Speeding up a relay

Go To Last Post
54 posts / 0 new

Pages

Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi - I'm looking at using some relays in an upcoming project. Datasheet here.

They're a bit slow for my taste. What I'm wondering is this: could I overdrive the coil of a relay like the one I linked to and make it speed up? For example, could I use the 5V part, drive it for the first 5ms at 12V, then drop the voltage on the coil back to 5V? The idea would be to slam on the relay, but then throttle back to avoid overheating it.

Also - am I right in thinking that there is a resistor in the relay that is acting as a current limiter? Can you buy relays without that and then use your own current limiting system?

Thanks!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes,Yes,No,No

Charles Darwin, Lord Kelvin & Murphy are always lurking about!
Lee -.-
Riddle me this...How did the serpent move around before the fall?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
The idea would be to slam on the relay,
What % of life would that take off it?

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Just a thought, not to be contrary to JS...

What is the first to go on a relay?
The contacts or the coil?
My guess would be the contacts...

Perhaps slamming it on would reduce arcing and increase life...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
Quote:
The idea would be to slam on the relay,
What % of life would that take off it?


You can't even hear the poor relay scream, as it is in a vaccuum!

The 5V version has a 28 ohm coil resistance. You are going to draw about 178 mA when it is on. Seems high for a small reed relay, but then you selected a high vaccum, high isolation version.

You can certainly drive it harder, but to what effect? Your driver, if you impliment it as stated above, becomes much more complex, and your switching transients become a greater problem. And you knock 1 mSec off of a 3 mSec switching time... Will it really make that be of a difference?

The worst part is you will be running it well out of spec! As JS noted, what will it's life time, or mean-time-to-failure become?

If the timing is that critical then perhaps you need to re-think your overall approach. (Digital switching?)

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

One of the things about driving hard is contact bounce. You will get LOT (and lots and lots) of it.

Electromechanical relays are simply not high speed devices.

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
They're a bit slow for my taste.

So give us some figures which you consider reasonable?

Quote:

am I right in thinking that there is a resistor in the relay that is acting as a current limiter

There are no resistors in relays, but the copper wire has ohmic resistance. But it is not the resistance that controls the operate speed, but the relay coils inductance! Oh by the way, there are no inductors in the relay either!:P

A relay can be operated quite safely at a higher voltage with a short pulse of suitable duration where peak current is limited by the inductance of the relay as well as the resistance of the coil. (remember time constants viz. Ï„=L/R)
The duration of the pulse should be such that the nominal maximum relay current is not exceeded. This is how many stepper motors are actually operated of typically 24 volts when the coil is actually rated at 3 Volt.

A nice way to do this is to charge up a capacitor to say +12 Volt and then dunp it in series with the +12V supply to get +24 Volts of operating voltage. When the capacitor discharges the holding voltage is +12V.
This will give more contact bounce & shorter life.

This is often done in amateur radio where surplus microwave relays, which are commonly 28 Volts can be operated from +24 Volt pulse and then held operated on +12 V. In this case the life of the relay is not shortened!
Usually another relay is used to charge up the capacitor and I do not know or have not thought of a "solid state" method for doing this. I really think a solid state method will be prohibitively complex & expensive.

Endorsing Jim's comment relays are never going to be fast. Perhaps there is an alternative to relays!

Charles Darwin, Lord Kelvin & Murphy are always lurking about!
Lee -.-
Riddle me this...How did the serpent move around before the fall?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:
One of the things about driving hard is contact bounce. You will get LOT (and lots and lots) of it.

Electromechanical relays are simply not high speed devices.

Jim


Jim - bounce is my big fear. I'm working on a board that will have this relay on it (among many other things). I will have two drivers for the relay - a normal driver that just holds it at a set voltage, and one that gives it a higher than rated voltage for a short (and adjustable) bit, and then switches seamlessly (via diode) to a lower voltage. It'll be a pretty straight forward circuit.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

LDEVRIES wrote:
Quote:
They're a bit slow for my taste.

So give us some figures which you consider reasonable?

Quote:

am I right in thinking that there is a resistor in the relay that is acting as a current limiter

There are no resistors in relays, but the copper wire has ohmic resistance. But it is not the resistance that controls the operate speed, but the relay coils inductance! Oh by the way, there are no inductors in the relay either!:P

A relay can be operated quite safely at a higher voltage with a short pulse of suitable duration where peak current is limited by the inductance of the relay as well as the resistance of the coil. (remember time constants viz. Ï„=L/R)
The duration of the pulse should be such that the nominal maximum relay current is not exceeded. This is how many stepper motors are actually operated of typically 24 volts when the coil is actually rated at 3 Volt.

