Author Topic: Lab with NO grd, what about gear?  (Read 1769 times)

0 Members and 1 Guest are viewing this topic.

Online bd139

  • Super Contributor
  • ***
  • Posts: 6922
  • Country: gb
Re: Lab with NO grd, what about gear?
« Reply #25 on: October 16, 2018, 06:13:03 am »
If you connect it to an isolation transformer, there's still mains voltage across two points. There's just no path between one terminal and ground through you.
 

Offline FriedMule

  • Regular Contributor
  • *
  • Posts: 158
  • Country: dk
  • Can make even the simplest task look imposible.
Re: Lab with NO grd, what about gear?
« Reply #26 on: October 16, 2018, 01:21:02 pm »
I can see how that would protect me!
What about a common wire so all gear have same voltage like a common ground?
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 2713
  • Country: us
Re: Lab with NO grd, what about gear?
« Reply #27 on: October 16, 2018, 02:09:40 pm »
First I will apologize to the safety rules zealots.  Second, I will emphasize that if you don't know exactly how all of this works you can do something dangerous by accident.  So be very thoughtful while reading this.  And then I will proceed with some opinions.

There are two reasons for the earth ground in our test gear and supplies.  One is for safety.  The other is to provide an element of electrostatic shield which can reduce capacitively coupled noise in our systems.  It may also provide a lower impedance path than other grounds directly reducing simple ohms law noise from ground currents.

a.  The safety arguments are valid, but you can limit risks adequately to satisfy most people without earth grounding.  The most important protection is against hot to case shorts.  These are actually uncommon, and can be further mitigated by periodic checks of insulation integrity.  Even when such a short occurs, health hazard is only serious when there is a path to neutral or earth ground through the body of the operator.  In a normal indoor location standing or sitting in a chair you have a very high resistance path to any of the danger paths.  A higher risk occurs from contact with metal benches and the like, but again periodic isolation tests can mitigate the risks.

The only time significant likelihood of injury occurs is when this path is low impedance - in a bathtub, on a wet floor or other circumstance.  And even then, only if the occurances mentioned previously that make tying your ground to a water pipe an inadequate solution are present.

Remember that there is no such thing as completely safe.  Things which comply with electrical codes can still hurt you.  They are somewhat more safe than some practices that violate the codes.  And the codes are designed to keep things safe enough for children and knuckle dragging adults.  You may well be safer operating your equipment outside of approved code but with your eyes open and brain active than the general public is while operating code approved installations.

b.  The noise issues may or may not be important to you.  It depends on what you are doing and how you have set up your equipment.  But if it is an issue it is easily resolved by connecting the grounds of all of your instruments together.  If your instruments have chassis grounded BNC (not uncommon) it is merely a matter of hooking them up with BNC or other shielded grounded connector.  This slightly increases the safety risk because it effectively combines the chance of a failure in any of the connected instruments.  But this chance is low, and you did decide to periodically check isolation didn't you?  Now we just arguing about inspection interval and how much risk you will accept.
 

Offline FriedMule

  • Regular Contributor
  • *
  • Posts: 158
  • Country: dk
  • Can make even the simplest task look imposible.
Re: Lab with NO grd, what about gear?
« Reply #28 on: October 16, 2018, 06:32:45 pm »
CatalinaWOW hmm how risky I want it to be?!

I am willing to know that any lead with voltage and current can hurt me, but I am not willing to get Einstein hair by touching any of my gear!
In other word I want to treat it as anyone would do a blender ore something like that, you don't get hurt by using the machine normally, but don't be an idiot and stick your hands down to the knives!! :-)

So I do get an isolation transformer, make a common wire (floating ground) and everything is fine?

EDIT: About noise, I would low to get as little noise as possible, I will be working a lot with LF analog circuit but will be "fighting" common noise and HF from interfering.
« Last Edit: October 16, 2018, 06:40:45 pm by FriedMule »
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 2713
  • Country: us
Re: Lab with NO grd, what about gear?
« Reply #29 on: October 16, 2018, 08:18:53 pm »
CatalinaWOW hmm how risky I want it to be?!

I am willing to know that any lead with voltage and current can hurt me, but I am not willing to get Einstein hair by touching any of my gear!
In other word I want to treat it as anyone would do a blender ore something like that, you don't get hurt by using the machine normally, but don't be an idiot and stick your hands down to the knives!! :-)

So I do get an isolation transformer, make a common wire (floating ground) and everything is fine?

EDIT: About noise, I would low to get as little noise as possible, I will be working a lot with LF analog circuit but will be "fighting" common noise and HF from interfering.

Think it through.  In normal situations you will be fine, with and without the isolation transformer.  The isolation transformer provides protection against one class of faults, not all faults.  It also intoduces new possible faults to the system such as winding shorts. 

The ore grinder analogy is a bit strained, but much of the discussion here is not about sticking your hands in the blades, but about what happens when the shield that stops you from doing that fails.
 

