Good start Chris.
This wasn’t endearing to me, but had he asked properly, I would have been glad to answer instead of ask him a higher question, one I was sure he hadn’t ever thought of.
The SCR latches with very little effort, DC in one way will turn on the SCR, the reverse to the gate cancels the unit out, or dropping the voltage feed.
To hold them on, they require very little current.
The make simple logic gates if done properly.
The Triac is good for passing AC with just a small signal.
It can be held on with the use of the SCR if necessary, or just a continuos signal.
Power Latching IC
Q: "What is the battery voltage you plan to use?
And Vcc of the uP?"
A: "They are both 3.3v dc."
This limits your options. Forget the SCR, it will have a voltage drop between 1 and 2 volts, depending on load current and temperature. Also, your 50 mA load is a maximum value, which might drop low when minimum to let the SCR turn off.
You may be able to tolerate a voltage drop of 100 mV in the switching device. This will be a 3 percent loss with a 3.3 volt battery.
Two possible devices would be either a P-channel MOSFET or a bipolar PNP transistor. The P-channel MOSFET will probably be more expensive, but it will also be more efficient.
The gate drive for the MOSFET will be about 3 volts. "Logic level" FETs may have specifications that will guarentee a minimum source-drain resistance at this level. You need a value less than 2 ohms. Look for a low source-drain breakdown voltage, this usually correlates with low source-drain resistance.
The other alternative is a bipolar PNP transistor. This will be less efficient, since the base current will be about 5 to 10 percent of the collector current to insure saturation. This results in a 5 to 10 percent power loss, which is usually undesirable in a battery powered unit. Low saturation voltage can be obtained by operating the transistor far below its peak collector rating. For a 50 mA load, look at transistors with a peak current rating in the .5 to 1 A range.
The latching action can be obtained by using an output pin on the microcontroller to drive the switching device. As noted previously, a discrete inverter will probably be needed to allow the switch to remain turned off with no voltage supply to the microcontroller. The output pin needs to be set high before the initial button is released. This will probably take a few teneths of a second. The program in the microcontroller will have to maintain this pin high until it is desired to power down, which is done by programming the output pin low.
And Vcc of the uP?"
A: "They are both 3.3v dc."
This limits your options. Forget the SCR, it will have a voltage drop between 1 and 2 volts, depending on load current and temperature. Also, your 50 mA load is a maximum value, which might drop low when minimum to let the SCR turn off.
You may be able to tolerate a voltage drop of 100 mV in the switching device. This will be a 3 percent loss with a 3.3 volt battery.
Two possible devices would be either a P-channel MOSFET or a bipolar PNP transistor. The P-channel MOSFET will probably be more expensive, but it will also be more efficient.
The gate drive for the MOSFET will be about 3 volts. "Logic level" FETs may have specifications that will guarentee a minimum source-drain resistance at this level. You need a value less than 2 ohms. Look for a low source-drain breakdown voltage, this usually correlates with low source-drain resistance.
The other alternative is a bipolar PNP transistor. This will be less efficient, since the base current will be about 5 to 10 percent of the collector current to insure saturation. This results in a 5 to 10 percent power loss, which is usually undesirable in a battery powered unit. Low saturation voltage can be obtained by operating the transistor far below its peak collector rating. For a 50 mA load, look at transistors with a peak current rating in the .5 to 1 A range.
The latching action can be obtained by using an output pin on the microcontroller to drive the switching device. As noted previously, a discrete inverter will probably be needed to allow the switch to remain turned off with no voltage supply to the microcontroller. The output pin needs to be set high before the initial button is released. This will probably take a few teneths of a second. The program in the microcontroller will have to maintain this pin high until it is desired to power down, which is done by programming the output pin low.
-
Chris Smith
- Posts: 4325
- Joined: Tue Dec 04, 2001 1:01 am
- Location: Bieber Ca.
Hi there,
I have to think the SCR idea isnt a bad idea, it's just not as easy to
implement as it seems at first, and then you have to implement
some means of turn off as other posts have already mentioned.
I've used the "transistor to short the SCR out" method back in
the early 80's in an attempt to implement a high current cheap
switching regulator that could use an inexpensive high current SCR
(inexpensive relative to everything else that was available at that
time and could also handle the high current) and a relatively cheap
transistor to turn it off. The transistor doesnt have to be able to
handle the full rated current continuously but it does have to be able
to handle the peak. Anyway, what i found out is that besides the
minimum holding current rating of the SCR there is also a minimum
turn off time, which means the transistor has to stay on for a min
time period in order for the SCR to actually stop conducting. Too
short and the SCR stays on, too long and the transistor heats up.
This turn off time turned out to be too long for most SCRs that were
available back then so the project was abandoned.