A nice way to do this is to charge up a capacitor to say +12 Volt and then dunp it in series with the +12V supply to get +24 Volts of operating voltage. When the capacitor discharges the holding voltage is +12V.
This will give more contact bounce & shorter life.

This is often done in amateur radio where surplus microwave relays, which are commonly 28 Volts can be operated from +24 Volt pulse and then held operated on +12 V. In this case the life of the relay is not shortened!
Usually another relay is used to charge up the capacitor and I do not know or have not thought of a "solid state" method for doing this. I really think a solid state method will be prohibitively complex & expensive.

Endorsing Jim's comment relays are never going to be fast. Perhaps there is an alternative to relays!


Lee - so you don't think there are any resistors whatsoever in relays? I was looking at a family of relays earlier that had a number of different coil voltage ratings. I was thinking they might make one coil, then add different resistors for the different voltage ratings. Otherwise, each coil will have to be different, giving it different electrical specs, different performance, etc. The relays were all speced to perform identically.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

DocJC wrote:
Quote:
Quote:
The idea would be to slam on the relay,
What % of life would that take off it?


You can't even hear the poor relay scream, as it is in a vaccuum!

The 5V version has a 28 ohm coil resistance. You are going to draw about 178 mA when it is on. Seems high for a small reed relay, but then you selected a high vaccum, high isolation version.

You can certainly drive it harder, but to what effect? Your driver, if you impliment it as stated above, becomes much more complex, and your switching transients become a greater problem. And you knock 1 mSec off of a 3 mSec switching time... Will it really make that be of a difference?

The worst part is you will be running it well out of spec! As JS noted, what will it's life time, or mean-time-to-failure become?

If the timing is that critical then perhaps you need to re-think your overall approach. (Digital switching?)

JC


JC - I'm not too worried about killing these things. They aren't going into a critical application and if they fail no damage will be done.

I'm planning on using these as they're rated for a very large voltage. If you can find me a 10KV MOSFET, BJT, IGBT, etc I'd be stoked as I'd much prefer to use a solid state solution! Instead, as far as I can tell - the highest rated transistors on the market are 4KV parts from Ixys.

I don't mind the extra complicated driver. More fun for me!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
so you don't think there are any resistors whatsoever in relays?

I am certain of it! Resistors produce heat & waste energy without creating any magnetic flux!

The number of turns & the diameter of the wire is carefully designed for each coil voltage so that the specs are much the same.

The electromechanical characteristics is determined by the
Amp & Number of turns (N.I), hence the N.I product must be the same.
The operating current I is determined by the applied voltage
and if the operate current is to be the same, the coil resistance R must be different.
For R to be different the length of copper wire or the cross sectional area must change.
Changing the length or cross sectional area of the wire will change the number of turns(N)
The operate lag is determined by the inductance (L) & resistance(R). We want ratio of L/R constant.
The inductance (L) is determined by the number of turns squared & the reluctance of the magnetic path.

So relay design is not trivial. If it was then there would be no need for Electrical Engineers.

Charles Darwin, Lord Kelvin & Murphy are always lurking about!
Lee -.-
Riddle me this...How did the serpent move around before the fall?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I'm not too worried about killing these things.

I am concerned you will kill yourself. You seem to know little about electricity and trivialize things. You will be rewarded with a Darwin award in due course. Those assisting you and giving advise need to be careful of possible litigation.
You remind me of this guy!
http://www.youtube.com/watch?v=Nl6_KSd9fhw
If you are easily shocked, don't watch!

Charles Darwin, Lord Kelvin & Murphy are always lurking about!
Lee -.-
Riddle me this...How did the serpent move around before the fall?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

LDEVRIES wrote:
You seem to know little about electricity

Seriously? Did you really just say that?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
am I right in thinking that there is a resistor in the relay that is acting as a current limiter? Can you buy relays without that and then use your own current limiting system?

Quote:
I was thinking they might make one coil, then add different resistors for the different voltage ratings.

