Anyway, I suspect these relays will work fine for my zero energy switching application.
Mine haven't arrived yet. But I will thoroughly abuse a few to see how good they are.
BTW, another method of making low energy switching work is called contact "Wiping", I think. Lots of relays do this. As the armature moves the contact slide over each other a bit to rub off corrosion products. You can see this in many open frame relays if you close them with your finger.
Duane J
Duane J
Thanks for sharing your previous experience, it's very good to learn those things, from one of The Sources.:nerd:
In fact, I was also concerned with the possibility of contact adhesion, due to the intended extra-specs monodirectional current inrush.
I'm not sure if the gold clad contact finish will be enough to ensure it will not happen.
I'm also aware of the "Wiping" effect of the sliding movement of contacts, but, in the present case, as they are miniature relays, originaly intended to up to 2 A currents, I'm not sure that the total sliding movement and tangencial forces involved, will both ensure that adhesion will not occur or can be healed by that strategy.
The twin crossbar design of the contacts can be helpfull in achieving this, but as you stated in your post, thoroughly abuse a few parts will sure let you confident about their performance.
I'm sure we'll soon heard a lot about this, from you!
I got mine yesterday afternoon.
I soldered it to a 12 pin dual row header. The pins on the header were bent out so I can plug it into a plugboard.
I wired up an oscillator
This oscillator runs, on my relay #1, at about 380Hz.
So far it has logged about 186 million cycles and appears to be going strong. (Started 13/11/18 10:30pm)
The 122Ω relay coils in the oscillator draw about 41mA of current each.
I am stopping this test.
Note! I used a RED LED as the snubber which collapses the field a bit quicker than a silicon diode and it flashes when the magnetic field is collapsing.
Clearly a snubber is required though to prevent fairly high kickback voltages.
An observation:
I can't find in the spec where it says the coils have polarity other than in the schematic where the coils are labeled with a + and -.
There is no built in snubber diode, as in some relays, so one is allowed to operate the coils in opposite polarity.
However, if I reverse the 5V to -5V, and reverse my snubber LEDs, the oscillator locks up.
To make it work the contacts need to swapped. Whats going on?
I think the relay armature is a permanent magnet.
So, the relay can be switched back and forth with only one coil by reversing its polarity.
The other coil does the same thing, just the opposite way.
Or it can be driven with a 4 MOSFET bridge. MOSFET bridges have snubbers built in.
This is interesting!!!
This relay can be switched in a number of different ways.
What a bonus!!!!
And being your oscillator's frequency 380Hz, or 22800 cpm. is about 126,67 times the 180 cpm value, used by Panasonioc to determine mechanical lifespan of their product.
Checking the operate and release time (with diode) shows values between 2 and 3 ms, and your data is showing us about 1.3 ms for each operation, a 100% improvement over factory data.
Outstanding!
Thanks for the mild current flowing thru the contacts in your test fixture, its electrical life isn't being yet hardly challenged, since those tests were conducted at 20 cps at their lab facilities, but in a 2A@30Vdc resistive load landscape.
Perhaps a slightly higher than normal temperature raise could be observed, due to the stress test conditions, but if things are not failing nor melting, you are in the right direction.
Our eyes and ears are poor observers, until near to catastrophic failures are about to happen.
I must agree with you, except for the diagram + and - markings, there are no references to the armature construction, but your thought about a permanent magnet makes perfect sense.
When you analyze the test diagrams (9, 10-1 and 10-2) and also wondering about the valuable adition to electromechanical hysteresis represented by a permanent magnet armature construction, IMHO we could find the reason why they didn't explicity stated its internal structure till that point.
These relays will operate faster and at lower voltages in sustained oscillation than they will from dead rest due to armature rebound. There is a PDF with a very detailed explanation of how they work at Digi-Key's part page for the relay (as well as a giggle inducing price point).
With 2200 uF capacitors, 33 and 22 ohm charge caps, and a 10-ohm discharge cap, this thing is stable within less than 1 percent at 7 Hz. It can be enabled and disabled in several creative ways which sync it up to a leading pulse edge. I'm thinking of building an async receiver-transmitter pair; early drawings suggest this should require three relays per bit (two for the shift register and one latch) and at most two or three for control. And they should be able to send and receive to an ordinary computer serial port, how about that?
