??? 05/25/06 03:40 Modified: 05/25/06 04:07 Read: times |
#117062 - Trying to improve the performance... Responding to: ???'s previous message |
Now, where I seem to have a better understanding of your application I can try to give some hints:
But first: When designing the layout of your board, you violated some important design rules! With your actual layout you can hardly handle ESD events! Cables leave your board at different locations, so that ESD is always running across your whole board! In the future look, that all the cables are leaving the board at only one location, best near the point where the supply voltage enters your board. Here a star ground point must be provided, where also the metallic enclosure of mainboard is connected to. Also the protection earth should be connected to this point, either directly or via a cap. Even if there must be many openings for the LCDs or 7-segment-displays, your application will always benefit from a metallic enclosure!! You can even put wire screens over the openings for the displays, soldered to enclosure, maintaining the shielding capabilities of Faraday cage. Today there are wire screens available for shielding purpose, which are so transparent that you can hardly see the wires! Where the cables enter the enclosure, have ESD absorbers connected from each line to enclosure, and this with shortest connections!!! What is the sense of this designing? If an ESD event hits a cable outside of enclosure, for instance a cable coming from the keyboard, then it will be shunted directly to enclosure and further to protection earth. It has no chance to enter the enclosure then with an unsane amount of energy. Especially if immediately afer entering the enclosure there are additonal filters, furtherly decreasing the ESD energy running to the according section of mainboard. The idea is, to shunt the ESD arround the mainboard, so that there's no reason for the ESD event to enter the enclosure and to run to any point on the mainboard! But with your design, you inject the ESD directly onto the mainboard!! Then, even sophisticated filtering will not solve anything, just because the ESD-energy hitting the board is unsane and cannot be handled adequately! You can add some more filtering or ESD absorbers and slightly improve the situation, but a true cure of the ESD susceptibility will be impossible by this. Only a proper layout of mainboard in combination with proper shielding and filtering will make yor application immune against ESD events! What can be done to improve your situation? Remove the keyboard from the mainboard, if it's mounted on it. If not possible, then keep it isolated at least, by using plastic screws or plastic distance roles. Keep a distance of at least 2cm between all metallic parts of keybaord and mainboard. Provide an additional rf-ground (radio frequency ground, because ESD contains far reaching harmonics!) near J11, where the power supply enters the mainboard. Means, remove the solder mask of solid ground plane near J11. (It's assumed, that solid ground plane is connected to signal ground (GND) of mainboard.) Don't feed the cables of keyboard directly to J20, but to this rf-ground. Do this outside of metallic enclosure surrounding the mainboard, if you use one. If not, keep the keyboard cables isolated from the mainboard by maintaining a safety distance of at least 2cm to it. Connect from each line of keyboard an ESD absorber to this rf-ground by using shortest connections (every millimeter counts!). Connect a 100nF cap from this rf-ground to metal chassis of your device (trainer, or whatever you call it). (It's assumed, that GND is elsewhere connected to protection earth also, perhaps at the power supply unit, as you told in an earlier post.) Via this cap the ESD will run, from the user back to the rollers of belt, which also must be connected to metal chassis of device. Guarantee, that there's a good connection from this point (earthed terminal of above cap) to all metallic stuff of your trainer (especially the bearings of rollers, etc.). For safety purpose everything metallic must have good connections to each other and to protection earth (that is provided via mains cable) anyway. Now, route the keyboard lines from the ESD absorbers near this rf-ground to J20, by keeping the cable wires internally of metallic enclosure surrounding the mainboard, if you use one. If not, keep the cable near the PCB. Feed the cable optionally through one or more soft ferrite sleeves (cylinders). Change the protection circuitry near IC8279 as follows: 1. Change R39, R38, R88 and R37 to 100R. 2. Change R36, R35, R87, R34, R64, R66, R61, R63, R60, R62, R67 and R65 to 10k. 3. Connect from each keyboard wire to GND a 10nF cap (best near the protection diodes' circuitry). 4. Make a break in the connection of each keyboard wire running from J20 to the protection diodes' circuitry and insert resistors: 100R for the lines Y0...Y3. And 220R for the lines RL0...RL7. 