Quote:
I will have two drivers for the relay - a normal driver that just holds it at a set voltage, and one that gives it a higher than rated voltage for a short (and adjustable) bit, and then switches seamlessly (via diode) to a lower voltage. It'll be a pretty straight forward circuit.

Quote:
I'm planning on using these as they're rated for a very large voltage. If you can find me a 10KV MOSFET, BJT, IGBT, etc I'd be stoked as I'd much prefer to use a solid state solution!

Quote:
I don't mind the extra complicated driver. More fun for me!

Enough said!

Charles Darwin, Lord Kelvin & Murphy are always lurking about!
Lee -.-
Riddle me this...How did the serpent move around before the fall?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Could you describe in a little more detail how you plan to apply a stepped voltage to the relay? I'm interested in what you are planning.

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:
Could you describe in a little more detail how you plan to apply a stepped voltage to the relay? I'm interested in what you are planning.

Jim

Sure. Ground one side of the coil. Other side would be connected with a p-fet to, say, 48V. Other side would also be connected, via a diode, to a p-fet that was connected to, say, 12V. Relay driver would turn on both FETs at once then turn off the 48V fet after a bit. Delay can be achieved with an rc and a couple NANDs.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think you will have some hidden problems with such a circuit. Those "body diodes" can create strange current paths. I am also really dubious that you will get significant speed-up of your relay.

In particular, you need to remember that the force on the reed in the relay is proportional to the current. Current does not turn on immediately, it ramps up like the voltage in an RC circuit. And, you also need to remember Mr. Newton (in addition to Mssrs Darwin, Kelvin, and Murphy)(; force only sets the acceleration of the reed. I suggest that you do some current vs time analysis to see what your proposed arrangement will get you. You can model the relay coil as an inductor in series with a resistor (which has a value equal to the DC resistance of the coil). I think you will be surprised. And, no, I do not know what the coil inductance is (technically, it will vary by perhaps 20%, depending on the physical position of the reed).

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
It'll be a pretty straight forward circuit.

A complete circuit diagram, detailing the biasing of the FET's and the pulse generation should not be difficult to supply then.
Impress me and I'll eat my words!

Charles Darwin, Lord Kelvin & Murphy are always lurking about!
Lee -.-
Riddle me this...How did the serpent move around before the fall?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If this is for a hobby project then I would just make the spring a bit weaker if you only need faster on time ;)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

And if you really will use higer voltages for a start I would charge a big cab to xxxV and dump it into the coil that way you don't have the hard volt change from 48V to 5V that can give a/some extra bounce/s.

edit
And feed the coil with 5V thru a diode at the same time!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

How about just driving the 5V relay with constant current driver from say 12V? I know the relays are not meant to be driven like that, and the current must be calculated via nominal coil resistance and nominal coil voltage. But on the other hand, the relays are not meant to be driven with giving it bigger than nominal voltages anyway. This would just keep the current always in safe level.

Edit: also the release time can be shortened by designing a better circuit than just a diode to suppress the inductive kick. RC snubber, zener or whatever, as long as the voltage spike is reduced to what the coil driving transistor can handle.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Some relay's has a magnet on the arm so you can actually push them up !

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Shove
plot (x/150)*(1-e^(-(150/0.5)*t)), x=5..48,t=0..0.01
into Wolfram alpha and have a look at the contour plot.
This equation shows the variation of current through a series resistor / inductor combo, with x representing your voltage across the combo.
Replace 150 for your actual coil resistance in ohms, and 0.5 for your coil inductance in henries.

The contour plot should give you a visual idea of how much your increasing voltage will help increase your force (directly proportional to your current). Then you have the added issue of the mass of the armature acting analogously to an inductor, as Jim mentioned, requiring a ramp up in speed. So the asymptotic effect you see in the contour plot from above is doubly present.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It just occured to me that instead of constant current driver that may be too complex to construct, how about just driving the 5V relay from 12V with resistor? First when power is applied, current is zero and coil has 12 volts over it. When current increases, it finally stabilizes to a point where extra resistor drops 7V and there is 5V over the relay. Can this speed up the relay?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Or put a relay across the resistor to control the supply voltage :) :) :)

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You use a hit and hold circuit. Charge an electrolytic cap via a resistor and the coil connects to that point downstream from the 12 volts. then pull a ground on the other side of the relay coil. The coil gets hit with the 12 volts and then the cap bleeds down to the voltage drop of the resistor and the coil resistance.

