Brand new workshop - installing faraday cage/magnetic shielding?

[Someone]

Thanks for those tips - that was the kind of thing I was thinking of. I'm a bit wary though about bonding the incoming mains supply to another earth besides whatever is provided as part of the household supply. I'm not sure how it is wired, and don't want to end up causing problems. I know in a previous house, there was routinely a 5-10V difference between a wireless ground (buried in the 'real' ground) and the mains earth potential - connecting them together caused a significant current to flow. I also understand there may be possible fault conditions at the local substation which can result in high earth voltages which can be dangerous if they are in reach of a 'real' earth, but harmless if not. I'm probably being paranoid, but I want to check very thoroughly!

I will research how the supply wiring and grounding works here in case there are any gotchas there.

If I stay totally battery powered and independent of the supply mains for my electronics - would it be more beneficial to tie the shield (and my worktop ground plane) to the ground return of my scope or other instrumentation - which may or may not be the same thing as the mains supply ground?

I would have mains available in this space - I would just isolate it (e.g. disconnect the cable where it enters the shield) when I want to run a sensitive test.

Grounding will be so specific to the instrument and measurement configuration its hard to give any specific guidance. But on the topic of mains you will be in for some big arguments with electricians and the regulatory authorities since you will want your protective earth bonded at the border of your cage and to the neutral at the same point. Voltages induced on the neutral will appear all through measurements and even if you have a single run back through the house to the main neutral bonding (aka star grounding) it will still pickup interference, enter the high cost solution where local authorities are strict about the bonding location:

Does the "armored" electrical cable make a difference?

Sure can, but its expensive and you still end up with issues about where you bond the shield of the cable to.

[GK]

I once had to repair the solid state driver amplifier for one of a pair of tube-output 450kW 500kHz RF power amplifiers that were operating at Flinders university. They were the 0.9MW RF excitation for a rotomak fusion experiment thingie. The plate supply was provided by half a room full of paper-in-oil (PCB) capacitors charged to ~16kV. The grid drive power to the power output tubes was ramped up as the plate supply capacitors discharged, to maintain a constant power output for some milliseconds or so. Then there was another adjacent lab space chock full of hundreds of electrolytic capacitors in rack frames that were discharged simultaneously into the quadrature (IIRC) magnetic field coils of the rotomak.

For shielding most of the whole thing was enclosed inside a wooden frame supporting earth-strapped chicken wire mesh, like a big bird cage.

All up that was a lot of hardware to get back on line for just a little purple flash of light 

[lesaid]

I am trying to remember the '80/20' rule here, and that my bigger problem at the moment is 50Hz pickup, which I believe needs magnetic shielding. So - the strategy I'm forming is to accept that I need some kind of small magnetically shielded enclosure - as large as possible at sensible cost, but not the whole room! But RF is still an issue, and anything I can do cheaply and easily to shield the whole place should generally make life easier. So, putting aside the magnetic enclosure for the moment, does the following sound reasonable ...

Line the entire space, against the brickwork and concrete, with heavy-duty aluminium foil - available cheaply in 100M rolls aimed at the food service industry, and heavy enough not to tear accidentally - not like the lightweight foil used normally in the kitchen. This could be done in overlapping strips all the way around the space, including floor and ceiling, before the cladding goes up - the shield would be largely out of sight/touch and not at risk of damage or being inadvertently connected to unwanted things. The exception would be around the window and doorway. Strips would overlap, and be taped or glued in place. There would inevitably be some breaks and holes in it - e.g. for fixing cladding through it to the wall, but smaller than the holes in chicken wire.

Need to check if there might be damp problems with a waterproof lining against brick - if so, I'd use chicken wire instead of foil.

The window would have a chicken-wire screen that could be clipped over it (to the underlying shield) but be easily removeable (the window doubles as a fire escape). Similarly, the door would need to be lined and somehow flexibly bonded over the hinge. Maybe it won't need separate bonding opening side given that the door can closely abut the frame when closed - perhaps even with a little overlap on the edge of the screen, and I'm not trying to protect against very short wavelengths.

For the mains electricity supply, I wonder if I can sidestep some of the issues by providing separate 'clean' and 'raw' supplies on the workbench - such that the clean supply ...

