating any adjustable regulator more than 5 amps. I will let you know what I can find.

    How will this new power supply connect to the rest of the circuit? Is it only connected to the P-FETs along with a common ground and decoupling caps, etc.?

    OK - Now what is the final list of components I will need to buy.
     
  14. Mike - K8LH

    Mike - K8LH Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    3,637
    Likes:
    109
    Location:
    Michigan, USA
    Yes, that will isolate the two supplies and your logic circuitry should enjoy clean glitch free power.

    No, I only mentioned that I would be throwing a 5v regulator on my 'test' board while I experiment and play with that LM350.

    I agree, that's a good idea.

    They're out there, I just haven't had the opportunity to use them yet and so I can't make a recommendation. Sorry.

    Yes. The new supply will power the new adjustable regulator which will power the display through the anode switches. You'll adjust the pot' on the adjustable regulator PCB to set the regulator to the correct voltage for full brightness.

    Samples?

    I haven't found a source for sampling ULN2803's or MIC5821's but I did get some Allegro A6821 samples without any trouble last year (the A6821 is a pin-for-pin compatible MIC5821).

    I ordered samples of the adjustable versions of the 3-amp LM350's (TO-220) and the 5-amp LM1084's (TO-220) from National a short while ago without any trouble.

    I ordered and received Fairchild FDN304PZ samples about a year ago.
     
    Last edited: Apr 16, 2007
  15. Mike - K8LH

    Mike - K8LH Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    3,637
    Likes:
    109
    Location:
    Michigan, USA
    Also please note that Murphy's Law suggests we'll find the flaws in the current design that will require changes to the parts list or bill of materials immediately after you order and pay for those parts (grin).
     
  16. tdg8934

    tdg8934 New Member

    Joined:
    Oct 15, 2006
    Messages:
    147
    Likes:
    0
    Location:
    Somerset, KY
    Did you know about:

    www.findchips.com

    It will list all the vendors / distributors for almost any part you need. This is how I knew that Digikey sold the MIC5821.

    Do you know if I can get free samples of these? If not it is ok but just thought I would ask.
     
  17. Mike - K8LH

    Mike - K8LH Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    3,637
    Likes:
    109
    Location:
    Michigan, USA
    I already said two messages ago that I had not found a source for sampling MIC5821's but I was able to sample the Allegro version of the MIC5821 (the A6821) from Allegro.
     
    Last edited: Apr 16, 2007
  18. tdg8934

    tdg8934 New Member

    Joined:
    Oct 15, 2006
    Messages:
    147
    Likes:
    0
    Location:
    Somerset, KY
    Thanks Mike. I wasn't clear if you meant the samples were free or not. I just ordered 6 free samples of the Allegro A6821SA-T. I have seen some vendors that have Samples but they are not free or at a partial shipping cost or something.
     
  19. Mike - K8LH

    Mike - K8LH Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    3,637
    Likes:
    109
    Location:
    Michigan, USA
    Yep, you're right. I just got the order confirmations from National for the regulator samples and they wanted $18.00 so I cancelled the orders.

    Bummer!
     
  20. tdg8934

    tdg8934 New Member

    Joined:
    Oct 15, 2006
    Messages:
    147
    Likes:
    0
    Location:
    Somerset, KY
    I just ordered 10 free samples (max) of the Fairchild (2.4 A) FDN304PZ. I need to find another source to get 10 more. I will look.
     
  21. Mike - K8LH

    Mike - K8LH Well-Known Member

    Joined:
    Jan 22, 2005
    Messages:
    3,637
    Likes:
    109
    Location:
    Michigan, USA
    Wait a few days and order again (grin).

    BTW, I also received samples of the following parts from Fairchild long ago;

    PN2907A (TO-92)
    2N4401 (TO-92)
    2N4403 (TO-92)

    LM2931A, 5v, 100ma, LDO regulator (TO-92)
     
    Last edited: Apr 16, 2007

Share This Page