EDIT: I see that LDEVRIES had already posted about this above.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

LDEVRIES wrote:
Quote:
It'll be a pretty straight forward circuit.

A complete circuit diagram, detailing the biasing of the FET's and the pulse generation should not be difficult to supply then.
Impress me and I'll eat my words!

I needed to draw it up anyways, to make sure my reasoning was sound.

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

alwelch wrote:
You use a hit and hold circuit. Charge an electrolytic cap via a resistor and the coil connects to that point downstream from the 12 volts. then pull a ground on the other side of the relay coil. The coil gets hit with the 12 volts and then the cap bleeds down to the voltage drop of the resistor and the coil resistance.

EDIT: I see that LDEVRIES had already posted about this above.


I thought about doing a circuit like this - but it'd be frequency limited. For slow switching it'd be a very elegant solution.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Jepael wrote:
It just occured to me that instead of constant current driver that may be too complex to construct, how about just driving the 5V relay from 12V with resistor? First when power is applied, current is zero and coil has 12 volts over it. When current increases, it finally stabilizes to a point where extra resistor drops 7V and there is 5V over the relay. Can this speed up the relay?

This would combat the inductance of the coil, but it would not increase the field generated by the coil. So it'd make for a small increase in speed, but not as much as a solution that increases the current beyond the part's specification. However, it dominates what I came up with in terms of parts count!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Jepael wrote:
How about just driving the 5V relay with constant current driver from say 12V? I know the relays are not meant to be driven like that, and the current must be calculated via nominal coil resistance and nominal coil voltage. But on the other hand, the relays are not meant to be driven with giving it bigger than nominal voltages anyway. This would just keep the current always in safe level.

Edit: also the release time can be shortened by designing a better circuit than just a diode to suppress the inductive kick. RC snubber, zener or whatever, as long as the voltage spike is reduced to what the coil driving transistor can handle.


Running it at a constant current would be similar to Jepael's solution, in that it'd help overcome the inductance of the part but would not generate a stronger field. So it's an improvement, but you can do better!

Can you suggest a faster circuit for shutoff? While drawing up the circuit I posted earlier I realized that the ideal solution would be to have an H-bridge driving the coil, so you could slam the relay closed and slam it open.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

nleahcim wrote:
Jepael wrote:
How about just driving the 5V relay with constant current driver from say 12V? I know the relays are not meant to be driven like that, and the current must be calculated via nominal coil resistance and nominal coil voltage. But on the other hand, the relays are not meant to be driven with giving it bigger than nominal voltages anyway. This would just keep the current always in safe level.

Edit: also the release time can be shortened by designing a better circuit than just a diode to suppress the inductive kick. RC snubber, zener or whatever, as long as the voltage spike is reduced to what the coil driving transistor can handle.


Running it at a constant current would be similar to Jepael's solution, in that it'd help overcome the inductance of the part but would not generate a stronger field. So it's an improvement, but you can do better!

Can you suggest a faster circuit for shutoff? While drawing up the circuit I posted earlier I realized that the ideal solution would be to have an H-bridge driving the coil, so you could slam the relay closed and slam it open.

You may find the TI TPIC2603 interesting. I use it to pull ground on solenoid coil. What is unique is they designed it for automotive fuel injection stuff and it has a unique circuit to help turn off the coil fast by allowing back emf to reach 60 volts and then it clamps it actively instead of using a free wheeling diode across the coil. They may offer other devices like this that would fit your needs. I think you may have to dig around for some app notes to read about the flyback details.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I am a bit mystified at which point in your circuit you are measuring the voltage. I assume that the relay is supposed to be a 12 Volt relay, yet the graph saturates at +5 & +1.8V ?
Either the high side FETS are not saturated or Kirchoff's voltage Law has been violated! :shock:
Have you actually built the circuit & made it work?
By how much was the operate time decreased?
Whilst you commented in your original post that the relay you were considering operate time "was not to your taste", I could not find an operate time in the data sheet, so I don't understand what you based your original comment on :?

Charles Darwin, Lord Kelvin & Murphy are always lurking about!
Lee -.-
Riddle me this...How did the serpent move around before the fall?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

LDEVRIES wrote:
I am a bit mystified at which point in your circuit you are measuring the voltage. I assume that the relay is supposed to be a 12 Volt relay, yet the graph saturates at +5 & +1.8V ?
Either the high side FETS are not saturated or Kirchoff's voltage Law has been violated! :shock:
Have you actually built the circuit & made it work?
By how much was the operate time decreased?
Whilst you commented in your original post that the relay you were considering operate time "was not to your taste", I could not find an operate time in the data sheet, so I don't understand what you based your original comment on :?

