Lightning

[Robert Atkinson 2Participant @robertatkinson2]

And this accident involving a Do228 commuter aircraft.
http://www.aibn.no/Aviation/Reports/2007-23-eng

The thread on the rose joint end of the elevator control rod (part of a turnbuckle like adjustment) was blown out by the lightning current. The design had a bonding lead across the joint but this had not bee maintained properly.

The report gives a bit of background on the currents involved.  As I’ve said before lightning protection of aircraft and avionics is one of my professional areas of expertise. Probably the hardest protection is fuel tanks when 200,000 amps is flowing through the skin without causing sparks of more than 200 microjoules inside. This includes rivets, pipe joints etc.

Full direct effects testing is always interesting. Pin injection testing for secondary effects (basically the OP’s issue) is relatively boring particularly if there are a lot of pins with different circuits. I have seen a unit catch fire during pin injection testing but fortunately it was someone else’s test being carried out at the same test house.

Robert.

[duncan webster 1Participant @duncanwebster1]

I’ve been a bit quiet of late as I’ve been having great difficulty uploading the software to my nanos. I bought them from aliexpress described as CH340 USB input. Turns out they are not, they are FTDI but a knock off version, so windows won’t talk to them. There is a  work around by using Linux,  but something to beware of.

This could well explain my difficulty uploading to minis as well.

Not bought opto isolator yet, too many other irons in the fire

[duncan webster 1Participant @duncanwebster1]

I’ve had to send the Nanos off to my very kind volunteer as I can’t get #2 son’s Linux laptop to speak to my interweb to download the IDE, so in the meantime I’ve been trying to get my head round Robert’s suggestion of 13 June. The picture has disappeared from the entry on the forum if I look on my PC, but it still there on my phone, Weird. Anyhow, I’ve managed to retrieve it, and almost understand it. What I don’t understand is why there are 2 off 330 ohm resistors rather than just one 680 ohm, and is it likely that the CTR (current transfer ratio) is the same on each channel. As Robert suggests I can use a spare channel to calibrate the other 2 as long as this is the case.

All getting a bit complicated, but easy teaches you nothing and so is boring. Just in case others can’t see the schematic, here it is again

[Robert Atkinson 2Participant @robertatkinson2]

Hi Duncan,

Good spot! I used two 330R resistors rather than a single 680R for a couple of reasons. The first was so there is an impedance in series with the LED in the opto-coupler and it’s protective Zener for surges coming from either the 12V supply end or the track connection end. This is mostly for very fast pulses were the stray capacitance can make a difference.
Secondly it divides the voltage across and power dissipated in the two resistors making them less likely to fail.

If you want to try something across the tracks I have some new old stock 100W 47V stud mounted Zeners…

Robert.

[duncan webster 1Participant @duncanwebster1]

I got all the signals fixed with the old design in time for the steam fair, so now I’m looking to the future. As the signals will not be operating when thunder is likely, I’m considering a combination of SOD and RobertA schemes, using a relay to ground the inputs when the power is off. I think a suitable one is HK19F dpdt, pcb mounted. I don’t understand the data sheet, in particular the coil resistance table, the max operating voltage of 9v for a typical 12v relay, and the minimum dropout. I suspect the latter 2 are mistranslations and max should be switched for min, but the table of coil resistance has me baffled. Anyone any ideas? What would be really neat is what I might call an untransistor, which conducts  when no external signal applied, but goes open circuit with a control signal. I’ve never come across such a device, but you clever chaps might know of something.

Second issue which has come to light is that rodents are attacking the cable from the signal to the JB. As the signals were built in batches, we have 2 types of cable, one has grey outer cover and has stayed flexible for a long time, the other is black and has gone very stiff, no doubt due to exposure to UV. Anyone got suggestion for a type and source, I need 6 cores, braided screen, around 1/4″ OD, tastes bad to rodents. Current is trivial, 80mA max at present but would increase if relay fitted (which might rule that out)

[duncan webster 1Participant @duncanwebster1]

Just done some more googling and found This relay, which is smaller, and only consumes 1mA, so belay the first question, in the second, I want to buy it by the metre, not full drum, which probably limits the field.

[Joseph Noci 1Participant @josephnoci1]

 

Second issue which has come to light is that rodents are attacking the cable from the signal to the JB. As the signals were built in batches, we have 2 types of cable, one has grey outer cover and has stayed flexible for a long time, the other is black and has gone very stiff, no doubt due to exposure to UV. Anyone got suggestion for a type and source, I need 6 cores, braided screen, around 1/4″ OD, tastes bad to rodents. Current is trivial, 80mA max at present but would increase if relay fitted (which might rule that out)

 

Use a cable with shielded twisted pairs, all with an overall shield and insulated outer. Use the shields of the signal pairs as true grounded shields and connect the outer shield to 50VDC thru a 10K few watts resistor in series with a red led.

Connect both ends of the shield ( if you can) to ground via a 100v 1uf mica or mylar, etc cap.

