PDA

View Full Version : Noise, Interference, and unexplained electrical behaviour



g4ghp
23-04-2014, 09:21 PM
Hello everybody, after briefly introducing myself a couple of years ago, and welcomed by all, I fell ill and have been out of circulation since. I have been taking weekly trips to London to the hospitals there for on going treatment. While having this treatment, I was called into the local hospital down here in Poole, (still on the temporary staff payroll) where they asked me to perform some tests on provisional newly acquired life support equipment. My previous job before becoming ill was medical clinical engineering, and responsible for checking out new medical kit (electronic) to be purchased by the NHS in the future. It has been a real eye-opener for me over the last 28 years what I have seen and experienced in the screened test room facility. Myths have been disproved, and anything to do with reliability, noise, interference, and construction, along with solutions has been my bread and butter. Spectrum analysis shows you without doubt the hidden threats and the stuff you can't grasp, hear, or sea under normal conditions
Having read some of the posts over the last couple of years regarding interference, noise, or basically equipment not behaving itself, I wondered would there be enough interest for me to put down on paper my findings, with possible solutions, and construction ideas.
I am in the middle of trying to get my ideas down for a small cnc machine, nothing too big and mainly for woodwork and some aluminium machining. Doing everything the wrong way round ,:hysterical: I have already got some of the electronic components......mainly out of curiosity. So to wet the appetite, I purchased some Leadshine components along with the hbs86 drivers from Zapp, and played around with them and was quite impressed with the closed loop performance. They did exactly as they said on the tin.... and did not fail. Then having taken other makes of drivers and steppers into the screened room I subjected them to just about everything I could in terms of noise, and interference, from fast rise time spikes on the mains supplies, rf narrow band, and wide band signals , of high power, diverse modulation, running from 1mhz to just over 10 ghz....simulated both mobile phone interference, and also high power pmr and all the commonly used ISM frequencies. Also for good measure, 4 different makes of tower pc's, inside the screened room. ( for the clocking generated wideband hash)
The outcome was the Leadshine never failed at all, quite puzzling to me, but the other drivers and steppers failed or missed steps continually. So...... more investigation into the drivers. the run of the mill drivers, (some from Arc euro trade) and other suppliers seemed to have similar failure characteristics mainly losing steps... or even freezing completely. So as you can imagine I was quite baffled why the Leadshine drivers should be any different. It was only after taking the hbs86 driver apart I realized that the decoupling of all the circuitry was not relying on the standard .01 or .001 capacitor which frequents most of the standard drivers and digital pcb's. The Leadshine driver uses both notch and both pi and t decoupling made up of a chosen inductor and capacitor combination, effectively allowing a far more broadband decoupling to be achieved than the standard run of the mill. I would never have believed it if I had not taken the driver apart. So there we are, another hidden bonus for the Leadshine driver and stepper. They cost a lot more than the run of the mill steppers and drivers, but for me that reliability is worth it ...and I think at the end of the day you get what you pay for.
I would like to say right now, I am not associated with Zapp, or Leadshine in any way at all. And my findings are just that ..my findings only so no axe to grind or anything. Many of you will not suffer with any of the above, it really depends on the background noise and interference at the location of the cnc or the workshop. Unfortunately it is hard to know what the background noise is in your workshops, and that will always be the fly in the ointment.
If there is enough interest, I might get into writing my findings on other matters which have a direct relationship to the noise and interference, and how we can simply with good construction methods remove a large part of the unseen threat, and maybe show the Achilles heel in trusting the CE approval to readily.

Neale
23-04-2014, 09:58 PM
Is that more sophisticated decoupling/filtering on the power lines, or is there also something on the signal lines to try to keep noise out? I would guess that it's power only due to effects on pulse shaping, etc, but curious to know. Does it also imply that with lower-spec drivers, it might be worth adding some filtering on the power lines at the driver terminals (or even just inside the box for those brave enough)?

JAZZCNC
23-04-2014, 11:39 PM
Hello everybody, after briefly introducing myself a couple of years ago, and welcomed by all, I fell ill and have been out of circulation since.

Ah great you posted it, thanks Bren. Catch you later mate.

g4ghp
24-04-2014, 11:47 AM
Is that more sophisticated decoupling/filtering on the power lines, or is there also something on the signal lines to try to keep noise out? I would guess that it's power only due to effects on pulse shaping, etc, but curious to know. Does it also imply that with lower-spec drivers, it might be worth adding some filtering on the power lines at the driver terminals (or even just inside the box for those brave enough)?

Hi Neal, the coupling was actually on the main pcb, and it would be quite difficult to modify an existing pcb as space, and track layout would be a nightmare. But you have hit the nail on the head with regard to the power supplies and both the signalling and supply lines. more robust measures can be taken and often improve a poor situation.

Take for example the standard linear psu that many people build, using a toroidal transformer, bridge rectifier and electrolytic capacitor. Standard toroidal transformers are very susceptible to external interference of varying type and magnitude. They easily become one of the prime sources of spreading interference within a cabinet set up......typical of many of the cnc installations. So, how do you resolve this problem?. First have the mains introduced to the power supply chassis via a fully grounded and screened mains socket/ input filter, providing enough attenuation to interfering mains spikes. Secondly the toroidal transformer spec. For example the psu I put together recently required a 2x25volt secondary to give me approximately 70 volts. pay a bit more money and get a toroidal that has both a electrostatic screen, and a emi/rfi shield, so then you know the transformer is not going to be susceptible to any external interference. Next, put the whole thing (psu) in an aluminium box (rf tight) and take the output from the psu via a substantial feed through capacitor, capable of handling the maximum current and a bit more. Then for final measure, wrap the output lead through a emi ferrite ring to kill off any other external noise of rf influences.

So there you go, it all makes sense, and is easily achievable but it does cost a bit more, and there is a bit of chassis bashing to be done , but you now know that if you are having problems of interference or noise, your main psu is unlikely to be responsible.

irving2008
24-04-2014, 12:39 PM
Takes me back to my first job with Plessey Avionics EW labs in Ilford (long since gone), used to have a big Faraday cage to do Tempest testing on various military equipment. I agree with your last comments, though some might consider overkill for 80% of cases.

Irving (ex-g8ikv)