Since this is simply a power supply on/off operation though the
turn off time isnt that critical. Keep the transistor on for say 100ms
and the SCR should turn off.
The other problem however is that an SCR drops a bit more voltage
than we would like to see in low voltage applications (as mentioned
before in this thread). This alone means it might not work that well.
Also, Chris, i've never heard of the idea of turning off an SCR by
reversing it's gate drive. Perhaps you can explain how to get this
to work. Triacs allow plus and minus gate currents but that's to
turn them on more efficiently in different quadrants of operation.
Another idea i dont think was mentioned yet is to simply use another
pin on the uC, assuming we have control over the programming too.
The uC drives the MOSFET (N chan, P chan, take your pick) and
another pin on the uC is used for the switch. The entire operation
is controlled via the software programming. The feasibility of this
idea depends on what technology the uC is you are using.
Another simpler idea is to use a relay. Not a regular relay, but a
latching type relay. One button press latches the relay closed,
and it needs no more power. Another buttong press latches the
relay open, and it needs no more power.
The MOSFET approaches will be more reliable however for long term use.
Of course there is always the simple bipolar transistor idea too.
Many transistors passing such a small current like 50ma will easily
saturate to 0.2 volts or less. The transistor is turned on via a
switch closure through a resistor and then the uC fires up and
mimics the switch closure so when the switch is opened the
transistor stays on. The uC is then under full control of the transistor
and can turn it off whenever needed.
I'll call this solution the "25 cent solution" because that's about what
it would cost using a general purpose switching transistor like the
2N4001 or 2N4003 (NPN or PNP whatever works for you).
I have to think the SCR idea isnt a bad idea, it's just not as easy to
implement as it seems at first, and then you have to implement
some means of turn off as other posts have already mentioned.
I've used the "transistor to short the SCR out" method back in
the early 80's in an attempt to implement a high current cheap
switching regulator that could use an inexpensive high current SCR
(inexpensive relative to everything else that was available at that
time and could also handle the high current) and a relatively cheap
transistor to turn it off. The transistor doesnt have to be able to
handle the full rated current continuously but it does have to be able
to handle the peak. Anyway, what i found out is that besides the
minimum holding current rating of the SCR there is also a minimum
turn off time, which means the transistor has to stay on for a min
time period in order for the SCR to actually stop conducting. Too
short and the SCR stays on, too long and the transistor heats up.
This turn off time turned out to be too long for most SCRs that were
available back then so the project was abandoned.
Since this is simply a power supply on/off operation though the
turn off time isnt that critical. Keep the transistor on for say 100ms
and the SCR should turn off.
The other problem however is that an SCR drops a bit more voltage
than we would like to see in low voltage applications (as mentioned
before in this thread). This alone means it might not work that well.
Also, Chris, i've never heard of the idea of turning off an SCR by
reversing it's gate drive. Perhaps you can explain how to get this
to work. Triacs allow plus and minus gate currents but that's to
turn them on more efficiently in different quadrants of operation.
Another idea i dont think was mentioned yet is to simply use another
pin on the uC, assuming we have control over the programming too.
The uC drives the MOSFET (N chan, P chan, take your pick) and
another pin on the uC is used for the switch. The entire operation
is controlled via the software programming. The feasibility of this
idea depends on what technology the uC is you are using.
Another simpler idea is to use a relay. Not a regular relay, but a
latching type relay. One button press latches the relay closed,
and it needs no more power. Another buttong press latches the
relay open, and it needs no more power.
The MOSFET approaches will be more reliable however for long term use.
Of course there is always the simple bipolar transistor idea too.
Many transistors passing such a small current like 50ma will easily
saturate to 0.2 volts or less. The transistor is turned on via a
switch closure through a resistor and then the uC fires up and
mimics the switch closure so when the switch is opened the
transistor stays on. The uC is then under full control of the transistor
and can turn it off whenever needed.
I'll call this solution the "25 cent solution" because that's about what
it would cost using a general purpose switching transistor like the
2N4001 or 2N4003 (NPN or PNP whatever works for you).
LEDs vs Bulbs, LEDs are winning.
Hello Al,
There was a variation on the SCR called a GTO (gate turn off) device developed by General Electric. The device was turned off by a reverse polarity pulse applied to the gate. It wasn't very practical, since the gate pulse needed was a substantial fraction of the load current that was flowing. For a moderate size of GTO, the turn-off pulse could be fractions of an ampere or posssibly amperes.
This area may have been revisited by one of the Japanese semiconductor companies. I can't recall actual details, but one company may have developed devices with higher turn-off gain than the GE devices.
