| Mahmood Elnasser 12/25/09 03:18 Read: 1377 times Amman Jordan |
#171915 - LED Display Panel reliability |
Hi All
I made and LED panel that gets data through a GPRS modem once every 10 minutes, then writes this data to the panel. Hardware wise I made every precaution to avoid noise like having ferrite beads on all wires coming in and out of the boards including power supply lines and 2 separate power supplies, a linear regulator for the drivers and control with a separate SMPS for the LEDs themselves. Data is being sent to each display panel block using RS-485 and used solid ground plane for all the boards. since the data goes to the constant current led drivers serially. is there away to know that the data was clocked correctly into the registers? I had one thought to clock the data into shift registers 5 times then read back the data, if the last 4 times data was correct then there is a good probability that the fifth one is correct, then accept it, otherwise clock the data five times more. But this approach doesn't guarantee that the displayed data is correct. Another thought since the data arrives once every 10 minutes to be displayed, I could keep refreshing the data every 1/2 second say in case corrupted data was occasionally clocked for some reason like power glitch. This way the erroneous reading will be shown for a very short time saving embarrassment. Do you have any other thoughts that can make my Panel close to perfection? Mahmood |
| Topic | Author | Date |
| LED Display Panel reliability | Mahmood Elnasser | 12/25/09 03:18 |
| Serial out/serial in | Per Westermark | 12/25/09 08:11 |
| you do not say | Erik Malund | 12/25/09 08:13 |
| TB62706BFG | Mahmood Elnasser | 12/26/09 02:43 |
| Decision by majority... | Kai Klaas | 12/26/09 17:04 |
| I would use CRC | Daniel Contarino | 12/26/09 18:19 |
| I think you missed the point | Erik Malund | 12/27/09 06:58 |
| Reading back the parallel outputs then? | Kai Klaas | 12/27/09 08:54 |
| actually not | Erik Malund | 12/27/09 08:57 |
| "If" it is identical data... | Kai Klaas | 12/27/09 09:04 |
| Only the bits in error will generate any flicker | Richard Erlacher | 12/27/09 10:04 |
| But he wants perfection! | Kai Klaas | 12/27/09 10:54 |
| all the above | Erik Malund | 12/27/09 10:58 |
| So, there is a chance for flickering after all? | Kai Klaas | 12/27/09 11:05 |
| If incorrectly designed | Per Westermark | 12/27/09 11:10 |
| trust and trust | Erik Malund | 12/27/09 10:04 |
| trust | Mahmood Elnasser | 12/27/09 10:51 |
| layout, layout, layout | Erik Malund | 12/27/09 11:04 |
| Some hints... | Kai Klaas | 12/27/09 12:08 |
| excellent proposal! | Richard Erlacher | 12/27/09 14:13 |
| Principle of function... | Kai Klaas | 12/27/09 14:42 |
| Main EMI Source | Mahmood Elnasser | 12/28/09 08:26 |
| Thoughts | Andy Neil | 12/28/09 08:37 |
| I do not know your part, but ... | Erik Malund | 12/28/09 09:34 |
| Duplicate thread | Mahmood Elnasser | 12/28/09 13:13 |
| Replies | Mahmood Elnasser | 12/28/09 13:14 |
| double posted | Mahmood Elnasser | 12/28/09 12:59 |
| reply | Mahmood Elnasser | 12/28/09 12:59 |
| RC-filtering and Schmitt-Triggers will help! | Kai Klaas | 12/28/09 10:07 |
| response to above | Erik Malund | 12/28/09 15:09 |
| Why??? | Kai Klaas | 12/28/09 15:49 |
| maybe today | Erik Malund | 12/28/09 19:58 |
| Hardly... | Kai Klaas | 12/28/09 20:20 |
| no | Erik Malund | 12/29/09 07:38 |
| Wild speculations... | Kai Klaas | 12/29/09 08:42 |
| Mahmood | Erik Malund | 12/29/09 08:55 |