The 50V will tickel the rodent and he won’t like it. If a short developes, the led will light and you can go hunting…

 

[duncan webster 1Participant @duncanwebster1]

Thanks, but I’ve only got 12v , and not enough cores to generate 50v by buck-boost and distribute it, and anyway, it wouldn’t work when the system is switched off, which is most of the time

[Joseph Noci 1Participant @josephnoci1]

Slip the cable inside some 20mm plastic conduit? If you are asking after cable, I assume it means you have to replace/route it, so routing conduit should not be much more effort?

[duncan webster 1Participant @duncanwebster1]

That’s not a bad idea, I can actually get 8mm ID stainless flexible conduit (ebay), which would also keep the sun off, I’ll get hold of some, thanks

[not done it yetParticipant @notdoneityet]

Personally, and I’m not an expert, I think the suppression/diversion needs to be on the incoming supply (mains power line).

if the offending strike was close, I doubt that some damage to electronics can be easily prevented.

Re rodent damage – that could be the insulation material being grub to the little Bs.  A lot of cars have sustained damage to wiring because of this.

[duncan webster 1]

On duncan webster 1 Said:

I’ve

On duncan webster 1 Said:

I’ve been a bit quiet of late as I’ve been having great difficulty uploading the software to my nanos. I bought them from aliexpress described as CH340 USB input. Turns out they are not, they are FTDI but a knock off version, so windows won’t talk to them. There is a  work around by using Linux,  but something to beware of.

This could well explain my difficulty uploading to minis as well.

Not bought opto isolator yet, too many other irons in the fire

had to send the Nanos off to my very kind volunteer as I can’t get #2 son’s Linux laptop to speak to my interweb to download the IDE,

Bit Late in the game, but, I had the same issue and all it took was installing the CH340 USB driver. The IDE was happy with that.

[Joseph Noci 1Participant @josephnoci1]

Duncan, There is also a cable type issue – what type USB connector is on the board – mini/micro USB or Type-C?

If its the same as the previous Ft232 Nano’s you were using, then the cable you used for the Ft232 will be fine. But if it is now a type-c and you are using a new, different cable, ie type-C rather than mini/micro usb, make sure the Type-C cable is a DATA cable, and not just a charge cable. Still need the CH340 driver though.

 

[duncan webster 1Participant @duncanwebster1]

I’ve already got the CH340 driver as I’ve downloaded to nano with mini USB successfully. The USB cable I use is the one I use to transfer data to and from my phone. The suspect nanos have pattern FDTI chips, and W11 won’t recognise them. A friend has successfully programmed them using Linux computer. I’ve since borrowed a Linux computer and successfully programmed using the aforementioned USB lead.

This issue also affects the device  I use to program arduino mini. It works on W10 and Linux, not on W11

The current cable issue is from the PCB to the junction box. The grey wire always was more flexible than the black, and the black has become much stiffer with age, grey unaffected. I’ve not checked whether rodents eat the black (Sundays job) but they have nibbled holes in the blue water pipe feeding neighbouring farmers water tank

[Speedy Builder5Participant @speedybuilder5]

I was just about to chuck an old iPad away, charged it up to make sure it was clean.  It took some time to get to 100% (24 hours).  Yes, it looked clean, then I had an idea – download the Farm Boy Hit and miss plans onto it and keep in the workshop.

After a bit of mucking about, I used AirDrop (Apples short distance device to device copy) to copy the plans from laptop to iPad.

Saves dashing between armchair library and workshop with the printed plans.

Bob

[not done it yet]

On not done it yet Said:

Personally, and I’m not an expert, I think the suppression/diversion needs to be on the incoming supply (mains power line).

if the offending strike was close, I doubt that some damage to electronics can be easily prevented.

Re rodent damage – that could be the insulation material being grub to the little Bs.  A lot of cars have sustained damage to wiring because of this.

On not done it yet Said:

Personally, and I’m not an expert, I think the suppression/diversion needs to be on the incoming supply (mains power line).

if the offending strike was close, I doubt that some damage to electronics can be easily prevented.

Re rodent damage – that could be the insulation material being grub to the little Bs.  A lot of cars have sustained damage to wiring because of this.

The system is solar powered.

[duncan webster 1Participant @duncanwebster1]

I got the system up and running in time for the steam fair, but it came back to bite me a couple of weeks ago. Not lightning this time, one of the ULN2003s failed and instead of being an open collector with a 1k pull up to 5v it started outputting the ~12v supply to the preceding signal. Arduinos don’t like 12v on their inputs, so yet more repairs, made considerably easier by having the Arduinos on sockets this time.

I’ve tried my best to lay out a new board incorporating opto isolators but failed, there isn’t room in the box. I do have 3 inputs on one of the signals, so it’s a lot more complicated than the schematic shows. I have therefore dreamed up the attached, which I hope incorporates the best features. It now has 100 ohm at the input, capacitor to ground, 10k from the nominal 2.5v point to the actual input and a 4.7v Zenner and a cap from the input to earth. When the system is switched off, all the inputs are pulled to ground via a relay. To avoid having a 3 pole relay, in the case of 3 inputs 2 of them are grounded via diodes. In the normal 2 input scenario I’ll just leave out the diodes and put a link in place of one of them.

[Author]

Posts