There was a variation on the SCR called a GTO (gate turn off) device developed by General Electric. The device was turned off by a reverse polarity pulse applied to the gate. It wasn't very practical, since the gate pulse needed was a substantial fraction of the load current that was flowing. For a moderate size of GTO, the turn-off pulse could be fractions of an ampere or posssibly amperes.
This area may have been revisited by one of the Japanese semiconductor companies. I can't recall actual details, but one company may have developed devices with higher turn-off gain than the GE devices.
-
Chris Smith
- Posts: 4325
- Joined: Tue Dec 04, 2001 1:01 am
- Location: Bieber Ca.
Seems there are several reasonable solutions. Is the goal to conserve battery power by shutting down a section of a device or is it to power down a remote circuit monitored by the uProc? One solution needs to skimp on power and eliminate voltage drops as much as possible while the other can use a little more power in one location to be more creative at the other end.
One could ensure regulated voltage at the load by putting a LDO regulator right there with it after the switch and power with a larger source like 5V (which you may not have). A FET or Bipolar transistor are sufficiant to serve as the switch with this built in regulator. This is a nice 2 component solution but may waste a little more power than necessary while operating depending on the LDO. This is non latching but with the addition of another transistor or two, can be made to be so.
Fairchild makes a line of LDO regulators with a shutdown pin but a quick look at the datasheet looks like they require more parts than a typical 3 pin but a deeper search may show an LDO with an eneble pin or a method of achieving that with a typical 3 pin LDO.
MrAl suggestion of using another uP pin reminded me of another trick that works well with low current loads. You can sometimes power a load with an output pin directly. Providing the output VOH is 3.3V as it usually is for CMOS and that its output current IOH is greater than that needed by the load, it should work. 50mA would be too much but 5mA maybe and at 100uA likely. This too is non latching but you can do the same thing with the output pin of a latch.
One could ensure regulated voltage at the load by putting a LDO regulator right there with it after the switch and power with a larger source like 5V (which you may not have). A FET or Bipolar transistor are sufficiant to serve as the switch with this built in regulator. This is a nice 2 component solution but may waste a little more power than necessary while operating depending on the LDO. This is non latching but with the addition of another transistor or two, can be made to be so.
Fairchild makes a line of LDO regulators with a shutdown pin but a quick look at the datasheet looks like they require more parts than a typical 3 pin but a deeper search may show an LDO with an eneble pin or a method of achieving that with a typical 3 pin LDO.
MrAl suggestion of using another uP pin reminded me of another trick that works well with low current loads. You can sometimes power a load with an output pin directly. Providing the output VOH is 3.3V as it usually is for CMOS and that its output current IOH is greater than that needed by the load, it should work. 50mA would be too much but 5mA maybe and at 100uA likely. This too is non latching but you can do the same thing with the output pin of a latch.
I checked my old 1972 GE SCR manual. It mentions that a "GTO" can be turned off by injecting a reverse current of a value double that the device is conducting into the gate, however the structure of the junctions in a GTO device is different from a regular SCR.rshayes wrote:Hello Al,
There was a variation on the SCR called a GTO (gate turn off) device developed by General Electric. The device was turned off by a reverse polarity pulse applied to the gate. It wasn't very practical, since the gate pulse needed was a substantial fraction of the load current that was flowing. For a moderate size of GTO, the turn-off pulse could be fractions of an ampere or posssibly amperes.
This area may have been revisited by one of the Japanese semiconductor companies. I can't recall actual details, but one company may have developed devices with higher turn-off gain than the GE devices.
It also says on a different page that there is no gate control of an "SCR" once it is turned on. You can turn it off by using extra circuitry to commutate the current. ie: the current through the device must be reduced to zero for a few microseconds.
-
Chris Smith
- Posts: 4325
- Joined: Tue Dec 04, 2001 1:01 am
- Location: Bieber Ca.
Gate turn off SCRs were used in things like the KV 17/19 TV [Trinatron] and from memory it controlled the Fly back or some such device that controlled a voltage of 1100+ volts or so.
I repaired mine in the early 80s with this device.
It was as big as a To3, it was huge, and it took a lot of power to controll it.
It was not practical for much else, and All I found on the subject these days was the part # NTE 276 .
5 amps at 1250 volts, huge gate turn on specs. [Several volts/ 100 plus milli]
I repaired mine in the early 80s with this device.
It was as big as a To3, it was huge, and it took a lot of power to controll it.
It was not practical for much else, and All I found on the subject these days was the part # NTE 276 .
5 amps at 1250 volts, huge gate turn on specs. [Several volts/ 100 plus milli]
By the way, guys, I HATE thyristors with a passion. They are unreliable and vary wildly in characteristics.
Light dimmers in my house: Half of them have had to be replaced due to triac failure.