• is provided by a 'pure sine wave' inverter from 12V or 24V batteries, charged from the 'raw' mains and independent of the existing electronics battery
• is used solely for test instrumentation (mainly a scope). The electronics under test is already battery powered and lighting would be from the window, supplemented by low voltage LEDs. Everything else (main lighting, soldering iron, power tools etc.) I should be able to do without during a test.
• during sensitive tests, the raw supply is disconnected by unplugging a connector at the shield boundary - hence no connections at all through the shield during a test - effectively, the clean supply is just a harmonically clean UPS supply
• can have it's neutral bonded to the shield since it would otherwise float, while the shield can also be bonded however the regulations and safety considerations require, to the household earth. There should be no current flowing between the clean supply and the shield or household earth.

A 500W inverter aimed at the solar power market is not too pricey, and claims to synthesize a harmonically pure sine wave supply from 12V (or 24V). This should be good enough for me - even if it isn't quite pure, it should be completely immune from anything lurking on the regular mains supply (not connected to it) and hopefully, would have a built-in filter to remove any high frequency component from the waveform. If this turned out to be a problem in practice, I could add a low-pass filter. I doubt if the inverter would radiate enough power directly to be a problem - I'm sure it has it's own screening. However, it would be easy enough to put it into it's own little shielded enclosure grounded to the main shield if necessary.

The above scheme does not connect to a buried 'rod' earth - I know this was recommended by one of the comments - but given the nature of what I'm now proposing - does this really matter? I don't really understand why the shield needs to be earthed externally to anything other than for safety, or to match earthed cables passing through the shield. If my test is totally self-contained within the shield, with no connections through it at all, and the test and workbench ground plane are themselves grounded to the shield from the inside - does it really matter to the test what the whole setup is grounded to externally?


BUT - even this might be over the top. The only piece of mains-powered test equipment that I anticipate needing is a scope, and I need to buy a new one anyway - my existing one is no longer fit for purpose (an old USB device whose noise floor and calibration are both deteriorating badly with age). I need to check if there are any good scopes out there that have an option to run directly from 12 or 24V DC. If so, then perhaps I don't need any 230V supply at all during a test, and can dispense with all the cost and complexity beyond a simple connecter at the shield boundary that can be unplugged when necessary. This approach would also keep 50Hz further away at the same time.

What have I missed?  And thank you everyone again for your time and comments!

[Someone]

I am trying to remember the '80/20' rule here, and that my bigger problem at the moment is 50Hz pickup, which I believe needs magnetic shielding.

You'll find 50Hz comes through in unexpected ways through grounds and the mains supply, your idea of an inverter sounds like a nice way to sidestep the law but it'll need to be outside the cage and build your own little distribution for it that grounds to the cage.

BUT - even this might be over the top. The only piece of mains-powered test equipment that I anticipate needing is a scope, and I need to buy a new one anyway - my existing one is no longer fit for purpose (an old USB device whose noise floor and calibration are both deteriorating badly with age). I need to check if there are any good scopes out there that have an option to run directly from 12 or 24V DC. If so, then perhaps I don't need any 230V supply at all during a test, and can dispense with all the cost and complexity beyond a simple connecter at the shield boundary that can be unplugged when necessary. This approach would also keep 50Hz further away at the same time.

I have a cute little tektronix storage scope that runs on 12/24V DC, its even sharp enough to capture serial packets for decoding by eye. Perhaps not such a silly idea.

[lesaid]

Thought I'd post a quick update after all the helpful comments a few months ago.

I took the decision to install a faraday cage as part of a new workshop since it seemed possible to do it easily as part of the initial build and fitting out. If I didn't do it then, I'd never be able to do it.

The room is now surrounded by a shield which at first sight, appears far more effective than I'd expected, and was cheap and easy to install. The total cost of materials was under £100 and it added only a few extra hours of labour to the job of a simple garage conversion.

The walls, floor and ceiling are lined with overlapping strips of aluminium foil. The window was covered with a removable wooden frame strung with a length of aluminium gauze fly screen, and ringed with the gasket. The door was also lined with foil and ringed with the neoprene gasket, set up to 'mate' with aluminium carpet bars (designed to edge and join carpets in doorways). The shield is mostly invisible from inside the room - the foil is all between the wall/floor/ceiling lining boards and the masonry.

