Most relays on the market have minimum coil activation power of 100 mW (or in mW range) and loads rating from micro Watts to Watts. I am looking for a relay that has coil activation power in low microWatts range and loads range of mW. Has anyone run into that kind of relay? Relay have a property of being full on / off - no leakage. Unless there is an alternative circuit that simulate its relay properties very well - nanoWatts losses/leakage is acceptable.
Adam,
It would be best to either describe the application or indicate that amount of leakage that is acceptable. If you are simply trying for DC isolation, then then you probably want to look at solid state relays (e.g. LH1529AAC-ND from Digikey).
For RF isolation, the choices will be much more application specific.
Adam, I am designing field energy harvesting systems in my research and had similar questions and found an interesting solution which I would like to share with you. Instead of focusing on POWER, I started focusing on ENERGY.
Check out this BI-STABLE relay from Fujitsu: AL5WN-K. It is a latch that needs only 6ms of 100mW power to latch, and once it latches, it is latched forever (sometimes these relays are called LATCHING RELAYS). To latch it in the other direction, you have to apply the energy in the other direction, only for 6ms again ...
This means that, although you are consuming 100mW power during the latching, it is so short that, its ENERGY consumption is only E=100mW*6ms=600uJ (600 micro Joules, less than a milli Joule). This, of course, would only work in an application where the switching back and forth is very rare. In my application it was ...
So, let's say, you are only switching once every minute. Effectively, your switching activity is 6 ms every 60 s. So, you are only consuming 600uJ in 60s, which makes your effective power consumption 600uJ = P * 60s ==> P=10 microWatts. There is your micro Watt relay :)
Hope this helped, or gave you some useful ideas ... Tolga
Thanks Tolga,
Your answer is very interesting. I will look into that with enthusiasm and see if it solves my problem. Yes, my application is energy harvesting and it is a stationary system. Furthermore, I will try relays you listed and add some new designs. In return, I would like to share solution that I have not tried yet (I have to build it and I am waiting for parts right now). My solution is to use an analog ammeter that has a conducting moving arm (25 uA range scale) . Power-wise it is very elegant however, every solution comes with trade offs. It has a slow response to input signal and it needs to be stationary which is not a problem for my case.
In energy harvesting, the bi-stable relay solution should definitely work. You will need a little bit of engineering to get the relay working in both directions. There are also DUAL COIL versions of the same relay. One for SET and one for RESET. This is a lot easier to get to work. If you use the single coil one I told you about, you need two sets of NMOS and PMOS transistors to SET, RESET. You completely lost me on the moving arm, and ammeter. Typically, current measurements are done by using a SENSE RESISTOR. You determine the size of the sense resistor by figuring a 25 to 50 mV voltage drop. The simplest ever solution ... Then, you use a CURRENT SENSE AMPLIFIER to convert the little voltage drop in the sense resistor into a single-ended voltage by using a current sense amp like MAX9634, which only consumes 1 uA to do its job ! Are you trying to measure current ?
If you look at the old analog multimeter, they have analog panel display that has different scales for measuring whatever you wish (voltage, current or resistance). I decided to look and buy the analog panel for a specific job, in my case, it is ammeter that has 25uA scale range or less with build in shunt resistor. The moving arm is non conducting therefore it needs to be painted with a conducting paint. When a current is sensed , say 10uA, it moves the arm to display 10uA. There, I install two pins each connected to a wire. When arm hits two pins, two wire are shorted or connected just like relay. As I mentioned before, the compromise is that this method has slow response, needs to be stationary and is bulky. However, power-wise it is elegant and efficient. Another trick is yours that has its own advantages. I will do both since I already ordered parts for my solution and yours is interesting enough that it will keep me awake every night if I do not try it.
Oh, Ok. I got you now. Very cool. So, you are using an analog ammeter as a relay. What are you trying to "switch" ? Why not use a transistor ? IF you can describe your app, I could give you a solution. Answer a few things please :
1) Do you need "isolation" between the switching signal and the load
2) How much is the load ?
3) How much is the switching signal ?
I have a weird feeling, all you need is a transistor ?
Hi Tolga,
1) Yes
2) Load is equivalent to a battery, D cell, AA cell etc.
3) Switching signal will be as lowest as I can make. Preferred 100uW or less.
Transistors leak current. They have (switch on) resistance. For example, if I am charging battery from capacitor using capacitors and using a transistor as a switch in between, I noticed incomplete power transfer. Furthermore, (when switch is off) transistor off mode should be perfectly isolated which is not. That is why relays have some benefits that transistor do not have. I know relays have a come back trend in mem R & D. I know that GE developed some MEMS relays and I talked to a developer for details unfortunately he said that they developed relays for their own products only. RADANT MEMS company developed some relays too, however they are not suitable for my application. So far, I am sticking to analog panel and latching relays. Its good to try many solutions and pick the simplest and efficient one.
