Announcement

Collapse
No announcement yet.

Easson ES-8 lost led segments help

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    The seat of the problem is going to depend on how the units are set up.

    The most likely arrangement for a single multi-digit device, is for there to be a transistor for each digit. That turns on the digit, and the processor then turns on whatever segments are needed to be on for that digit. Then the processor turns on the next digit, and turns on the correct segments for it, etc, etc.. For that type, the common point is the transistor that turns on the digit. If it does not work, the digit will not turn on.

    Cause can be a bad transistor, OR an open base drive coming from the processor to the transistor, typically a resistor that is not soldered, or has opened for some reason. Can be the pin on the processor having a bad solder joint, or even the processor port having a bad pin.

    When several essentially identical displays exist, they can all be driven by a single transistor, or from a single port pin, in which case a problem there can affect the same digit on all displays.

    Since your unit has a bad digit on every display, the suspicion is on the drive transistor, or the transistor base drive from the processor.

    Best scenario is a single transistor for all, because a fault there can affect all. You would look at that transistor and its PWB traces to find the problem

    It gets more involved if each display has its own digit drivers, because then the common point moves back toward the processor, which hopefully has a bad solder joint, or an open PWB trace. An actual bad port would pretty much mean replacing the unit.
    Last edited by J Tiers; 10-20-2016, 08:34 PM.
    CNC machines only go through the motions

    Comment


    • #17
      0.6 volts is the expected voltage drop across a diode. If something is drawing too much power then I would expect that voltage which isn't enough to light the digit displays. It depends a lot on how this voltage was measured. If jumping the connection to one of the good sets of digits results in dimming of any other digits at all that would indicate too much power draw due to some sort of low resistance defect somewhere pulling too much current. If there is no dimming or significant voltage drop when it is jumped then it is most likely some sort of open or poor turn-on of the transistor. In that case just jumping power should be a reasonable work-around since any good transistor should be able to handle that much more load. It would be a very poor design that didn't provide significant current headroom on the control transistor. I am making a lot of assumptions on how it is wired so I may be wrong.
      Free software for calculating bolt circles and similar: Click Here

      Comment


      • #18
        That line of diodes is more likely a set of steering diodes. A common way of driving multiple segments was to use a line of such one way devices so only the things down stream of the steering diodes see the power.

        With a multiplexed situation like this where the segments are being strobed in sequence it would be natural to read a lower "average" value of some sort since some segments will be getting a blip of power and others won't. It's promising that bridging that joint produces light on the dark digits.

        I'd get one another switching diode that is the same or better than what you have there on the board and try bridging D15 with the new diode. If the segments light up then it's the board's diode that is bad. If not then it's something else.
        Chilliwack BC, Canada

        Comment


        • #19
          It looks like the anodes of the diodes are connected to the ground plane through thermal spokes at the top of the board. The cathodes are connected to pins of the display. It is not clear what these diodes are supposed to do. If the display is common anode, those cathodes would be driven high with a 1/8 duty cycle, so if it uses 5 VDC, the average voltage would be 5/8 or 0.625 volts. If it is a common cathode display, these would be pulled low with a 1/8 duty cycle. The voltage during the OFF times might not be well defined as the lines would be floating with an "open collector" or "open drain" circuit.

          If the digits light up when you connect the cathode of the suspect diode to ground (by shorting it), then that confirms that it is common cathode.

          Here are some descriptions and circuit diagrams of multiplexed LED displays, which may help with understanding and troubleshooting:

          http://saroselectronics.com/multiplexed-led-display/ (3 digit common cathode)



          http://embedded-lab.com/blog/lab-11-...-led-displays/ (4 digit common anode)



          http://sentex.ca/~mec1995/tutorial/7seg/7seg.html (general information)

          https://tams.informatik.uni-hamburg....-display8.html (may be helpful)
          http://pauleschoen.com/pix/PM08_P76_P54.png
          Paul , P S Technology, Inc. and MrTibbs
          USA Maryland 21030

          Comment


          • #20
            A clear pic of the ENTIRE PWB might be helpful in deciding exactly how it is set up.