Offline FriedMule

  • Regular Contributor
  • *
  • Posts: 158
  • Country: dk
  • Can make even the simplest task look imposible.
Re: Lab with NO grd, what about gear?
« Reply #30 on: October 17, 2018, 12:34:10 pm »
I know that it was a a bit fare out but what I ment was: Security level should be, if I am idiot, its my own fault, if the gear fails, it shall not influence me. :-)

The system in Denmark do the following: It compare the in and out power and if there are the slightest irrigularity it triggers a breaker and the whole house goes dark!
So if a case is live and i touch it, the breaker will trigger in ms.

What I am concerned about is some says that no ground is the best when measuring electronic i.e. Dave's "how not to blow up your scope". On the other hand, there are also a lot about preventing noise, static electricity and so on.

So the "simple" question: "what would you do if you was in the same situation, lots of power but no earth.
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 2713
  • Country: us
Re: Lab with NO grd, what about gear?
« Reply #31 on: October 17, 2018, 01:12:05 pm »
What I would do is go ahead and use it without grounds.  But that isn't really useful to guide you.  I have both experience and knowledge that you are apparently lacking, and a hard won understanding of the cosequences of error.

Dave is correct, you can blow up your scope.  If you don't understand the mechanism and avoid it you can do this in many ways, and no single rule will avoid it.

This also plays into noise.  Much noise results from unplanned and/or uncontrolled signal paths.  Which includes multiple returns to ground.  While a brute force, ground everything approach it is quite possible to reduce noise without minimizing.  And making it really difficult to figure out what to do to further reduce it.  "Which path did that ground current take, anyway."

I'm sorry I can't give you a quick and pat answer, but there really isn't any substitute for thinking through all of the consequences of different choices.  A couple of examples.  How much total power will your lab consume.  All instruments, lights, fans solder stations and the like.  The how much does an isolation transformer to support that load cost.  Or, how diligent would you be in periodically checks of case isolation if you choose that method for protecting yourself?  How diligent would you need to be to satisfy your safety requirements.  How do the answers to those questions relate to your standard of "If I make a stupid mistake that's my tough luck, but I don't want to be hurt by someone else's error.
 

Offline FriedMule

  • Regular Contributor
  • *
  • Posts: 158
  • Country: dk
  • Can make even the simplest task look imposible.
Re: Lab with NO grd, what about gear?
« Reply #32 on: October 17, 2018, 04:11:13 pm »
Wow what a lot of great questions!!

As you rightly says, I am to new to this, my first lab, therefor all my "problems" :-)

About noise and signal path, I am thinking of using one power outlet and then one power strip, all without grd connection.

I do not understand the following: "brute force ground" and "Which path did that ground current take, anyway."

About the consequences, that's it, I do not know the consequences or my options at all.

About testing case isolation, all I know is that if I touch a case with voltage in it, the breaker will shut off in milli seconds.

EDIT: would a cheap max 500W isolation transformer be ok? (https://www.amazon.de/gp/product/B0067K0ESA/ref=ox_sc_act_title_1?smid=A3JWKAKR8XB7XF&psc=1)
« Last Edit: October 17, 2018, 04:13:27 pm by FriedMule »
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 2713
  • Country: us
Re: Lab with NO grd, what about gear?
« Reply #33 on: October 18, 2018, 12:22:32 am »
Your ground fault breakers cannot detect all faults in equipment on the output side of an isolation transformer.  Whether that matters depends on details of your lab set up.  Again thinking is the only solution.  What would the current path be if case isolation failed and you touched the case.  Think of all the other things you might be touching at the same time.

A 500W isolation transformer may or may not be fine.  I have no idea how many devices you have in your lab, or what their power requirements are.  If you are testing high power stereos or amateur radio transmitters that unit you are testing could consume all of that with nothing left for test instruments or anything else.
 

Offline Ice-Tea

  • Super Contributor
  • ***
  • Posts: 1099
  • Country: be
    • Freelance Hardware Engineer
Re: Lab with NO grd, what about gear?
« Reply #34 on: October 18, 2018, 12:29:41 am »
About safety: In Denmark the circuit breaker trips if there are any different in the power between phase and neutral so if you touch a metal case, it will trip in a tiny fraction of a second.

Funny thing though: the RCDs are ussually for 300mA. Too bad that a rule of thumb says things get hairy from, say, 100mA or so.
An engineer never has a problem. He just needs more time.
eBay shop with all the gear you need!
FS: TTi TSX1820P, Agilent Infinium 54815A, 54825A, R&S CMU 200  Tektronix CSA8000B, HP 8594E, 8595E, 8662A, 8663A, R&S SMHU, IRF 2051, Marconi 6201B 8GHz
 

Offline 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1040
  • Country: de
Re: Lab with NO grd, what about gear?
« Reply #35 on: October 18, 2018, 12:46:31 am »
Dunno about other countries, but in Germany the RCD has to trip for leakage currents <= 30mA if the RCD is used to protect from injury. 300mA are only allowed for fire prevention (i.e. for installations that a normal person could never touch).
Trying is the first step towards failure - Homer J. Simpson
 

Offline Ice-Tea

  • Super Contributor
  • ***
  • Posts: 1099
  • Country: be
    • Freelance Hardware Engineer
Re: Lab with NO grd, what about gear?
« Reply #36 on: October 18, 2018, 01:16:47 am »
Dunno about other countries, but in Germany the RCD has to trip for leakage currents <= 30mA if the RCD is used to protect from injury. 300mA are only allowed for fire prevention (i.e. for installations that a normal person could never touch).