Don't mention it, UN. I got a great deal and I'm not trying to pad my retirement account. Let me know if you want more, I will probably have stock for awhile :-)
Well, they might be a bit smarter if they halved the quantity as well as doubled the price.. $100USD investment in a pile of relays is a bit awkward... And $200 USD is out of my ballpark completely.
It would even be more appealing if that had the doubled price and a $50USD quantity.... the last thing I need is go into stockpiling hundreds of dollars in hundreds of relays.
This somewhat changes my point of view on whose goldmine they are referring to ;-)
Yeah Bean, the sale was always tagged as ending on the 19th. It only ran for about a week and a half. $200 is the regular price, which is still a darn good deal.
Loopy, if you look at the pictures I posted on the previous page, you'll see that changing the quantity would involve a *lot* of labor on EG's part. The package is a reel intended for dispensing to a pick-and-place machine. They would have to unwind the reel (hopefully without releasing the loosely bound cover plastic), somehow count the relays out, cut and repackage them. It just wouldn't be practical.
However, I do have 500 which I got at the 22 cent each with shipping price, so if you'd like a few PM me or post here and we can work a deal.
Fair enough, the 40 cents each is a 'normal' reel price... anything less should have a higher price for the cost of carrying a partial reel. I just presumed the 20 cents sale was too cheap... these are listed under their clearance section. SO I got a mixed message.
What is normal at the Goldmine.. no latching relays? They seem wary to keep some on hand for small quantity sales. I'd happily pay $2 for single units. I see the more conventional 5VDC relays are over $1 USD in single units.
It is an old dilemma, electronic components are really intended to supply an industry that wants to sell thousands if not millions in one sale. Hobbyist and students can ready about a lot of nice stuff, but getting just one often forces us to pay a big premium or to grab up oversupply that comes and goes.
Okay, you all win 40cent versus $4. None of the local supply houses can even comprehend what a latching relay is. A year or so ago, I was looking for some 12a 120VAC latching relays... just to see. Such are not the 5VDC type, and run $15-20 USD each. They are really great for a centralized home automation panel next to the Mains for the home.
Regrettably, those may never become a clearance item anywhere.
I still am amazed that the 'general public' perceives the conventional 5VDC as a good solution for a micro-controller driven relay when the latching version can be far superior in conserving power to the microcontroller. With 5VDC latching relays, a battery powered microcontroller can go to sleep for periods of inactivity, momentarily wake up to check status and tweek a relay if necessary, then go back to sleep.
I love the concept, but really have no projects to build.
With 5VDC latching relays, a battery powered microcontroller can go to sleep for periods of inactivity, momentarily wake up to check status and tweak a relay if necessary, then go back to sleep.
You just described a fairly productive day in my life.
With 5VDC latching relays, a battery powered microcontroller can go to sleep for periods of inactivity, momentarily wake up to check status and tweek a relay if necessary, then go back to sleep.
The general public probably doesn't realize how many of these things they own. My battery powered digital thermostat uses a latching relay. Since it needs only a brief pulse to turn the air conditioner on or off, and the rest of the electronics are a low power microcontroller and LCD display, the batteries last more than a year.
In a few days I after sleeping on the final layout I'll be sending this off to ExpressPCB. It's arranged to mount the relays with their pins bent straight down for thru-hole, which results in a 0.2 inch row spacing, too narrow and the pins too short for most breadboards.
This bad boy is going to cost more than the relays did but will save me a bunch of time in the first round of experiments I want to do. It has mounting room for 34 relays, 16 LED's, 24 tact switches to form 12 nonvolatile latching inputs, and the components to form a free-running oscillator.
In addition the ground pins of all coils are bussed, and the common pins of each contact are bussed separately for each group of 10 relays on the right. The tact switches are wired to set the top relays as inputs, and the LED's are arranged to reflect top relay status with an easy resistor placement. The bottom-right 20 relays also form a shift register which can be disabled to use the relays separately by cutting some top traces.
That's a lot of wiring I won't have to do. The prewiring is arranged so that the board can be either a receiver, shifting bits into the registers and latching them to the top row, or a transmitter which latches bits in from the top input row and shifts them out. My plan is to make it compatible with a regular RS-232 port although at a low baud rate and with appropriate level conversion.
I'll probably order six of these because the one-time setup and shipping costs about USD $75, after which each board only costs around $20, and I might think of something else to do with the general plan or just want to make a neater pair once the experimental set is working.
Any suggestions? Or anyone want to buy in on a sample? I can also supply the ExpressPCB source file if anyone's interested.
