I am trying to embed an arduino into a machinery which produces electric sparks at high fequency (about 8000V, 10KHz). The test point where I get the signals to control is at 15V DC.
I put Arduino into an aluminum sealed box. Signals get into the box by a shielded cable, a voltage divider and a low-pass filter. The box and the cable shield are grounded.
However, any time a spark occurs, Arduino freezes up.
Any suggestion, please?
Thank you.
Antonio Zanardo
Ferrite cores around all wires entering and leaving the box. Earth the box.
Riva:
Ferrite cores around all wires entering and leaving the box. Earth the box.
. . . and make sure that the cable shield is bonded to the case at the point of entry. A shielded cable entering through a hole acts as an antenna.
Russell.
Is the box really sealed? Good RF shielding is not easy, you typically need copper spring-strip
or braid for all mating surfaces. A slit in a box acts about the same as a round hole of the same
dimensions, so you cannot just use a standard box with a lid, the lid-box interface will be
completely see-through to RF.
The best approach is mount the electronics as far away from the source of interference and
condition the signals on the cables with feed-through capacitors or suchlike.
Tank you to all of you for your answers.
I have just finished all the mods you suggested, but the feed through capacitors as I was unable to get them from my usual supplier.
As for the box, it is an IP67 one, that means waterproof. When the lid is seated on the box it makes a very good electric contact, anyway I connected internally the lid to the box itself with a cable.
I'll let you know if all those mods will do their job: tomorrow I'll do another test.
Thank you again.
Antonio Zanardo
I see lots of high tech answers, and they look great.
But backstep, and add lots of caps on the rails, may also help.
How close to the spark is the box (arduino) ?
What function is the arduino expected to do?
What kind of equipment produces 8,000 volt sparks at 10khz? Maybe try to suppress those sparks, several methods available.
Hi
Could be playing Tesla Coils.
Tom... 
AntonioZanardo:
As for the box, it is an IP67 one, that means waterproof. When the lid is seated on the box it makes a very good electric contact, anyway I connected internally the lid to the box itself with a cable.
Connecting the lid to the box with a cable won't do anything. It needs to be connected all round for effective shielding. Your sparks, even though they are at 10 kHz will be generating much higher frequencies.
If your box is waterproof, Iguess it will have a rubber seal. You could try replacing that with copper braid. However I suspect that your main problem is interference getting in along cables. As has been said, decouple wwith feedthrough capacitors and earth the braid of screened cables at the point of entry not with wires.
Russell.
russellz:
If your box is waterproof, Iguess it will have a rubber seal. You could try replacing that with copper braid. However I suspect that your main problem is interference getting in along cables. As has been said, decouple wwith feedthrough capacitors and earth the braid of screened cables at the point of entry not with wires.Russell.
Thank you , Russel
That's what I did with the copper braid. As for the feedthrough capacitors, I was unable to get them from my dealer.
On the other hand, I need to detach the signal cable from the box by using a screw-in connector, therefore I can't figure out how I could make use of those feedthrough capacitors.
Anyway this morning I made a new test having placed a couple of ferrite cores at the input of the low pass filter and used the copper braid. The signal cable is the only one that enters the box.
The outcome of several test runs was almost 100% good.
A 16/2 LCD display is embedded in the box to show how Arduino works.
However during the test I read some odd characters , but I know for sure that Arduino code ran 100% right, having ascertained that by connecting the box to a PC and carefully analizing the data collected. None of them were corrupted.
In other words, apparently Arduino is much less sensitive to HF than the LCD. Arduino may be working as expected, but the LCD doesn't, thus it's hard to say which one is faulty. Actually I had to replace a HF damaged LCD, but not the Arduino board. And also the new LCD sometimes shows garbage, while Arduino still runs Ok!
Could the display suffer HF much more than Arduino? In that case, how can I protect the display from HF?
Many thanks for all your answers.
Antonio Zanardo
AntonioZanardo:
When the lid is seated on the box it makes a very good electric contact
What's relevant is that the electric contact is seamless all the way round the lid, no
gaps in good contact, which requires springs or braid, cannot be achieved with rigid parts.
Hi,
Can you please post a copy of your circuit, in CAD or a picture of a hand drawn circuit in jpg, png?
How are you driving the high frequency from the logic output of the arduino?