            But the system is generally as PSTP's pics show. The question is if there is one set of transistors for all, or if each display has its own. That matters as to where the problem may be.
            CNC machines only go through the motions

            Comment


            • #21
              Originally posted by J Tiers View Post
              A clear pic of the ENTIRE PWB
              To achieve this, select the macro (flower icon?) setting for the camera. Position the board in an area with LOTS of ambiant light with object of enabling shutter speed to be above 1/100th when f: is 3 or higher. If a tripod is available, use it and set the camera timer so there is no wiggle from manual shutter activation.

              .

              Comment


              • #22
                I also did this for a living for many years. The fact that jumping D15 causes the dead displays to come on is a strong lead. I would give a 95% probability that the problem is with D15 or TR6. Both look like they are soldered properly. I had a rule while troubleshooting things like this where I was not sure of the exact problem. After an hour of checking and measuring and head scratching, I took my best guess and replaced it. That, at least, eliminated one possibility.

                In this case, I would start with the diode because it is a through hole mount and probably easier to replace. It is probably a $0.05 part and even Radio Shack would not charge over $0.50 for it. If it is not marked, a 1N914 or 1N4148 will probably be OK. Be sure to install the new diode in the same polarity as the original one. If the diode tests OK after removing it from the board then go for TR6. That may be more problematic finding a replacement if it is not marked, but tell us the supply Voltages and I can probably give you a good generic replacement. Again, probably a $0.25 part.

                Changing these two parts will be faster and with your time included, less expensive than all the head scratching you and all of us can possibly do.



                Originally posted by bob_s View Post
                Sorry!
                Yeah, yer blind uncle.
                It looks like diodes should go with the transistor to the RIGHT.
                So TR6 is the suspect device. I can't make out any writing smaller than the printing on the circuit board.

                so if 15 ma/seg *7 segs * 3 devices = 315 ma

                If there's enough clearance above the board you may be able to get a regular leaded device to replace the surface mount device.
                Paul A.
                SE Texas

                And if you look REAL close at an analog signal,
                You will find that it has discrete steps.

                Comment


                • #23
                  PS: On checking that diode after it is removed from the circuit, you only need an Ohm meter which has a 9 Volt battery. A single 1.5 V battery may not provide enough forward Voltage to turn a diode on. Using a mid range resistance scale, you check the diode with the meter leads attached to it in both polarities. One way should read low resistance and the other way should read higher resistance. Low = one end of the scale and high = the other end.

                  For a diode, you only need to disconnect one of the two leads from the board for it to be detached from the circuit. A trick that many troubleshooters use with components with accessible leads is to just cut one lead in it's middle. If it tests bad they then remove both leads from the board and replace it. If it tests good, they just solder the broken lead back together. Not the neatest way, but it is fast.
                  Last edited by Paul Alciatore; 10-21-2016, 03:39 PM.
                  Paul A.
                  SE Texas

                  And if you look REAL close at an analog signal,
                  You will find that it has discrete steps.

                  Comment


                  • #24
                    My cheap (free) Harbor Freight DMM has a diode test function that reads the forward conduction voltage with a current of a few mA. Silicon diodes read about 0.6 to 0.7 volts, and Schottkys read about 0.2 to 0.4 volts. It will even light an LED, dimly, including some white LEDs. You can usually check diodes in-circuit unless they are connected to low resistance or transformer windings, which is unlikely in this case.

                    The diode check can also be used to test transistors. For a bipolar junction transistor (BJT), an NPN will show a diode drop with (+) connected to the base and (-) to either collector or emitter. PNP is the opposite. There should be no conduction from collector to emitter. For a MOSFET, there will be a parasitic diode from drain to source. An N-channel will conduct with drain (-) and source (+). There will be no conduction from the gate to drain or source.

                    You can further test it (for NMOS) by touching the ohmmeter probes to the device with (+) on the gate and (-) on the source, then read the resistance from drain to source. The gate capacitance will hold enough charge for a few seconds to keep the device in a conductive state.