The materials were mostly found from shopping around on ebay.

The initial test was far better than I had expected. Unsurprisingly, even before the window and door screens were in place, 'medium wave' broadcast signals were greatly reduced. However, when all was done, at the point of closing the door, FM broadcast radio (around 100MHz), two cell phones and a landline cordless phone all went dead. The loss of signal was dramatic and sudden - if the door was open the slightest crack - even if it appeared closed but hadn't quite been shut properly, everything worked normally. It was only when the last crack was closed that everything abruptly lost signal. I don't have the equipment to measure properly how effective this faraday cage is, but it certainly seems to work.

I don't know yet how useful this will be for what I'm working on, and I won't be in a position to find out for a few weeks. But I thought I'd let you know how it turned out so far, and that it is possible to create a seemingly effective faraday shield around a room on a shoestring budget and without a lot of effort, so long as it's done at the same time as other building work and not being retrofitted afterwards. And it seems to have been worthwhile taking the time to be super-pernickety about leaving no gaps and covering all apertures (e.g. for an extractor fan) with aluminium gauze to keep the shield continuous.

And even if this turns out to be a waste of time for my electronics developments, it has been an interesting exercise in itself!

Thanks again to all those that offered their comments and opinions.

[hammy]

Can you please post some pictures? 

[Fsck]

eevblog demands pictures.
Very cool.

[lesaid]

Can you please post some pictures? 

Sure - I'm now kicking myself for not thinking to make a photographic record during the work!  But I'll take some as soon as I can and upload something in the next day or two.

[Smokey]

Whatever you do, don't skimp on big foam cones.  More cones = more awesome!

[lesaid]

Whatever you do, don't skimp on big foam cones.  More cones = more awesome!

Darn it - now why didn't I think of that before I used up all my glue on tin-foil!

Taking photos this afternoon for upload soon.

[lesaid]

OK - after a bit of a delay, here are some photos. This is the first time I've tried to upload photos to this forum so I'm not sure how it's going to turn out but hopefully they'll arrive!

The workshop looks very rough in the photos - it is still a bare shell with no decoration, finishing or fitting out at all - but shows some of how the faraday cage was put together. Here are some of the features, illustrated by the attached pics.

The 'ingredients' that were used for the cage, from left to right in the pic, are

• a neoprene tube woven through with wires, used to make an electrical seal around the door, window shield and a removable panel around a screened enclosure
• A roll of aluminium gauze sold as 'fly screen'
• Carpet bars (sold to join carpets across doorways) - used to mate against the neoprene gasket
• Aluminium corner strip - used around the window shield and to make a seal over the door hinge when closed
• Two 100M rolls of catering heavy-duty aluminium foil

The workshop has been formed by partitioning off the main part of a single garage. It is a small space so hard to photograph the whole room! The 'view through the partition' is looking through a gap in the partition that was left to allow materials to be brought in and out during building and fitting out. Just visible to the right of the 'doorway' is a panel while will be screwed into place over the gap when all is ready. Visible around the doorway is a foil-covered frame which will mate with the foil on the centre panel when it is screwed into place. The partition is a sandwich of plywood, insulation and aluminium foil.

Care was taken to cover all breaks - e.g. the extractor fan is covered by gauze, compressed against the surrounding foil behind the wood frame.

The door was backed with gauze (though could have been foil) behind a protective sheet of plywood. It is edged with strips of carpet bar that overhang the edges all the way around by about 1/4 inch, against which the neoprene gasket has been glued (it is self-adhesive) and tacked at intervals for added security. Shown in the closeup of the door latch is a piece of the gauze wrapped around to make contact with the metal latch assembly, and through it, the metal doorhandles. There is a matching carpet bar fitted around the door frame against which the neoprene is compressed when the door is closed. The door hinge was a particular challenge - I couldn't find a way of making a permanent connection over the hinge, so instead, there is a strip of the neoprene up the door close up to the hinge, and also up the frame, such that when the door is opened wide, the two strips lie side by side without fouling each other. When the door is closed and the shield is required, a length of the aluminium corner strip is wedged into the right-angle by a few blocks of wood, pressing against both gasket strips and making a connection.