Lee - the square wave is the input signal. It ranges from 0-5V. The other signal is the inductor current. It ranges from 0-36ma. Note the alternate scale on the right side of the chart.

You're correct that I'm assuming a 12V relay. Specifically, I'm assuming that it has 1K DCR and 500mH inductance. (yeah yeah it's not totally accurate to model the coil as an inductor - but it's good enough for this simulation) These numbers are modeled after a relay I found in the lab.

I haven't been able to test this circuit in the real world, but it's simple enough that I would be pretty surprised if it didn't work. It will be tested eventually - probably within a month or so.

The parts I'm planning on using (which are not simulated here) quote 3ms operate time and 2ms release time. Will this circuit speed it up? I hope so.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

alwelch wrote:
You may find the TI TPIC2603 interesting. I use it to pull ground on solenoid coil. What is unique is they designed it for automotive fuel injection stuff and it has a unique circuit to help turn off the coil fast by allowing back emf to reach 60 volts and then it clamps it actively instead of using a free wheeling diode across the coil. They may offer other devices like this that would fit your needs. I think you may have to dig around for some app notes to read about the flyback details.

Interesting - by allowing it to go even more negative you can discharge the coil faster. I hadn't thought of that - but it makes perfect sense. I've added a zener in series (back to back) with the diode I already had in parallel with the relay coil. It does mean that I then have to size the breakdown voltage of my FET to be greater than supply voltage + zener voltage + diode voltage. But that's easy enough! Thanks for the tip!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

What's the gate voltage of M1/M2 when M3/M4 are closed?

What about putting several solid state switches in series?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Nephazz wrote:
What's the gate voltage of M1/M2 when M3/M4 are closed?

What about putting several solid state switches in series?


Vgs of M1 and M2 are -12V and -18V, respectively, when M3 and M4 are on.

Multiple SSRs in series might work... but I'd be worried that they wouldn't turn on exactly simultaneously, or their reverse leakage current wouldn't be matched well enough, and they'd die.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I agree that a Higher voltage pulse will speed up the pull in of a relay, just as it does in stepper motors but contact arcing occurs on release if the relay is used to switch a load when it's coil is energised.

The usual method of driver protection, Diode across Relay Coil, will slow the relay release time and extend arcing on release if the relay was delivering a load current when energised.

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Sorry, instead of

Quote:
What's the gate voltage of M1/M2 when M3/M4 are closed?

I meant: What's the gate voltage of M1/M2 when M3/M4 are open?

What about using a good old thyristor?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Nephazz wrote:
Sorry, instead of
Quote:
What's the gate voltage of M1/M2 when M3/M4 are closed?

I meant: What's the gate voltage of M1/M2 when M3/M4 are open?

What about using a good old thyristor?


In that case, Vgs of both FETs is 0V.

What advantage would a thyristor offer? I must admit I'm not very familiar with them as I think of them as mostly being for AC applications.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In my mind the gates are connected to nothing, so their voltage are in the "air" and floating randomly. Don't you need to explicitly put them on ground to guarantee turning off? But I'm not that good with analog electronics. (yet)

There are circuits which allow to turn off Thyristors at will, which would qualifie them for DC usage. They are used in high power, high current applications. I'm not sure about the voltage rating and speed though. How fast do you need to switch?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thyristors may be hard to get for such small currents.

Another option (even more back in time) would be a vaccuum tube. There are versions buid for several kilovolts. You may get a few russian surplus tubes for a resonalbe price.

I know there are working circuits that use series connected transistors (e.g. 5) to switch high voltages. As far as I remember the use they use small capacitors to make sure the switch at the same time. But there may be some small leakage.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

rberger wrote:
I agree that a Higher voltage pulse will speed up the pull in of a relay, just as it does in stepper motors but contact arcing occurs on release if the relay is used to switch a load when it's coil is energised.

The usual method of driver protection, Diode across Relay Coil, will slow the relay release time and extend arcing on release if the relay was delivering a load current when energised.