5. Connect directly from the cathodes of D58, D12, D9, D8, D13, D26, D28, D22, D27, D23, D24 and D25 some 100nF caps to GND. What is the benefit of these changes? The ESD is shunted directly to the chassis of your trainer now, by feeding the keyboard lines to the ESD absorbers near J11 (rf-ground). Via the 100nF cap from rf-ground to chassis of trainer the ESD can flow back to the rollers, without flwoing across the mainboard. The unavoidable rest energy of ESD events is furtherly filtered by the soft ferrite sleeves and the following RC filters consisting of 100R respectively 220R resistors and 10nF caps. These filters slow down the rise time of ESD pulses reaching the input lines IC8279, so that no longer unsane spikes can contamintate Vcc or ground. Also the potentials on the keaboard lines ending at U4 (74LS156) and U37 (IC8279) are limited to sane levels now. By these measures ESD events hitting the keyboard should no longer make trouble. But if ESD events hitting the displays or the mainboard itself are responsible for the hang-ups, then you need to use a fully closed Faraday cage arround the mainboard, using above mentioned wire screens mounted over the enclosure openings and soldered to enclosure. By the way, you can fabricate the enclosure by yourself by using copper foil or simliar. Kai |
Topic | Author | Date |
ESD problem | 01/01/70 00:00 | |
try isolating signal gnd from earth | 01/01/70 00:00 | |
thanks richard | 01/01/70 00:00 | |
Watchdog not really useful | 01/01/70 00:00 | |
hi kalpak | 01/01/70 00:00 | |
re watchdog, never said better | 01/01/70 00:00 | |
hi steve | 01/01/70 00:00 | |
that will NOT help | 01/01/70 00:00 | |
hello erik sir | 01/01/70 00:00 | |
activating the the watchdog while error | 01/01/70 00:00 | |
hello erik sir | 01/01/70 00:00 | |
Pick off brush ? | 01/01/70 00:00 | |
yes | 01/01/70 00:00 | |
this may help | 01/01/70 00:00 | |
So what? | 01/01/70 00:00 | |
Ouch | 01/01/70 00:00 | |
protecting the human | 01/01/70 00:00 | |
But.. | 01/01/70 00:00 | |
And what if... | 01/01/70 00:00 | |
ESDĀ¬precautions | 01/01/70 00:00 | |
static | 01/01/70 00:00 | |
Not from 10mm away! | 01/01/70 00:00 | |
Bad ground | 01/01/70 00:00 | |
Earthing problem! | 01/01/70 00:00 | |
It was observed... | 01/01/70 00:00 | |
TAKE CARE! | 01/01/70 00:00 | |
Is your design based on luck? | 01/01/70 00:00 | |
how do I | 01/01/70 00:00 | |
Some instructions and guideline | 01/01/70 00:00 | |
Thank You steve sir | 01/01/70 00:00 | |
two reasons not to | 01/01/70 00:00 | |
Posting Diagrams | 01/01/70 00:00 | |
It's a pdf-file | 01/01/70 00:00 | |
Where? | 01/01/70 00:00 | |
Oh, sorry! | 01/01/70 00:00 | |
Filename | 01/01/70 00:00 | |
Name | 01/01/70 00:00 | |
Post it to me. | 01/01/70 00:00 | |
Link | 01/01/70 00:00 | |
Perfectly! | 01/01/70 00:00 | |
How | 01/01/70 00:00 | |
Aha... | 01/01/70 00:00 | |
fix it and repost | 01/01/70 00:00 | |
Buttons? | 01/01/70 00:00 | |
No | 01/01/70 00:00 | |
sorry | 01/01/70 00:00 | |
Try | 01/01/70 00:00 | |
the simple way is | 01/01/70 00:00 | |
Thank you | 01/01/70 00:00 | |
For those who need a link to the code | 01/01/70 00:00 | |
after seeing the link | 01/01/70 00:00 | |
So, now you understand why "all of you o | 01/01/70 00:00 | |
Can't open cdr-files | 01/01/70 00:00 | |
Some questions? | 01/01/70 00:00 | |
Yes you are right | 01/01/70 00:00 | |
oops | 01/01/70 00:00 | |
Test | 01/01/70 00:00 | |
Yes | 01/01/70 00:00 | |
stop the frtiggin babble! | 01/01/70 00:00 | |
I guess | 01/01/70 00:00 | |
Photographs ? | 01/01/70 00:00 | |
steve sir | 01/01/70 00:00 | |
No photographs ? | 01/01/70 00:00 | |
Sorry | 01/01/70 00:00 | |
so, it is OUR job | 01/01/70 00:00 | |
I had | 01/01/70 00:00 | |
Ground, earth, soil,... | 01/01/70 00:00 | |
Spark gaps ? | 01/01/70 00:00 | |
maybe he was, but ... | 01/01/70 00:00 | |
Good for surges, but for ESD? | 01/01/70 00:00 | |
Thank you | 01/01/70 00:00 | |
Some more questions | 01/01/70 00:00 | |
Yes | 01/01/70 00:00 | |
Frame, rollers bearings | 01/01/70 00:00 | |
Kai, is it time? | 01/01/70 00:00 | |
Hallelujah! | 01/01/70 00:00 | |
The mess with... | 01/01/70 00:00 | |
Semi isolating semi conducting enclosure | 01/01/70 00:00 | |
I can't see why... | 01/01/70 00:00 | |
I never understood why either, but ... | 01/01/70 00:00 | |
Measurement | 01/01/70 00:00 | |
Forget the watchdog! | 01/01/70 00:00 | |
kai sir | 01/01/70 00:00 | |
Hallo? | 01/01/70 00:00 | |
kai sir | 01/01/70 00:00 | |
What do you expect from a watchdog? | 01/01/70 00:00 | |
Unpleasant. | 01/01/70 00:00 | |
Which is why I doubt it is ESD | 01/01/70 00:00 | |
Curiouser and curiouser | 01/01/70 00:00 | |
instruments | 01/01/70 00:00 | |
you can't | 01/01/70 00:00 | |
Diagnosis ... | 01/01/70 00:00 | |
the difference can be 'legal' | 01/01/70 00:00 | |
expensive is not always expensive | 01/01/70 00:00 | |
Exactly! | 01/01/70 00:00 | |
Killing shock ? | 01/01/70 00:00 | |
Who says that... | 01/01/70 00:00 | |
Trying to improve the performance... | 01/01/70 00:00 | |
Thank you kai sir | 01/01/70 00:00 |