There is a screened enclosure under the window with a piece of kitchen-worktop on the top. This is intended for non-electronic work - drilling, filing and suchlike. Inside will go whatever I want to run that is screened away from the general workshop - e.g. a computer, mains battery charger or whatever. One of the pics shows this enclosure with the foil lining being installed, along with a plywood 'floor' to isolate the contents of the enclosure from the foil itself. This foil screen is mostly isolated from the main screen except for a single connection at the back.

One of the pics shows the window screen removed and resting on the worktop. This is ringed with aluminium corner pieces, clamped against the gauze, which mate with neoprene which rings the window frame. Slip bolts at each corner hold it in place.

Also shown is the 12V LED strip and PSU which, when the shield is in use, will be fed from a 12V lead-acid battery. These LEDs are very impressive - I have a 5M strip dissipating about 12W per meter with double rows of LEDs, These have been cut into two equal strips and used as parallel 'striplights'. They light up the space as brightly as I could have wished, drawing in total, about 5A from the 12V supply.

The 'electronics' bench is not part of the build. It is a regular if solid table wooden table, topped with a layer of foil and a toughened glass sheet on top. This foil is grounded to the main shield at the same point where the whole thing is grounded to the house 'earth'. There is no bonding to neutral, as when the shield is in use, all cables through the shield are unplugged and foil-lined covers fitted over the relevant plugs/sockets. When the shield is active, there are no cables of any sort passing through it.

This post is long enough - I'll post another one with an update about how it has performed so far.

[lesaid]

First test (described in earlier post) used an FM broadcast receiver (around 100MHz) and two cell phones. These worked full normally until the centre panel in the partition was not only put in place but pressed firmly into place. Then abruptly, the cellphones and broadcast radio all lost their signals. There was no gradual deterioration apparent - it went from 'normal' to 'silent' as though someone had flicked a switch.

The centre panel is not yet properly screwed into place, but it is becoming clear that the seal around it is inadequate. I was having a wireless alarm sensor moved (elsewhere in the house), but had the alarm engineer do an experiment with signal strength in the workshop while he was here. His equipment measured the strength on a scale of 1 to 10 on a frequency of just under 900MHz. Putting the sensor inside the shield, about ten feet away from the 'base station' in an adjacent room, and with some omissions - e.g. the door hinge wasn't sealed, and the mains power plugs were connected, not isolated and shielded (this was a quick opportunistic test!) - the signal strength at first was '9' or '10' (i.e. - full strength). By pushing against the centre panel in the partition, the strength started to drop - until, pushing hard, it dropped below '3' (the level below which the sensor is deemed to be 'out of range' of the base station). The limit was how hard I could push on the panel to try to seal it. 

Because that panel is intended to be screwed into place and not be routinely removable, I hadn't put a gasket around it - was just relying on foil pressing against itself to make a connection. Gasket is quite expensive, so for now, I'm going to try to pad out the foil around the frame with a rolled 'sausage' of foil before finally placing the panel. Hopefully that will provide a much better seal.

To get that level of reduction however between a sensor and base station within ten feet of each other, in spite of the screen not being completely sealed, is quite reassuring. Combined with the first test when the seal was more complete, I am looking forward to testing it properly after all is fully sealed. This will be some time next week.

For a faraday shield on a shoestring budget - looking promising so far!

[lesaid]

The shielded room is now complete and properly sealable, and I have done some preliminary tests - posting a summary of the results here for anyone interested.

The measurements were rough, based simply on the 'S-meter' in the receiver, nominally 6db per unit from 1-9. While uncalibrated, one of the tests was repeated with a different receiver and gave the same relative results, so I believe the estimates are probably in the right ballpark.

The 'distant' signals were commercial broadcast stations. Those 15M distant were low power amateur radio transmitters in the garden. Attenuations of >50db were where the 'inside' signal was below the minimum scale of the meter, or in some cases, was completely inaudible.

Approx frequency --	TX/RX distance --	Attenuation
1.5Mhz	15M	0db
1.5Mhz	Distant	>50db
21Mhz	15M	>50db
21Mhz	Distant	>50db
29Mhz	15M	36db
103Mhz	Distant	>50db
131Mhz	Distant	>50db
435Mhz	15M	>50db

The first test listed showed no apparent attenuation at all - the signal was the same inside or outside. However, on a distant broadcast signal on a similar frequency, there was a massive attenuation, from strength '9' to completely indetectable - even by ear. Can anyone explain this? Is the near field harder to shield than the far field?