Hi Tolga,
Looks like I will be trying your approach soon. My idea of paint on analog panel moving arm does not work. There are many reasons for it. I used a nickel print and test it on masking tape if conductivity works. It works but not perfectly. It takes couple coatings. When painted on arm, the paint becomes brittle. And it is hard to make contacts align perfectly (manually). On top of that, I noticed something odd when doing this experiment. I know that aluminum foil makes a good conductor by measuring resistance with two probes of multimeter. But when I put that same aluminum foil on top of two probes, I get open circuit. However when I press gently the foil ,sitting on those probes, then multimeter indicates short. Aluminum oxides coating was a cause of insulation effect? What material should I be using to make a good conductivity without using force on probes (just a surface contact)? My guess is that I should be using silver coating.
Just FYI, Aluminum is a nightmare to solder. You will never be able to solder it ! You have to use COPPER. This is the only material that solders well. When you solder it, it soaks up TIN (Sn) which is the only material to reliably use. The only place where Al is useful is COOLING something. DO NOT USE SILVER ! IT is impossible to solder. melting temp is very high.
I am a little lost on your PAINTING ? approach ? I am having trouble visualizing.
What are you trying to really achieve ? If you are trying to magnetically switch something, what the industry continuously uses is REED RELAYs. They are super reliable, however, you will clearly not get uW. They are more like high mW.
If you look at the image of the analog panel meter for example, see link for image:
http://img.hisupplier.com/var/userImages/old/precision/precision$715163516.jpg
The moving arm is non-conducting which is good for my case. I paint the arm with nickel print by MG Chemicals. That is suppose to make this arm conducting. I set/glue two pins at 50uA location so that when I apply 50uA of current say at 2V , the arm will move to 50uA location and hit two pins thus creating a short because the arm was painted to be conductive then. Each pin is connected to each wire. The whole thing function as a relay or current analog panel meter relay or (CAPM Relay) I thing they have been made before in 60's but then they were abandoned. The neat thing about it is that this switch creates a perfect open/close and at micro-amps range. Of course, it has its own disadvantages too.
Ok. Just wanted to mention something : REED Relays. These relays are used in DOOR contacts (magnetic) etc ... I don't really have an idea at this point about the magnetic field that is necessary to move them, but, they are also mW range, and according to my very first idea, I don't know if there is such a thing as REED LATCHING RELAY ? I figured, if you check them out, it might give you ideas ...
I totally know the pure-analog type uA'meters ... In fact , I used them in my undergrad studies, back when the pure-digital equipment wasn't so widespread.
Could you please describe the FUNCTIONALITY you are trying to achieve ?
Adam and Tolga,
I would like to join in a little unusual way your interesting (for me) discussion by telling an incredible story from my life in 80's that is closely related to this issue...
At the beginning of 1984 I tried using a current reed relay (having a coil with only a few windings of thick wire) as an AC current consumption detector. To increase the relay sensitivity I placed a small magnet near the glass capsule and began slowly varying its location. At one moment, the vibrating reed contacts got stuck in the closed position. Of course, I was initially disappointed by that harmful effect but I soon realized that the element obtained a wonderful feature -- a kind of memory. When I aimed one pole of an external magnet at the one side of the capsule the reed contacts stuck in the closed position and remained in that state even after I had removed the magnet. And vice versa, when I aimed the same pole at the other capsule side the reed contacts opened and remained in that state even when I removed the magnet again.
What a wonderful element that was! It consisted of only a reed capsule and a magnet. Nevertheless, this Magic Switch simultaneously incorporated sensitive bipolar magnetic sensing, nonvolatile memorizing -- due to its hysteresis -- and switching! So I immediately applied it to magically arm/disarm the burglar alarm in the car that I had just bought...
Nevertheless, I managed to increase the reed relay sensitivity by an additional AC magnetization. Then a new idea suddenly aroused -- to affect an internally AC pre-magnetized reed switch by another external DC magnetic field. It turned out to be a wonderful element now consisting only of a reed relay and an adjustable resistor! But it had all the functions of a supersensitive bipolar magnetic sensing and pulse-length regulation.
I used to carry out the experiment of a magnetically controlled DC motor which deeply impressed my colleagues. For the purpose, I connected a small DC motor via the reed contacts of the relay to a reducing transformer. Then I supplied the relay coil via the variable resistor from the transformer and adjusted the resistance so that the contacts began tapping slightly. After that I hid the "device" in a big box and became a "magician". I aimed the one pole of an external magnet (my "magic wand") at a distance of about 30 cm from the box -- the motor started to turn slowly. After that I began moving the magnet slowly -- the motor changed speed. Then I aimed the other pole of the magnet -- the motor reversed the direction of the revolution...