What is the load?
What is the application, is it a Tesla Coil?
Tom.....
TomGeorge:
What is the application, is it a Tesla Coil?
Hi Tom,
Thank you for your interest.
First, I wish to anticipate the description of the plant where I wish to use Adruino. Afterwards I will post the circuit.
No, it is'nt a Tesla coil, but something similar to it.
Arduino is supposed to monitor a plant made by three TIG welding machines. Each machine has a high frequency-high voltage generator (10,000Hz, 8,000V).
At the very beginning of the the welding process, the electric arc is triggered by some sparks which last maybe 0.3 seconds. Then the actual welding goes on at 10-15V 80A DC current.
Each one of the three welders fires up in succession.
The entire process lasts about 45 seconds and repeats every 60 seconds.
The welder has a test point DOWNSTREAM the HF-HV generator, where I can get the variable the 10-15 voltage signal to be monitored.
To do that job, I wrote a code for Arduino Uno, which at home tests runs as expected.
In order to protect the system from HF, I made all improvements you guys suggested on this forum. Arduino works now definitely better, but the problem has not been completely eradicated.
I even set up a spark generator using an automobile coil to test the system at home. Arduino works perfectly in presence of those homemade (and weak?) sparks, but crashes when I put it next to the three industrial welders.
TomGeorge:
How are you driving the high frequency from the logic output of the arduino?
I am not sure to understand what you mean. My english is not my mother tongue.
What I am striving is just how to completely get rid of High Frequency bad effects.
I will post the skematics, anyway.
Tank you so much.
Antonio Zanardo
MarkT:
What's relevant is that the electric contact is seamless all the way round the lid, no
gaps in good contact, which requires springs or braid, cannot be achieved with rigid parts.
Hi Mark,
I replaced the rubber gasket with a piece of shielded cable, after taking off the external insulation. The gasket and the cable have, fortunately, the same cross-size, thus the cable seated perfectly all the way into the lid housing.
I believe that the two ends of the braid shoud be soldered to make a continuous conductive loop. I did that. Right?
Antonio Zanardo
Hi,
I am trying to embed an arduino into a machinery which produces electric sparks at high fequency
I am trying to embed an arduino into machinery.
The machinery produces electric sparks at high fequencich.
No problems, so you are just monitoring the test points of the three mig welders.
How have you got the welders connected to the arduino when it comes to gnd connections?
Have you tried just one mig on its own?
It is a hostile environment but microcontrollers in MIG gear is the way to go these days, so a solution has to be possible.
Tom....
Hi Tom,
actually the welders are of the TIG type (no MIG!) which needs high voltage sparks to trigger the arc. That's the problem.
I have tried just one welder at a time.
We are welding stainless steel. Stainless steel needs reverse polarity, that is NEGATIVE at the torch and POSITIVE to the piece to weld.
The welder test point is at NEGATIVE level just like the torch, thus I connected this one to the Arduino ground, while the bench cable is POSITIVE (+15V), so I connected this one to the to anolog input pin A0.
Of course, the signal passes through a voltage divider and a low pass filter before getting to Arduino.
As soon as possible I'll send a copy of the circuit.
Thank you.
Antonio Zanardo
Hi,
The voltage you are monitoring is a test point on the TIG control board, what is the voltage level between the test point and the CONTROL board gnd, this is what the test point voltage would be referenced against, not necessarily one of the output cables.
What does this monitored voltage represent and what do you use it for?
Tom.....
How fast do you need to detect transitions? 100 th of a second?
Use optocouplers.
Use RC ( or LC ) filters.
AntonioZanardo:
A 16/2 LCD display is embedded in the box to show how Arduino works.
However during the test I read some odd characters , but I know for sure that Arduino code ran 100% right, having ascertained that by connecting the box to a PC and carefully analizing the data collected. None of them were corrupted.
If you have a window in your screened box so that you can see the LCD display you no longer have a screened box! Can you move it further away from the welders making sure that the connecting cables are well protected with ferrite bead chokes?
Russell;
Hi Russelz,
You are right , of course.
But how else could I know what is going on inside the box?
At the beginning I connected Arduino to a PC, but no data were visible on the monitor as the PC immediately froze up.
Anyway I am posting a sort of layout-circuit of the gear, hoping it may serve to better understand the problem.