As expected, at the lower frequencies, it made little difference whether the door was left open or closed. At the other extreme, at 435Mhz, if the door was left open, there was no measured attenuation at all - the room might as well not have been shielded.

All these tests were carried out with no cables passing through the shield, and nothing active inside apart from the battery-powered receiver. I was surprised how much noise appeared to be generated inside the receiver itself - at a level that was immediately swamped by background noise when outside the shield, but which became apparent inside. On other noise sources - I didn't get anything noticeable from live mains cables, so long as there were no switched-mode PSUs active. However, I did get noise from an Ethernet cable attached to a hub outside the shield, even though the computer inside the room was switched off. Unplugging that cable removed the noise. I also had noise from a 'smart surge protector' distribution board which must contain electronics to monitor the current flow - it shuts down if overloaded. It was not in use at the time - nothing plugged into it - but if it was live, it radiated noise.

Although I don't really need to know, I am now curious to find out just how much attenuation this room does provide at these various frequencies - any suggestions as to inexpensive ways of measuring this with more confidence in the accuracy and a greater dynamic range? I've started wondering about whether the answer is to try to build something myself - perhaps a kind of broadband wavemeter. Or a front-end to my USB-attached scope/spectrum analyser - but that wouldn't cover more than perhaps up to 50Mhz or so at the outside.

Comments or ideas welcome - as is probably obvious, I've never tried to do anything quite like this before!

[mzzj]

Can anyone explain this? Is the near field harder to shield than the far field?

Indeed.
http://www.learnemc.com/tutorials/Shielding01/Shielding_Theory.html

Intresting project.
BTW, If your shielding effectivenes seems to fade away with the oncoming  years it could be because your  aluminium foil is getting oxidized and connection between overlapping sheets is getting worse.

I bet similar construction to folded-seam tin roofing would be really good also in shielding....

[TerraHertz]

This is a really interesting thread. I had considered shielding my workshop too, but there wasn't time during the original construction phase, and now it's TOO LATE.
I'm hoping OP will eventually talk about the nature of the low signal level experiments he's doing. Sounds fascinating!

Can anyone explain this? Is the near field harder to shield than the far field?

Indeed.
http://www.learnemc.com/tutorials/Shielding01/Shielding_Theory.html

Intresting project.
BTW, If your shielding effectivenes seems to fade away with the oncoming  years it could be because your  aluminium foil is getting oxidized and connection between overlapping sheets is getting worse.

If I'd seen this thread earlier, I'd have suggested lesaid cough up the extra cash for copper/brass foil and mesh, since there's no way to make a good, permanent electrical seal with aluminium, due to the surface oxide layer and corrosion. Especially since it seems he's placed it in contact with brick outer walls, which are probably only single layer and will get very damp in wet weather. Every Faraday cage room I've ever seen used sheet and foil that could be soldered along seams, and all seams definitely were soldered. Continuously, not just spot soldered. Any break in the conductive surface becomes a slot antenna.

But it's obviously a bit late now.

[Lurch]

If you are worried buy a 3 phase mains isolator and run the PE through it as well,

I would seriously recommend not doing that. Protective conductors should never be switched, and as this is in the UK it is also part of the standards that need complying with. If isolation of the PE is required a 3 pole plug and socket arrangement is what is required, from a safety and compliant POV.

As this is on the UK though buried earth mats, whether by design or not, need special considerations before just connecting them to the mains incoming earth willy nilly. Various other factors need taking into account.

[lesaid]

BTW, If your shielding effectivenes seems to fade away with the oncoming  years it could be because your  aluminium foil is getting oxidized and connection between overlapping sheets is getting worse.

If I'd seen this thread earlier, I'd have suggested lesaid cough up the extra cash for copper/brass foil and mesh, since there's no way to make a good, permanent electrical seal with aluminium, due to the surface oxide layer and corrosion. Especially since it seems he's placed it in contact with brick outer walls, which are probably only single layer and will get very damp in wet weather. Every Faraday cage room I've ever seen used sheet and foil that could be soldered along seams, and all seams definitely were soldered. Continuously, not just spot soldered. Any break in the conductive surface becomes a slot antenna.