| No fun to poke about in a fog... | Kai Klaas | 12/29/09 09:06 |
| The clock on top | Mahmood Elnasser | 12/29/09 13:26 |
| yes | Mahmood Elnasser | 12/29/09 13:06 |
| He's not required to scroll | Richard Erlacher | 12/29/09 09:52 |
| absolutely, but | Erik Malund | 12/29/09 14:50 |
| You're right, Kai, but the frequent refresh costs less | Richard Erlacher | 12/29/09 09:48 |
| Again... | Kai Klaas | 12/29/09 09:56 |
| it is, indeed! | Richard Erlacher | 12/29/09 18:38 |
| not only then | Erik Malund | 12/30/09 08:26 |
| PWM with the TB62706?? | Kai Klaas | 12/30/09 09:23 |
| YES, of course there is | Erik Malund | 12/30/09 12:33 |
| Yes, I understand, but... | Kai Klaas | 12/30/09 13:46 |
| Good question | Mahmood Elnasser | 12/30/09 14:07 |
| Digital potentiometer? | Kai Klaas | 12/30/09 15:48 |
| the problem here is | Erik Malund | 12/30/09 15:59 |
| It seems they can be hooked together... | Kai Klaas | 12/30/09 16:36 |
| Look at this too... | Kai Klaas | 12/30/09 16:42 |
| Ok, then show me ... | Erik Malund | 12/30/09 17:43 |
| And write? | Per Westermark | 12/30/09 19:25 |
| :) | Erik Malund | 12/31/09 08:20 |
| Mahmod doesn't have multicolor LEDs! | Kai Klaas | 12/31/09 08:48 |
| well, what is this then? | Erik Malund | 12/31/09 09:35 |
| Question for Mahmod | Michael Karas | 12/31/09 10:56 |
| Looks like... | Kai Klaas | 12/31/09 12:07 |
| Yes | Mahmood Elnasser | 12/31/09 14:50 |
| Why from the same chip??? | Kai Klaas | 12/31/09 11:20 |
| limiting color choices to 'per chip' | Erik Malund | 12/31/09 11:43 |
| What flexibility? | Kai Klaas | 12/31/09 12:03 |
| well, if ... | Erik Malund | 12/31/09 13:17 |
| Dials | Mahmood Elnasser | 12/31/09 15:11 |
| just to clarify | Erik Malund | 12/31/09 17:56 |
| Length of chain shouldn't really matter | Per Westermark | 12/31/09 20:26 |
| Single color LEDs | Mahmood Elnasser | 01/01/10 01:22 |
| Vled only | Erik Malund | 01/01/10 07:10 |
| Two questions... | Kai Klaas | 01/01/10 09:56 |
| Answers | Mahmood Elnasser | 01/01/10 13:58 |
| What range? | Kai Klaas | 01/01/10 17:15 |
| Blue | Mahmood Elnasser | 01/02/10 03:11 |
| I would do it like that... | Kai Klaas | 01/02/10 09:59 |
| thanks | Mahmood Elnasser | 01/02/10 14:47 |
| Why? | Kai Klaas | 01/02/10 15:24 |
| Board to Board Hopping | Michael Karas | 01/02/10 16:54 |
| Of course... | Kai Klaas | 01/02/10 17:36 |
| one advantage, the other is .. | Erik Malund | 01/02/10 18:33 |
| Not 40xx Chips As Generic Family | Michael Karas | 01/02/10 19:39 |
| Clock skew? | Per Westermark | 01/03/10 02:27 |
| Still very very helpful... | Kai Klaas | 01/03/10 07:51 |
| There are lower-resistance solutions | Richard Erlacher | 01/03/10 14:26 |
| On-resistance is beyond any doubt... | Kai Klaas | 01/03/10 15:38 |
| I've used several '66s | Michael Karas | 01/03/10 18:59 |
| without any opinion of the 4066 option | Erik Malund | 01/04/10 08:01 |
| I didn't intend to suggest that it was wrong, Kai | Richard Erlacher | 01/04/10 09:46 |
| All the filter is supposed to do is limit the slew rate | Richard Erlacher | 12/28/09 17:18 |
| no, not the slew rate | Erik Malund | 12/28/09 20:06 |
| Scope plot? | Kai Klaas | 12/28/09 20:37 |
| Maybe later on | Mahmood Elnasser | 12/28/09 23:15 |
| Yes, that's true, but his worry was about his clocking | Richard Erlacher | 12/29/09 09:46 |
| Great | Mahmood Elnasser | 12/29/09 12:52 |
| suggestion | Jingxin Li | 01/22/10 00:06 |
Problem Solved!!!!!! | Mahmood Elnasser | 03/03/10 14:53 |