                    Generally, I have found that most problems are caused by bad connections, such as solder joints, crimp connectors, and IDC ribbon connectors, especially when the problem is intermittent. Other problems can be contamination on the surface of the PCB, from soldering flux or ionic substances (salt), that become conductive with humidity. I have also found problems on PCBs where tracks have an invisible hairline crack under the solder mask, or vias that are incompletely plated through. These defects can be detected by reading continuity between points, and then narrowing down by scratching through the mask. Tracks can be fixed with a piece of fine wire, and vias can be fixed by poking a component lead or wire through the hole and soldering both sides.
                    http://pauleschoen.com/pix/PM08_P76_P54.png
                    Paul , P S Technology, Inc. and MrTibbs
                    USA Maryland 21030

                    Comment


                    • #25
                      I found a video showing a partial tear-down of the ES-8:



                      He points out some 74HC373 ICs that are used for the displays. They are octal D-type transparent latches, likely used to hold the segment information for each display. His unit is a bit different, having thru-hole transistors and some SIP resistor networks, so probably an earlier version.
                      http://pauleschoen.com/pix/PM08_P76_P54.png
                      Paul , P S Technology, Inc. and MrTibbs
                      USA Maryland 21030

                      Comment


                      • #26
                        The first place to check for bad connections is any place on the board with larger wires or device leads installed or connected. They don't warm as fast when soldered and will often have microscopic cracking around the connection. It often isn't visible to the naked eye. Easiest is to go around and resolder those connections when possible and do it on both sides when you can.
                        Free software for calculating bolt circles and similar: Click Here

                        Comment


                        • #27
                          He said that if he jumps D14 to D15 (not something I would have recommended) that the bad displays do come on. I think that would rule out (99%) a bad connection. I still think his problem is with D15 or TR6. If not there, then perhaps the signal from the IC that is feeding them is bad.

                          He did not say which end of the diodes he jumped together. Cathode/banded end or anode? That info. would also be a help.



                          Originally posted by Evan View Post
                          The first place to check for bad connections is any place on the board with larger wires or device leads installed or connected. They don't warm as fast when soldered and will often have microscopic cracking around the connection. It often isn't visible to the naked eye. Easiest is to go around and resolder those connections when possible and do it on both sides when you can.
                          Paul A.
                          SE Texas

                          And if you look REAL close at an analog signal,
                          You will find that it has discrete steps.

                          Comment


                          • #28
                            With all due respect.....

                            I dunno how anyone can deduce the purpose of the diode until we get a better idea of how the thing is set up. I can think of several possible uses and locations for the diode, generally related to preventing spurious illum ination by back-currents into unlit segments.

                            We need the pic, or other info as to how many digit drivers are present. The latches will hold the segment data, but here we have all 3 displays dead at once. Obviously the segment data is not the issue.

                            The one in the video, if you look at about 6:31, 6:54, and later, appears to have one set of digit drivers. Might even have the diodes, it is never quite in focus in a closeup. If (could be a big "if") it is the same circuit as the OP's, then his has one set also. His could just be a surface-mount version of the same circuit.

                            That suggests looking at whether the transistor for that digit is good, and whether it is getting a proper signal to turn it "on" when wanted. The diode may be a red herring. If the base/gate of the transistor is getting digit drive, then the transistor is either bad, or something in the drain/collector circuit is not getting connected.

                            If base/gate drive is not present, then the processor itself is bad, or wiring from it to the transistor is open/shorted.
                            Last edited by J Tiers; 10-22-2016, 12:10 AM.
                            CNC machines only go through the motions

                            Comment


                            • #29
                              He said that if he jumps D14 to D15 (not something I would have recommended) that the bad displays do come on. I think that would rule out (99%) a bad connection.
                              No, it doesn't. That is why I asked about possible dimming. It there is an open somewhere then jumping power over shouldn't cause dimming. If there is a semi-short or low resistance failure then it is a different story.
                              Free software for calculating bolt circles and similar: Click Here

                              Comment


                              • #30
                                Willing to bet it is a bad transistor or lack of a signal getting to it. Sometimes there are marking on the transistor that can lead you to what it is. Really hard to read, usually.

                                From the pic the lower right pin on the transistor is the base. Measure the voltage here to ground, the end of the diode away from the black ring. It should be about the same across all of them. If there is no voltage you need to trace back, the base of the transistors are probably tied to one set of those resistors to the right which are probably pulling down the line from the IC that is driving the transistors to the right of the resistors. If there is no signal out of that IC, it could have a bad output.

                                Comment

                                Working...
                                X