I did wonder about that when deciding how to do this. But remembering my very rusty school chemistry lessons from decades ago, I recollected that aluminium, while a very reactive metal, forms a protective oxide layer within minutes if not seconds, which is impervious to moisture - otherwise it would corrode away very quickly? Some solutions will penetrate that layer - if I remember rightly, solutions containing mercury were among them.

I did try and found, as I expected, that aluminium foil in contact with itself formed a good conductive path in spite of the presumed presence of this oxide layer - so not sure why it should degrade over time?  And old aluminium foil is still foil and shiny - it doesn't corrode away unprovoked by other chemicals?

On the damp front, things aren't as bad as they might have seemed. This garage is integrated with the house - only one wall is external. That wall is a cavity wall, and on the inside, is a layer of thermal insulation, then the foil, and then the MDF lining. So on that particular wall, the foil is not in contact with the brick. For the internal walls, one is the partition, which is an ordinary 'stud partition', faced by MDF on both sides, a layer of thermal insulation in the middle, and the foil between that insulation and the facing MDF on the inside.
 
But if it does degrade over time, so be it - it can't be helped now. I am not a professional relying on this to make a living. I am a hobbyist and able to let my mind and projects roam wherever whim takes me - and it keeps taking me off at tangents!  This shield started off in the belief that it would make my life more convenient for projects I wanted to do - and early signs are that it will, though as mentioned by several people (whose opinions I greatly respect), I should be able to get by without anything so elaborate.  I didn't build the workshop for the screen - I added the screen as a cheap and easy 'extra' since I had the opportunity.

And I have to admit - the comments about how difficult and expensive it would be did act as a little bit of a challenge to see what I could achieve simply, just by taking a bit of care and without spending a lot!

I also have added an old, heavy steel filing cabinet, the top drawer of which I intend to use as a secondary faraday cage and providing a degree of magnetic shielding as well. Some experiments driving a coil with 50hz AC outside, and trying to pick this up with a second coil inside, seem to show that it is reasonable as a magnetic shield. Not sure now how much I'll need that, but it is certainly there to fall back on if the room shield does start to degrade.

One unexpected problem, and hidden benefit, though obvious with hindsight, is that all these electrical seals are also excellent draft seals too - so when things are closed, there is no 'natural' airflow. The room quickly becomes very stuffy and hot, particularly if I'm running a computer in there, but it also means that my measurements settle down better, I believe in part, because there is little airflow to cause temperature gradients.

I am now redesigning the gauze window screen to allow the window to be opened while the screen is still in situ, and also looking for a better (and quieter) extractor fan so I can ventilate the room more quickly when I need to!

[lesaid]

If you are worried buy a 3 phase mains isolator and run the PE through it as well,

I would seriously recommend not doing that. Protective conductors should never be switched, and as this is in the UK it is also part of the standards that need complying with. If isolation of the PE is required a 3 pole plug and socket arrangement is what is required, from a safety and compliant POV.

Thank you for the warning!

After consulting with professionals about both safety and standards compliance, I haven't done that. The shield is simply grounded to the house 'earth' at one point, in a similar way to what is done for plumbing metalwork.

There is no fixed mains cabling inside the shield at all. When I am wanting mains power and don't care about screening, I have the original mains sockets that were already fitted to the garage. These have been boxed in, with the foil shield being continued in the removable cover of the box, so that the sockets are outside the shield. When I am not using the shield and need mains power inside, I remove the cover and plug in whatever I need to. When I want the shield, I disconnect everything, replace the cover and use a lead-acid battery for 12V. If I need mains power at that time, I'll look at acquiring a 'harmonically pure' inverter to run off the battery. But I'm trying to avoid needing to.

Oh - and another safety issue before someone mentions it - the household system is fitted with a modern, central ELCB at the consumer unit, and the distribution board I plug in to those sockets when necessary also has it's own built-in ELCB and overload protection (shuts down if the current draw exceeds whatever's set on the board).

[Lurch]

also looking for a better (and quieter) extractor fan so I can ventilate the room more quickly when I need to!

In line centrifugal fan. Higher airflow than a standard wall mounted axial fan and also can be located away from the cage so should be quiet in the work area regardless. Decent ones aren't that noisy though, but noticeable.

[lesaid]

In line centrifugal fan. Higher airflow than a standard wall mounted axial fan and also can be located away from the cage so should be quiet in the work area regardless. Decent ones aren't that noisy though, but noticeable.

Thank you - I'm searching over the next few days and will look out for that type. Currently looking at types aimed at commercial kitchens rather than domestic bathrooms, mounted separately from the area being ventilated, claiming around 37-44db noise and shifting over 400 cubic metres / hour - roughly four or five times as much as a typical bathroom fan. They are a bit more expensive but seem not too bad if you shop around.

I'm guessing this is the kind of thing you're thinking of?  This is a much larger space than the average bathroom, and I want to be able to shift the fumes from solvents and soldering as well as adding to the general ventilation.

By the way, I do have vents through the wall, allowing outside air to be drawn in by the fan - since there are few other ways for air to get in when the window is shut!

[lesaid]

I'm hoping OP will eventually talk about the nature of the low signal level experiments he's doing. Sounds fascinating!

Thanks for your interest
Here is an outline - I can drill down into more detail if you want me to.

I have been returning to an old passion for electronics after decades of working in IT. I am not formally trained in electronics, though used to be a radio 'ham' a long time ago. I am basically following my whims, wherever they take me! I became interested in tiny voltages and currents after musing on the behaviour of AC signals when there were very few electrons per cycle. I quickly realised that I knew nothing about how to work with this sort of thing, and decided to get some experience with DC first.

So after trying and failing to build a voltage amplifier front-end for a multimeter (noise issues), I breadboarded a basic transimpedance amplifier and found it worked. So I built a the same thing 'properly' in a diecast box, based on the LMC660 quad op-amp. The unit was intended as a throw-away prototype experiment, and no effort was made towards low noise, precision or anything else, other than common sense around minimising leakage currents at the inputs and keeping the display bit outside the screened area.

Because I had four op-amps in the package, I built two identical copies (though one has a much longer time constant), each with a front-end that outputs 1mV for 1pA input current. Each one also has a switchable x1000 amplifier stage, which should cause 1fA to generate 1mV. With a 3½ digit display reading up to 200mV, that meant a full scale of 200pA and a resolution of 0.1pA - or fA with the x1000 switched in.

To my astonishment, after a little careful offset adjustment, the readings would settle neatly to zero (pA) when open circuit. It calibrated to within about 10-20% using a 1G resistor in series with a 100mV source to produce 100pA. This was in keeping with the 'ordinary' low precision components used.

With the x1000 stage switched in, still with nothing connected, the reading eventually (after several minutes) settles to a fluctuation +/- about 5fA, but with a few fA offset and a very slow drift. With an inch or two of copper wire hanging out of the input BNC socket, the fluctuations in an open environment would settle after a while to +/- a few hundred fA. But in a car, well away from power cables and houses, and the windows shut, it would settle eventually to +/- 15fA or so. Out in the open air, the reading is permanently end-stopped. In the new shielded workshop, behaviour matches that in the car. However, if I move a muscle, move my foot half an inch on the floor or whatever, even from several metres away, the reading will fluctuate wildly and only settle again after thirty seconds or so has passed.

Next steps include

• build something to let me calibrate this with more confidence
• investigate noise sources (e.g. thermocouple effects from air currents, Johnson noise, temperature drift, static electric fields from clothing etc.) and find ways to minimise them and hopefully work around them so the device becomes more useful in a non-shielded environment
• see how far I can push this design, given the 2fA input bias current of the LMC660 - how stable is this current - how well can it be offset to read values approaching or below the bias current - if I can get the noise down enough to make that worthwhile
• try to reduce the time constant to get faster settling, and start the learning curve towards eventually playing with tiny AC signals
• maybe think about a design with more stages and lower value feedback resistors, to reduce the resistor noise
• redesign and build a more precision, low-noise and outdoor-proof version, preferably with the ability to have a remote display so that no operator needs to be close to the device in order to read it! And add another digit to the display so that the fA range overlaps the pA one- will then give a full scale of 2pA/2nA instead of 200fA/200pA.
• start to think about small voltages instead of currents

I have a long and ever-increasing list of topics that I am hoping to investigate with this - some frivolous, some more serious. But the real goal is the learning journey that I'm travelling rather than the eventual results.

• Seeing if I can tell when a rain cloud comes overhead before it starts to rain, based on changes in the atmospheric currents
• Helping a friend who’s doing PHD research into freshwater crustaceans, and who would love to have a means of detecting these animals in a stream – apparently they produce a tiny electric field in the water which might be detectable, if they could be separated (probably in software) from all the other distractions
• Finding out if there is any truth to the urban myth from decades ago that waterfalls are full of healthy negative ions as an antidote to nasty unhealthy positive ions from cathode ray tube TVs
• Testing a hypothesis of mine, that on a wire carrying a current around a hairpin bend, there should be a tiny potential difference between the surface of the wire on the inside and outside of the bend – the outside being negative. I don’t know how to start to calculate the potential difference that might arise across the width of the wire though, or what current it might sustain. However, I’d like to try to measure it – it might be within my reach. But this one might be a topic for a different thread!

My time is going to be reduced soon though, as I've signed myself up for a university degree course in mathematics and physics, which starts in a few weeks. Perhaps in years to come, I'll have a more fundamental understanding of all this stuff!

[SeanB]

If you want a cheap fan try a local air conditioning company and see if you can get a used indoor unit for a ducted unit. Then you strip out the coil and mount it with ducting as needed. This has a good low noise variable speed fan and ducting in and out as well. Plus a 24VAC transformer and relays for selecting fan speeds as well.

[Lurch]

In line centrifugal fan. Higher airflow than a standard wall mounted axial fan and also can be located away from the cage so should be quiet in the work area regardless. Decent ones aren't that noisy though, but noticeable.

Thank you - I'm searching over the next few days and will look out for that type. Currently looking at types aimed at commercial kitchens rather than domestic bathrooms, mounted separately from the area being ventilated, claiming around 37-44db noise and shifting over 400 cubic metres / hour - roughly four or five times as much as a typical bathroom fan. They are a bit more expensive but seem not too bad if you shop around.

I'm guessing this is the kind of thing you're thinking of?  This is a much larger space than the average bathroom, and I want to be able to shift the fumes from solvents and soldering as well as adding to the general ventilation.

Yeah, something like that. Not sure where you are exactly but something of that performance is probably easily obtainable from a decent electrical wholesaler. For instance a Manrose 150mm in line centrifugal fan moves 420m^3/h @ 51db. Spend a bit more and Vent Axia do a quieter one with similar performance. No need to go too exotic. Not sure about fume extraction though, not had much luck with standard fans and filters for that. Be interested to know how you get on with this bit of the project. I found it makes a lot of difference deciding on what ducting to use, ideally you want a smooth internal wall, especially for distances over a couple of metres. The standard flexible ducting is pretty much useless for this. Of course if you don't bother with the 3 stage filtering it probably helps!

By the way, I do have vents through the wall, allowing outside air to be drawn in by the fan - since there are few other ways for air to get in when the window is shut!

Valid concern!

[lesaid]

For instance a Manrose 150mm in line centrifugal fan moves 420m^3/h @ 51db. Spend a bit more and Vent Axia do a quieter one with similar performance.

Thanks. Manrose and Vent Axia were the two makes I was looking at - and a wide range are stocked by several local trade retailers. Though I suspect that once I know what I want, I might be able to shop around for them more cheaply on Amazon.

I will only need at most 1M of ducting between the fan and the workshop vent though it will have to go around a right-angle bend, and I don't see why I should need any filtering. The soldering and solvent fumes are nothing unusual and would normally just go out of the window. I'm only worried here because of the chronic lack of ventilation in this workshop, and the wish to keep the window shut during the winter frosts!!  And as things stand, I can't open the window while the gauze shield is in place over it, though I'm about to try a redesign that will allow me to open and close the window without having to take the shield on and off all the time!

I've got a lot on for the next couple of weeks so it will be into October probably before I can take this forwards, but I'll post an update as soon as I have one.

If you want a cheap fan try a local air conditioning company and see if you can get a used indoor unit for a ducted unit.

I'll scout around and see what I can find - I hadn't thought of that. Though a lot will depend on how easily it can be mounted on a ceiling and how much work is required to adapt it. I am very space-constrained!  Thanks