It's arranged to mount the relays with their pins bent straight down for thru-hole, which results in a 0.2 inch row spacing, too narrow and the pins too short for most breadboards.
This is how I have mounted my test relays.
I used a 2 row 12 pin header. I pulled out 2 of the pins where not needed.
Then bent and formed the header pins to get 3 tenths spacing for my Plug Board.
As you did, I straitened the relay pins and soldered to the formed header.
BTW, this is my prototype solar tracker using the relays.
Am now attempting to get the PWM working.
That is a DS203 pocket scope in the background.
I really like it for hacking circuits together.
Sure, I have a high performance digital scope but that can't fit on my lap watching the TV.
Took a bit longer than I expected probably due to the holidays. Top board shows a relay seated to test the pin hole pattern in the spot pre-wired for an oscillator. Bottom board is flipped vertically to show reverse side. All prewired functions can easily be disabled by cutting a few traces to make every spot general-purpose. With those buses pre-run that's a lot of wires I won't have to breadboard, yay!
Looks great! And, just to fill out this form so that it will not be flagged as too short, allow me to say "thank you" once again for an outstanding bargain!
Yeah erco saw those after the goldmine email arrived today. The thing is, the double coilers are FAR easier to work with because for the most part you can ground the minus side of every coil and use a single contact to either set or reset. With the single coil you have to reverse polarity, which means two contacts are involved to set or reset and with DPDT that doesn't leave any more contacts for more complicated logic functions.
I also note these are 4.5V instead of 5V. I bet that's because if you're driving them with bipolar transistors, you need two transistor voltage drops (an H-bridge) instead of just one.
OTOH, since they latch, I think you can control more single coil relays with fewer pins. I'm pretty sure I can individually control 28 single-coil relays with the 8 I/O pins of a humble BS1, no extra hardware.
What's the minimum pins requirement for 28 dual-coil relays (no extra hardware)?
It would be true you could use the Prop itself as the H-bridge but it doesn't quite kick the 5V relays. 3.3V switches them but Prop pins don't because the resistance of the prop I/O drivers + only 3v3 to start with = no clicky. Might be different with the 4v5 coils on these single-ended guys (the double coil guys were had 5v0 coils) but I tend to doubt it.
IIRC prop pin drivers are about 40 ohms. The 5v0 coils are 125 ohms. They switch (barely) at 3v3 which means at a current of 26 mA. Add 40 ohms and you've got 160 total (worse 200 for an H-bridge, but let's just say 160) and on a 3v3 supply you've got 20 mA. They won't switch on that.
The 4v5 coils may be different but I doubt they're different enough to overcome the extra drop for two pin-drive transistors in series forming a reversible H-bridge.
No transistor h-bridge, pins connected direct to coils. 22 mA for a few milliseconds won't hurt anything. Use High-Z (input) to disconnect undesired coils, and only drive one pin high and one pin low at a time to switch a relay. Thus 28 different relays can be switched on or off when connected between pins:
Yeah Erco but you won't get 22 mA because prop pins add about 40 ohms resistance to the relay coil's 125 ohm resistance. You'd get 16-20 mA, and you need at least 25 mA to flip the relay. With nothing in series, 3v3 / 125 ohms = 26 mA. My tests show 25 mA is pretty much a minimum. You can make it oscillate as low as 20 but only once it's started by a higher current pulse, because the armature bounces. Below 20 mA even free oscillation doesn't work. The double-enders will not switch from prop pins at all, even with only one prop pin driving against ground.
This is all unless the 4v5 coils, which these new single coil relays have, are significantly different from the 5v coils in the double-enders. I tend to doubt they're *that* different.
Comments
Duane J
Thanks for sharing your previous experience, it's very good to learn those things, from one of The Sources.:nerd:
In fact, I was also concerned with the possibility of contact adhesion, due to the intended extra-specs monodirectional current inrush.
I'm not sure if the gold clad contact finish will be enough to ensure it will not happen.
I'm also aware of the "Wiping" effect of the sliding movement of contacts, but, in the present case, as they are miniature relays, originaly intended to up to 2 A currents, I'm not sure that the total sliding movement and tangencial forces involved, will both ensure that adhesion will not occur or can be healed by that strategy.
The twin crossbar design of the contacts can be helpfull in achieving this, but as you stated in your post, thoroughly abuse a few parts will sure let you confident about their performance.
I'm sure we'll soon heard a lot about this, from you!
Yanomani
I got mine yesterday afternoon.
I soldered it to a 12 pin dual row header. The pins on the header were bent out so I can plug it into a plugboard.
I wired up an oscillator
This oscillator runs, on my relay #1, at about 380Hz.
So far it has logged about 186 million cycles and appears to be going strong. (Started 13/11/18 10:30pm)
The 122Ω relay coils in the oscillator draw about 41mA of current each.
I am stopping this test.
Note! I used a RED LED as the snubber which collapses the field a bit quicker than a silicon diode and it flashes when the magnetic field is collapsing.
Clearly a snubber is required though to prevent fairly high kickback voltages.
An observation:
I can't find in the spec where it says the coils have polarity other than in the schematic where the coils are labeled with a + and -.
There is no built in snubber diode, as in some relays, so one is allowed to operate the coils in opposite polarity.
However, if I reverse the 5V to -5V, and reverse my snubber LEDs, the oscillator locks up.
To make it work the contacts need to swapped. Whats going on?
I think the relay armature is a permanent magnet.
So, the relay can be switched back and forth with only one coil by reversing its polarity.
The other coil does the same thing, just the opposite way.
Or it can be driven with a 4 MOSFET bridge. MOSFET bridges have snubbers built in.
This is interesting!!!
This relay can be switched in a number of different ways.
What a bonus!!!!
Duane J
Awesome news, for sure!
And being your oscillator's frequency 380Hz, or 22800 cpm. is about 126,67 times the 180 cpm value, used by Panasonioc to determine mechanical lifespan of their product.
Checking the operate and release time (with diode) shows values between 2 and 3 ms, and your data is showing us about 1.3 ms for each operation, a 100% improvement over factory data.
Outstanding!
Thanks for the mild current flowing thru the contacts in your test fixture, its electrical life isn't being yet hardly challenged, since those tests were conducted at 20 cps at their lab facilities, but in a 2A@30Vdc resistive load landscape.
Perhaps a slightly higher than normal temperature raise could be observed, due to the stress test conditions, but if things are not failing nor melting, you are in the right direction.
Our eyes and ears are poor observers, until near to catastrophic failures are about to happen.
I must agree with you, except for the diagram + and - markings, there are no references to the armature construction, but your thought about a permanent magnet makes perfect sense.
When you analyze the test diagrams (9, 10-1 and 10-2) and also wondering about the valuable adition to electromechanical hysteresis represented by a permanent magnet armature construction, IMHO we could find the reason why they didn't explicity stated its internal structure till that point.
Trade secret??? Who knows?
Oh, I've got it! YOU KNOW! Then, we know!
Yanomani
There is a good description of the permanent magnet armature operation, associated with polarized relays, at the following link:
http://electronics.stackexchange.com/questions/76666/understanding-the-polarized-relay-concept
Please note the excerpt, denoting the advantages of such construction:
"Polarized relays are usually small, have faster operate and release times, and are more resistant to shock and vibration forces."
It appears to be a case of "Face Of"!
Yanomani
With 2200 uF capacitors, 33 and 22 ohm charge caps, and a 10-ohm discharge cap, this thing is stable within less than 1 percent at 7 Hz. It can be enabled and disabled in several creative ways which sync it up to a leading pulse edge. I'm thinking of building an async receiver-transmitter pair; early drawings suggest this should require three relays per bit (two for the shift register and one latch) and at most two or three for control. And they should be able to send and receive to an ordinary computer serial port, how about that?
http://www.goldmine-elec-products.com/prodinfo.asp?number=G19618
Bean
It would even be more appealing if that had the doubled price and a $50USD quantity.... the last thing I need is go into stockpiling hundreds of dollars in hundreds of relays.
This somewhat changes my point of view on whose goldmine they are referring to ;-)
Loopy, if you look at the pictures I posted on the previous page, you'll see that changing the quantity would involve a *lot* of labor on EG's part. The package is a reel intended for dispensing to a pick-and-place machine. They would have to unwind the reel (hopefully without releasing the loosely bound cover plastic), somehow count the relays out, cut and repackage them. It just wouldn't be practical.
However, I do have 500 which I got at the 22 cent each with shipping price, so if you'd like a few PM me or post here and we can work a deal.
What is normal at the Goldmine.. no latching relays? They seem wary to keep some on hand for small quantity sales. I'd happily pay $2 for single units. I see the more conventional 5VDC relays are over $1 USD in single units.
It is an old dilemma, electronic components are really intended to supply an industry that wants to sell thousands if not millions in one sale. Hobbyist and students can ready about a lot of nice stuff, but getting just one often forces us to pay a big premium or to grab up oversupply that comes and goes.
Regrettably, those may never become a clearance item anywhere.
I still am amazed that the 'general public' perceives the conventional 5VDC as a good solution for a micro-controller driven relay when the latching version can be far superior in conserving power to the microcontroller. With 5VDC latching relays, a battery powered microcontroller can go to sleep for periods of inactivity, momentarily wake up to check status and tweek a relay if necessary, then go back to sleep.
I love the concept, but really have no projects to build.
You just described a fairly productive day in my life.
The general public probably doesn't realize how many of these things they own. My battery powered digital thermostat uses a latching relay. Since it needs only a brief pulse to turn the air conditioner on or off, and the rest of the electronics are a low power microcontroller and LCD display, the batteries last more than a year.
In a few days I after sleeping on the final layout I'll be sending this off to ExpressPCB. It's arranged to mount the relays with their pins bent straight down for thru-hole, which results in a 0.2 inch row spacing, too narrow and the pins too short for most breadboards.
This bad boy is going to cost more than the relays did but will save me a bunch of time in the first round of experiments I want to do. It has mounting room for 34 relays, 16 LED's, 24 tact switches to form 12 nonvolatile latching inputs, and the components to form a free-running oscillator.
In addition the ground pins of all coils are bussed, and the common pins of each contact are bussed separately for each group of 10 relays on the right. The tact switches are wired to set the top relays as inputs, and the LED's are arranged to reflect top relay status with an easy resistor placement. The bottom-right 20 relays also form a shift register which can be disabled to use the relays separately by cutting some top traces.
That's a lot of wiring I won't have to do. The prewiring is arranged so that the board can be either a receiver, shifting bits into the registers and latching them to the top row, or a transmitter which latches bits in from the top input row and shifts them out. My plan is to make it compatible with a regular RS-232 port although at a low baud rate and with appropriate level conversion.
I'll probably order six of these because the one-time setup and shipping costs about USD $75, after which each board only costs around $20, and I might think of something else to do with the general plan or just want to make a neater pair once the experimental set is working.
Any suggestions? Or anyone want to buy in on a sample? I can also supply the ExpressPCB source file if anyone's interested.
This is how I have mounted my test relays.
I used a 2 row 12 pin header. I pulled out 2 of the pins where not needed.
Then bent and formed the header pins to get 3 tenths spacing for my Plug Board.
As you did, I straitened the relay pins and soldered to the formed header.
BTW, this is my prototype solar tracker using the relays.
Am now attempting to get the PWM working.
That is a DS203 pocket scope in the background.
I really like it for hacking circuits together.
Sure, I have a high performance digital scope but that can't fit on my lap watching the TV.
Duane J
Took a bit longer than I expected probably due to the holidays. Top board shows a relay seated to test the pin hole pattern in the spot pre-wired for an oscillator. Bottom board is flipped vertically to show reverse side. All prewired functions can easily be disabled by cutting a few traces to make every spot general-purpose. With those buses pre-run that's a lot of wires I won't have to breadboard, yay!
http://www.goldmine-elec-products.com/prodinfo.asp?number=G19648
I also note these are 4.5V instead of 5V. I bet that's because if you're driving them with bipolar transistors, you need two transistor voltage drops (an H-bridge) instead of just one.
What's the minimum pins requirement for 28 dual-coil relays (no extra hardware)?
IIRC prop pin drivers are about 40 ohms. The 5v0 coils are 125 ohms. They switch (barely) at 3v3 which means at a current of 26 mA. Add 40 ohms and you've got 160 total (worse 200 for an H-bridge, but let's just say 160) and on a 3v3 supply you've got 20 mA. They won't switch on that.
The 4v5 coils may be different but I doubt they're different enough to overcome the extra drop for two pin-drive transistors in series forming a reversible H-bridge.
01 02 03 04 05 06 07 (7 relays)
12 13 14 15 16 17 (6 relays)
23 24 25 26 27 (5 relays)
34 35 36 37 (4 relays)
45 46 47 (3 relays)
56 57 (2 relays)
67 (1 relay)
total 28 relays
This is all unless the 4v5 coils, which these new single coil relays have, are significantly different from the 5v coils in the double-enders. I tend to doubt they're *that* different.