I modified the design to be a zener + diode to clamp the voltage at something higher than a single diode drop. Can you suggest a better method of de-energizing the coil?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Allright - time for an update. I fully tested the circuit today. At first - I was sad - it didn't work very well. Yes, it generated 48V, then 5V across the coil exactly as planned. Problem was that when it switched to just 5V the contacts bounced a second time. Damn.

So, I started tuning it. Turns out you just want to get the coil current just a little above the specified current - maybe double. And then you switch it back to its rated voltage. With tuning, I was able to get turn on time (time from signal to switch on relay to no more bounces) to be about 2/3 of what it is when driving the relay normally. Not as big of a win as I had hoped - but still significant. About half of the switch time is just the contacts bouncing, regardless of if you're driving it normally or not.

More impressive, however, were the results with the zener in series with the flyback diode. Turn off time went from 450us with no zener to 35us with the zener. There was no bounce with turn off. No surprise there.

There's more work to be done in making an ideal relay driver - but this is definitely a good start.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The fastest release will be without a diode etc. so if you have a driver that can take 500v and a 470V zener it can't be faster.
What is 2/3 of normal on time (comparede to 35us)?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

sparrow2 wrote:
The fastest release will be without a diode etc. so if you have a driver that can take 500v and a 470V zener it can't be faster.
What is 2/3 of normal on time (compared to 35us)?

The improved on time was about 1.25ms. Way slower than the off time, unfortunately. Like I said, there is still work to be done!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Have you considered reducing the voltage to 5v before it makes contact? By anticipating the approaching collision and reducing the drive, you might reduce the bounce.

Rick

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

When you drive the relay with an over current, the armature is pulled with a higher force. When you release the 48V, the force is decreased suddenly and the elasticity of whole system create a second bounce. You should release the 48V smoothly. One solution to achieve this is to use the proposed capacitor.
Connect a capacitor in parallel with a resistor and put them between the 48V drena mosfet and the relay.
Before turning the relay ON, the capacitor is discharged and the relay will get whole 48V at the beginning. As time passes, the capacitor will charge and the current through relay will decrease. If you adjust the capacitor and resistor values properly, you will not have bounces a second time.

Also, have you considered to always keep a small current through the relay, so when you turn it ON, the current will start to build up not from zero, but from a certain value. Anyway, when you turn relay OFF, you should not have this current passing the relay. This may complicate your driver, but I notice you are very motivated.
George.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ASS-U-MEing a glass body for the reed I still question the wisdom of the approach, hey but that's just me. :)

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

EDN June 10 issue, design note "Bootstrap circuit speeds solenoid actuation"
places a charged capacitor in series with the driven coil momentarily.
http://www.edn.com/article/509267-Bootstrap_circuit_speeds_solenoid_actuation.php

The "view as pdf" link give a better presentation.

Stan

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:
I think you will have some hidden problems with such a circuit. Those "body diodes" can create strange current paths. I am also really dubious that you will get significant speed-up of your relay.

In particular, you need to remember that the force on the reed in the relay is proportional to the current. Current does not turn on immediately, it ramps up like the voltage in an RC circuit. And, you also need to remember Mr. Newton (in addition to Mssrs Darwin, Kelvin, and Murphy)(; force only sets the acceleration of the reed. I suggest that you do some current vs time analysis to see what your proposed arrangement will get you. You can model the relay coil as an inductor in series with a resistor (which has a value equal to the DC resistance of the coil). I think you will be surprised. And, no, I do not know what the coil inductance is (technically, it will vary by perhaps 20%, depending on the physical position of the reed).

Jim

What about this experiment: connect a power NPN or MOSFET transistor in paralell with the coil of the relay that is to be activated fast. In series with this, have another relay or a coil with similar inductance.

Start the "sequence" by swithing the transistor on, shorting the coil of the relay. Then apply two times the rated voltage across the circuit. Current will ramp up, but the relay will not pull.

Now switch off the transistor, and the other relay (or coil) will try to maintain a constant current trough both coils. Hopefully the relay will pull faster because of a sudden constant current.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

A relay is an 'open loop' device. Seems like you could make a servo like relay with feedback from the armature position and really get tight position control. If the armature was a magnet, then h-bridge drive could control it in and out. If the armature is just iron, then it pulls in with either polarity, needs a spring to pull out. Since you cant modulate the spring force electronically, a magnetic servo relay seems like a good fast invention.

Imagecraft compiler user

Pages