Maybe you can find a relation between my 84's invention and your problem?
Regards,
Cyril
REED Relays are good for LOGIC responses ... They either TURN ON, or NOT TURN OFF. For more of an analog response, I absolutely love the CURRENT SENSE transformers. However, they are for very high currents , multiple-Amp. But, although I never got around to it, I meant to design a current sense transformer our of a very small inductor coil to sense the mA range. The beauty of these passive devices: No energy consumption until you use them. And, their BURDEN is pretty minimal ... However, Adam needs a uW device. So, keep talking ... We will find something Adam can use ...
Thank you Mechkov for joining this discussion. To answer Soyata's last question here is the link of something equivalent I want to make:
1) http://www.simpsonelectric.com/images/File/330020ADC.pdf
2) http://www.simpsonelectric.com/images/File/datasheets/3300ruggedseal_datasheet.pdf
However, I think that it is not passive meter relay. It uses external power source to power its internal relay. It is something like that that I want to make. I already have a 50 uA range analog panel meter ($15). I will go back to it again because I got some new ideas to test. Furthermore, I looked at latching relays and for me to use it, the requirement are such that it should consume about 20uJ (per 1 second cycle) or less. This is based on what I can achieve with my approach (analog panel if it will work). I can work with 50uJ per cycle. I found latching relay of 40 mW nominal coil power rating at best. Unfortunately, it require 5 ms pulse (if I am correct) to switch on/off and that is 200 uJ or 400 uJ per cycle which is too much. It will require a 0.25 us to meet criteria which is impossible.
Another approach is to use a very sensitive reed relay and put it in a close proximity of analog panel meter coil that turns the dial. If emf created by coil is strong enough, it should turn on the switch. Unfortunately I know nothing about reed relays and how strong magnetic field should be to turn on the switch. That's going to take some time too.
The first device is talking about TRANSISTORS etc ... inside the voltmeter. So, clearly, there is an ACTIVE circuit inside ... Your requirements for turning on a switch are very low power ... IT is looking to me like, it is time for you to explore MEMS !!! There are MEMS IC's around, definitely ... I haven't played around with them, though ...
I have looked into MEMS relay and there are few of MEMS and none of them are in my specs or in a way that can be utilized. See examples below:
1) http://ca.mouser.com/new/Omron-Electronics/Omron2SMES-01/
My maximum switching load is about 5V and 5mA. I can utilize it but then I have to add other sub circuits just to accommodate that switch.
2) http://www.cotorelay.com/redrock.php
I've got silver wire and it gives me the same negative results as ordinary wire. Then I did some research and learned a concept of wetting current. Below is the link to a definition of wetting current:
http://en.wikipedia.org/wiki/Wetting_current
The source of problem with my ammeter panel is that uA current I supply is enough to move dial's arm but not enough to provide CONTACT PRESSURE. I did a simple test by attaching one inch of 26 AWG sterling silver wire to each multimeter probe. Then I just drop a third silver wire on those other two. It simulates a low pressure contact well. Multimeter shows no conduction unless I press the wire. Of course, it could be just poor surface contact. So I wetted that wire with salted water. It worked occasionally. Then I though about the mercury, but then I realized that I am making things more complicated. I prefer a simpler approach. Currently I am running out of ideas of making good conducting contacts at low contact pressure. Any ideas are appreciated anyway. Now I am focusing on Soyata's suggestions which is latching relay. I am aware of very low power amps and timers to build controlled impulses to switch on/off latching relays. The best low power latching relay I found is 34 mW at 5mS pulses.
Adam, wow ! that's a wealth of info. First, this MEMS relay looks awesome. Exactly what you want with 30uW !
The 34mW and 5ms comes to about 150uJ, which is not bad, but, higher than your target 20-30uJ mark.
The wetting current is very common for relays. Almost every relay specifies a "minimum current to cause latching". However, I highly doubt that, this is the case for semiconductor relays, like the MEMS one you showed
OMRON - $192
http://www.digikey.ca/product-search/en/relays/solid-state-relays/1048664?k=mems
Coto - RedRock - $ 71
http://www.digikey.ca/product-search/en/programmers-development-systems/evaluation-and-demonstration-boards-and-kits/2622039?k=coto
Coto is cheaper. OMRON has an input (to activate switch) in my specs but switching loads are too low for my requirements. On top of that it is way too expensive.
Coto has an excellent load specs but inputs (to tun on the switch) is in mT. Which means that inputs are magnetic. I don't think that I can get 25mT by running uA range current under less than 5V through the coil. Still, the price is steep for hobbyist. Nevertheless, Coto looks promising.