AFAIK: in "wet" rooms, where injuries are most likely, 30mA or less is required. The "main" RCD, on which a "regular" room would be, 300mA.

So, yeah, what you said ;)

Note, however, that does not help you much anyway if you connect L and N with your body...
An engineer never has a problem. He just needs more time.
eBay shop with all the gear you need!
FS: TTi TSX1820P, Agilent Infinium 54815A, 54825A, R&S CMU 200  Tektronix CSA8000B, HP 8594E, 8595E, 8662A, 8663A, R&S SMHU, IRF 2051, Marconi 6201B 8GHz
 

Offline bsdphk

  • Newbie
  • Posts: 4
  • Country: dk
Re: Lab with NO grd, what about gear?
« Reply #37 on: October 18, 2018, 02:46:34 am »
Hi from another Dane :-)

First thing:  Determine if your installation has a Residural-Current-Interrupter (Da: HPFI relæ)

It almost certainly does, and that means that the missing earth is not likely to kill you.

If your lab is on the ground-floor and has an exterior wall, call an electrician and get a quote for a ground connection (Da: Jordspyd) for the outlet you plan to use (or a new one next to it.), it probably won't cost as much as you think.  If he asks why, tell him you need it for EMC/EMI reasons.

If you cannot establish a ground connection, for whatever reason, it is *probably* still a good idea to use 3-pin plugs and a distribution-strip which connects all your instruments earth together.

The good idea is that you do not get a voltage potential between your instruments, the "*probably*" is that if any of them has "Y" capacitors, then they will all float around 115VAC above earth.

The easy test is to run the back of a finger lightly along a metal edge on the instruments, if that buzzes, you really need true earth connection (or the courage to fix the power-filtering in the instrument yourself.)

An isolation transformer will just be a waste of time and electricity (unless you plan to work on old tube-based boat-anchors with transformer-less power-supplies).

Feel free to drop me an email.
 

Offline FriedMule

  • Regular Contributor
  • *
  • Posts: 158
  • Country: dk
  • Can make even the simplest task look imposible.
Re: Lab with NO grd, what about gear?
« Reply #38 on: October 18, 2018, 11:04:31 am »
Thanks a lot you all!!:-)

I do have a HPFI device, so as I understand it, I can connect all my gear to one distribution-strip with ground so that all my gear have a common connection and ignore that it is not grounded in the wall outlet. Unless that I do feel that sudle vibration in one of the cases.
I can maybe try to switch polarety on the plug if I get that vibrating feeling in my fingers.


If I am wrong do please shaut, before I play with the mains!! :-)
 

Offline IanMacdonald

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: gb
    • IWR Consultancy
Re: Lab with NO grd, what about gear?
« Reply #39 on: October 19, 2018, 05:13:06 am »
Think the point to note is that the main safety purpose of earthing metal cases is to prevent any internal fault from making the case live.  So for the average end user of a toaster or whatever with a metal case, earthing that case is a safety improvement over not doing so.

However, on a test bench where exposed high voltages are present, especially when working on valve (vacuum tube) gear or SMPS, the presence of earthed objects on the test bench is a safety hazard. This is because you can't get a shock from a single connection. So, you inadvertently touch the anode  pin of an EL34 with maybe up to 700v on it, and you might get a slight tingle if your shoes aren't all that good as insulators. However, go to adjust a control on your earthed scope, and you'll be paying Emily van Dort a visit as the earth completes the circuit.

Also, earth fault trips can only protect against current flowing from the mains live to earth. That is because they don't actually measure earth current, they compare the live and  neutral currents and note any difference.  If the HV came from DC PSU in the equipment, then it won't trip. Thus while an RCD/GFI on the bench is a good idea, don't assume it will save your life in all types of mishap.

Basically, when working on exposed circuitry you need to be constantly aware of the risk of a hand-to-hand shock, since this is the most dangerous. Best approach is never to use two hands at the same time. Especially, be careful to keep your non-dominant hand OFF any earthed metal. If you need to adjust test equipment whilst also holding a probe, do so by insulating knobs only without touching the panel.

The danger of not earthing test equipment such as scopes and sig gens is that an internal fault could make the case live. If you only ever work on low voltages this is the best arrangement. For high voltage work the best approach is to run the test equipment via an earth-free isolating transformer. So in fact you need two isolating transformers, one for test gear, and one for the equipment under test. Using the same one for both is not a good idea.

A cutoff button on the bench is also a good idea. Especially if someone else has to come to your aid, it means they don't have to stop and think about how to cut the power. or whether it is safe to touch you.

Oh, and beware of building site 'isolating transformers' which are actually center-tapped autotransformers. Correctly termed tool transformers, these are NOT for testbench use and would actually make things a good deal more dangerous than with no transformer.

Bottom line is that there is no absolutely safe way to work on exposed electricity. Avoiding accidents is a question of using your noggin and being 'ahead of' things that could go wrong. But, isn't that always the case anyway?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf