Register to reply

Would this relay design be useful?

by Curl
Tags: design
Share this thread:
Curl
#1
Jul8-14, 02:20 AM
P: 757
Hi
I had a neat idea for a relay design, basically its a relay that takes very tiny movements to open/close a circuit, I.e. I can make a 10 micron movement to create/close a 1mm gap very easily and in a small package (3mm by 10mm by 20mm say) and with a fairily large contact area (for that size it would be at least 100mm^2).

I could probably use a piezoelectric material to actuate it at pretty high frequencies, it looks like it should do 500Hz no problem.

I want to know if there would be use for such device, I'm a mechanical engineer and know almost nothing about relays except that they cost too much. It is mechanical so it will probably have some bounce, and just to be clear it doesn't simply use a mechanical advantage (like leverage) to create what I said.

If I am to build this it would take me some time and I want to know if it's worth it, I don't want to waste time making something that doesn't impress anyone. Also can someone let me know if something like this (or better) already exists? This is usually what happens when I design something, I find out someone has done it before and probably done it better

I appreciate some advice from those of you experienced in the field. Thanks!
Phys.Org News Partner Engineering news on Phys.org
DIY glove-based tutor indicates muscle-memory potential
Tricorder XPRIZE: 10 teams advance in global competition to develop consumer-focused diagnostic device
Study shows local seismic isolation and damping methods provide optimal protection for essential computing equipment
Simon Bridge
#2
Jul8-14, 02:58 AM
Homework
Sci Advisor
HW Helper
Thanks
Simon Bridge's Avatar
P: 12,961
Necessity is the mother of invention.
What is the problem you are trying to solve with this invention?

You should be thinking of this first - then think of the inventions, otherwise you are doing art.

There are already ultra-small relays: have you tried canvassing manufacturers to see what products are already on the market that fit your invention's parameters?

i.e. http://www.omron.com/ecb/products/pdf/en-g6j.pdf
Surface Mount DPDT Dim. WxLxH: 5.7mm x 10.6mm x 9mm

Clearly there is a market for small relays and a cheaper way to build one with similar specs may be competative.
You are quite a ways from market research though.
jbelectric777
#3
Jul8-14, 05:00 PM
P: 5
Sounds like a marketing idea.... So you know, they already make a W117 PCB (1000 ohm coil) relay/relays. Not sure if its w171 or w117 but theyre small and work. Is that about what you have?

thankz
#4
Jul8-14, 07:50 PM
P: 60
Would this relay design be useful?

if your looking for close contacts a reed relay might suit your needs
jim hardy
#5
Jul8-14, 08:55 PM
Sci Advisor
PF Gold
jim hardy's Avatar
P: 3,684
I could probably use a piezoelectric material to actuate it at pretty high frequencies, it looks like it should do 500Hz no problem.
Your relay might solve common mode troubles in analog multiplexing equipment.
What kind of contact life do you anticipate ? At 500 hz , 10^8 operations could take place in ~2 days...

For our plant computer multiplexer we used mercury wetted relay contacts that sampled once per second for thirty years, around 10^9 operations. They were still doing fine when the system was replaced due to obsolescence.
Baluncore
#6
Jul8-14, 09:55 PM
Sci Advisor
Thanks
P: 1,923
Relays are a liability, not because of the actuator, but because of their contact life.
Relays have now largely been replaced by semiconductor switches and opto-isolators.
Being inexperienced in a field tends to lead to excitement over reinvention of the wheel.
Cheers from the Devils Advocate.
Curl
#7
Jul9-14, 12:46 AM
P: 757
The reason I thought about it was because I was working on a project where I wanted to create a small spark using an ignition coil off a car. It worked fine using a normal switch except I noticed I got 2 sparks, one in the switch and one across my gap, so I needed a way to open the switch much faster.

What my design does is basically multiply the gap distance that you create. With a classic switch if you move it a distance X, then the gap for the current to jump is X. With my design, if you move the "button" a distance X, the gap for the current will be 100X or 1000X or whatever you want to build. And I repeat, it's not a reed or some sort of lever - the contact area can be arbitrarily large and there is no material deformation at all i.e. nothing to twist/bend.
The velocity is therefore multiplied too, so therefore you can create an appreciable gap to the current in very short times just using your finger. I stopped working on that particular project for another reason and so the my relay (I called it a switch) never got made but I remembered the idea (this was about 2 years ago) and thought it might be useful for making a relay.

If not I won't worry about it, I just didn't want to spend 40 hours of my graduate student time trying to build this thing if nobody cared about it.
Baluncore
#8
Jul9-14, 02:01 AM
Sci Advisor
Thanks
P: 1,923
Because you have not actually described you relay mechanism, we can only advise in generalities.

It is quite normal for an inventor to believe that their invention will be worth a great deal of money and that it must be kept a secret, but the original idea is worth only about 2% of the final product sale price. It is unlikely that it would be economic for you to patent it, or to defend the patent in the courts if you were challenged. As a guide, a full patent specification will cost about $10k to file and $100k minimum to defend. It will also cost you, in each country, annual fees of a few hundred dollars to keep it active.

For that reason I would suggest you fully disclose your idea, maybe here on the PF. That will prevent others from patenting it since it is no longer novel, it is prior art. It may be brilliant, in which case public disclosure will identify you as the inventor. It will give us the chance to identify if it is novel, available in the market or as a prior publication or patent.

To hedge your bets you could first apply for a provisional patent, then publish. That gives you 12 months to file a complete specification if it turns out to be brilliant, novel and marketable.
Curl
#9
Jul9-14, 03:01 AM
P: 757
Baluncore I never said anything about making a billion off of this; it's a very simple idea and I'm 99% sure that someone has already done this "trick" in design.
All that you said is true and I probably know it better than you, I don't get how you got the sense that I'm all in love with the idea. Truth is it took me about 5 minutes to think about this design and all I want is someone experienced in the field to tell me if it has any worth because right now I think it's a waste of time to attempt building it as I am busy with other things. I could also learn a thing or two about relays in the process.

That's all this is, I think I've been rather clear...
Simon Bridge
#10
Jul9-14, 07:22 AM
Homework
Sci Advisor
HW Helper
Thanks
Simon Bridge's Avatar
P: 12,961
That's all this is, I think I've been rather clear...
No - you have not described the relay at all ... you have not, for example, stated how it is that you get this multiplication of motion, only how you don't do it.
Is it a secret?
jim hardy
#11
Jul9-14, 07:39 AM
Sci Advisor
PF Gold
jim hardy's Avatar
P: 3,684
I wanted to create a small spark using an ignition coil off a car. It worked fine using a normal switch except I noticed I got 2 sparks, one in the switch and one across my gap, ...
i trust you used a condenser across the switch contact....
Baluncore
#12
Jul11-14, 01:34 AM
Sci Advisor
Thanks
P: 1,923
Quote Quote by Curl
Baluncore I never said anything about making a billion off of this;
Nor did I.


Quote Quote by Curl
All that you said is true and I probably know it better than you, …
I have been counselling inventors while assessing their ideas, prototypes and psychology for over 35 years.
So I doubt that you know it better than me.


Quote Quote by Curl
I don't get how you got the sense that I'm all in love with the idea.
Your apparent unwillingness to detail the key mechanism involved suggested to me the characteristic behaviour of a paranoid inventor, one who holds their cards so close to their chest that they will never be able to play them.

Paranoid inventors usually claim to have an idea that is going to make them a million dollars, an idea so valuable that it must always remain a secret. Others are just good story tellers, they aim to maintain the suspense for as long as possible, because that is the only way they can get the control and attention they need.

@ Curl:
You can demonstrate that my interpretation of your reticence is wrong.
The best way to do that is by describing your secret mechanism here.
Curl
#13
Jul11-14, 02:03 AM
P: 757
The reason why I did not show the details is because it's irrelevant. I came here to find out if a relay like this is useful, I didn't come here to get a design review.

The customer doesn't care how the device works they only care about what it can do, so I want to get opinions from that perspective.

If you're really that curious about what I thought, all this is essentially is stacking 1000 switches in series, and one motion opens/closes all the switches. So the motion required to close 1 switch is the same as for 1000 switches except now since they are in series the gap is 1000 times that of one switch. There, that's the idea. I don't even have any drawings of this and I don't know if its even worth my time to sketch some. There are also many ways to implement this and I haven't even decided on anything in particular. I just want to know if it's worth my time to think about this further, that's what I came here to ask.

I get ideas like this at least once every week so you must trust me when I assure you that I am not "paranoid" about it as you believe. I've been through this hundreds of times.
rollingstein
#14
Jul11-14, 03:18 AM
PF Gold
P: 330
Quote Quote by jim hardy View Post
Y
For our plant computer multiplexer we used mercury wetted relay contacts that sampled once per second for thirty years, around 10^9 operations. They were still doing fine when the system was replaced due to obsolescence.
Out of curiosity, what sort of system was this. i.e. Can you give more details of its function / design etc.
sophiecentaur
#15
Jul11-14, 06:12 AM
Sci Advisor
Thanks
PF Gold
sophiecentaur's Avatar
P: 12,186
Is this a relay or a switch that the OP is describing?
If, as he says,
I don't know if its even worth my time to sketch some
, it is hardly worth our time in sketching wrong pictures in our heads and answering questions that are not actually his question. Switches, as they connect or disconnect, behave non-linearly so the situation with multiple switches in series is very dependent on actual circumstances.
Baluncore
#16
Jul11-14, 08:29 AM
Sci Advisor
Thanks
P: 1,923
Quote Quote by Curl
The customer doesn't care how the device works they only care about what it can do, so I want to get opinions from that perspective.
What it can do depends on how it is implemented.

When in series, the contact resistance will be 1000 times greater than a single contact.

10um vibration is very common. I expect the contacts will be microphonic.

The reliability of the individual contacts will need to be very much higher than today's contacts.
If the reliability of each contact was 99.99% then having 1000 in series would give only 90.4% system reliable.
99.9999% reliable contacts would be unachievable, yet still give only a 99.9% system reliability.
That would certainly be unacceptable in the electronics world today.

The breakdown voltage of a contact in air is non-linear. For small gaps in series the sum of breakdown voltages will probably be significantly in your favour, but only if the voltage can be distributed evenly along your string of contacts. That will probably require a string of R and C elements across the series contacts to balance the DC and AC voltage drops.

Weird things can happen in series strings of switch contacts in a gas. It is counter-intuitive but the breakdown voltage over two switches, completely jumping the intermediate conductor, can be lower than the breakdown voltage of the two switch gaps in series. It is for the same reason that a spark does not form directly between two electrodes, but instead it will take an initial circuitous path of a geometric “arc”.

For small gaps in series or parallel you have a Marx Impulse generator trigger, where the UV flash of any breakdown will ionise all nearby contacts and so trigger them also. Is that an advantage or disadvantage?

Quote Quote by Curl, in post #1
… and with a fairily large contact area …
Here I assume it was proposed that the contact areas are being used in parallel. The advantage of huge switch areas is to conduct huge currents. The critical operation then becomes turning off the current in the presence of circuit inductance. But when multiple small contacts are employed in parallel, the last switch element(s) to open will arc and so be vaporised. I expect that each time the switch opens, more contact elements will be consumed until there are none left.



MOSFETs have now largely taken over the DC power switching field from relays. TRIACs continue to be used in AC phase control power systems, but they are being gradually replaced by PWM power factor correction systems using MOSFETs and IGBTs.

Switching very small RF currents often uses multiple gold plated fingers to reduce the resistance and impedance mismatch through the switch. Mercury wetted contacts were used previously for high reliability, low noise, low voltage signals. RF multiplexers now employ semiconductors. They are often embedded in the LNAs used for RF signals.

Time domain multiplexing is now used for digital telephone switching systems. Relay contacts in telephone systems were robust and could withstand the significant electrical currents induced in the overhead lines by atmospheric discharges. Optic fibre has now eliminated exposure to that environment.

I do not see many applications remaining for relays with electrical contacts.
But there may be an application for such an unusual distributed contact system.
Possibly consider implementation in a MEMS system.
Necessity is the mother of invention. Where would such a relay be advantageous?
jim hardy
#17
Jul11-14, 08:48 AM
Sci Advisor
PF Gold
jim hardy's Avatar
P: 3,684
Quote Quote by rollingstein View Post
Out of curiosity, what sort of system was this. i.e. Can you give more details of its function / design etc.
it was a power plant monitoring computer
that scanned and recorded a few hundred analog signals for datalogging, post trip analysis, and heat rate monitoring

some channels were scanned once per second, most less frequently
process signals are typically one to five volt unipolar, almost DC. One hz is a fast moving process for us. We also handled millivolt signals from thermocouples

The signals might have as much as forty volts DC common mode, so the relays connected a floating and guarded amplifier and ADC(14 bit) to them one at a time. Three pole relays- hi, lo and guard.
Most FET multiplexers can't handle common mode voltage more than about ten or fifteen volts so the mercury wetted relays were a great choice .

Also, FET multiplexers become electrically "soft" when they lose power so they can short out the signals to which they're connected, and it's sure embarrassing when a computer failure upsets the meters that the operators are using to control the plant.
A relay gives a mechanical disconnect on power fail - so it's Graceful Degradation", or " Fail Harmless".

Interfacing the computer to a plant is the most challenging part of such a project.

We needed no particular speed so used a dual slope integrator ADC with integration time of one line cycle for maximum line frequency rejection. Works great.
Another approach is to sample the signal with a much faster ADC perhaps four times in a line cycle and average them on the premise that since the average of line frequency sinewave is zero, symmetric sampling should cancel it out.

Another way around common mode voltage is to use "Flying Capacitor" , a capacitor tracks your input signal and is switched by a two pole double throw relay arrangement from the signal to the measurement equipment.

All those relays are expensive and slow. Your gizmo might be of interest to manufacturers of computer interface equipment = a fast relay offers fail safe isolation for critical applications.

RTP in Ft Lauderdale has been a leader for decades and their equipment is renowned for its reliability. We ran their 7430 front end from 1973 to around 2002. Analogic too has been around, you'll find more suppliers wit a search..

Your relay might fill a niche between high speed FET multiplexers and molasses slow relay types.
And it offers the graceful power fail behavior noted above.

good luck -

old jim
jbelectric777
#18
Jul11-14, 12:08 PM
P: 5
opti couplers and mosfets are the way to go but it sounds like he wants a form "A" single pole single throw dry set of contacts, transistors of any kind wouldn't work for him


Register to reply

Related Discussions
Can a relay use AC or DC? Electrical Engineering 15
Do i need a relay? Electrical Engineering 3
Help me in relay Electrical Engineering 5
Help with Relay circuit Electrical Engineering 3
Contactor vs. relay Electrical Engineering 1