At the TV studio where I worked, the lighting control solid state relays failed regularly.
In the car stereo shop that I owned, one model of cassette tape deck used unijuction transistor oscillators in the reel motion sensors. This is the only device that ever failed in this model.
And finally, years ago in 1964, my Dad's consulting engineering firm failed due to the unreliability of one device: The long defunct GE 13A8 SCS (silicon controlled switch.) I still have some of these TO-5 can 4 lead devices in my parts collection. Searching the internet today, I can find no history of this device even existing. (GE part # 4JD-13A8)
Light dimmers in my house: Half of them have had to be replaced due to triac failure.
At the TV studio where I worked, the lighting control solid state relays failed regularly.
In the car stereo shop that I owned, one model of cassette tape deck used unijuction transistor oscillators in the reel motion sensors. This is the only device that ever failed in this model.
And finally, years ago in 1964, my Dad's consulting engineering firm failed due to the unreliability of one device: The long defunct GE 13A8 SCS (silicon controlled switch.) I still have some of these TO-5 can 4 lead devices in my parts collection. Searching the internet today, I can find no history of this device even existing. (GE part # 4JD-13A8)
haklesup,
Let me clarify the use case here. I press a button and a micro processor is started. Once it finishes doing what its' programming tells it to, I would like to pull an IO line High and have the circuit power off the micro. It is really a case of conserving as much power as possible and to allow flexibility/variability in the programming (Rather than trying to fit it into some prefixed time).
Dale,
I say a 3.3v battery because even though batteries say they are 3.6v when you put any type of load on them, they seem to drop to between 3.3v and 3.4v. Didn't mean to imply a "new" battery type!
You all have really been great and I truly appreciate the attention this post has received. I am very appreciative of this.
Regards,
Karl
Let me clarify the use case here. I press a button and a micro processor is started. Once it finishes doing what its' programming tells it to, I would like to pull an IO line High and have the circuit power off the micro. It is really a case of conserving as much power as possible and to allow flexibility/variability in the programming (Rather than trying to fit it into some prefixed time).
Dale,
I say a 3.3v battery because even though batteries say they are 3.6v when you put any type of load on them, they seem to drop to between 3.3v and 3.4v. Didn't mean to imply a "new" battery type!
You all have really been great and I truly appreciate the attention this post has received. I am very appreciative of this.
Regards,
Karl
-
Chris Smith
- Posts: 4325
- Joined: Tue Dec 04, 2001 1:01 am
- Location: Bieber Ca.
Thyristors are the most widely used component in the general industry aside from the Transistor, used for AC when both ways are needed.
The SCR is also one of the most widely used items, making gate logic for simple logic a snap.
Like the FET, it goes back more than 3 decades now with almost a 99.9% reliability.
If your looking for really low logic switching and holding,...Id look up the specs for the LF11331, LF11202 etc, JFET analog switches.
Thats a start.
The SCR is also one of the most widely used items, making gate logic for simple logic a snap.
Like the FET, it goes back more than 3 decades now with almost a 99.9% reliability.
If your looking for really low logic switching and holding,...Id look up the specs for the LF11331, LF11202 etc, JFET analog switches.
Thats a start.
The GTO Switch was described in the 3rd edition of their SCR Manual, published in 1964. A data sheet listing a G5 and G6 series was given, but it noted that this was for reference use only, the devices had been withdrawn.
The 7th edition of the GE Transistor Manual, also published in 1964, has a section on the Silicon Controlled Switch (SCS). This was basically an SCR with a second gate connection associated with the anode. In other words, all four layers of the SCR were brought out to external pins. Data was given for types 3N81 to 3N86. The 4JD-13A8 was probably a house number for a developmental device previous to the JEDEC registration of the 3N81 series.
The SCS appears to have had similar properties to the GTO, with the additional feature that it could be triggered on or off with either gate, as well as with the conventional SCR methods. The turn off gain may have been poor, it isn't clearly specified in the tabulated data.
The 7th edition of the GE Transistor Manual, also published in 1964, has a section on the Silicon Controlled Switch (SCS). This was basically an SCR with a second gate connection associated with the anode. In other words, all four layers of the SCR were brought out to external pins. Data was given for types 3N81 to 3N86. The 4JD-13A8 was probably a house number for a developmental device previous to the JEDEC registration of the 3N81 series.
The SCS appears to have had similar properties to the GTO, with the additional feature that it could be triggered on or off with either gate, as well as with the conventional SCR methods. The turn off gain may have been poor, it isn't clearly specified in the tabulated data.
Most Uc's have a sleep or low-power mode,
Would it be posible to just use this mode instead of trying to externally power down?
Depending on the Uc family, some of the sleep/low-power current draws are on the order of microamps.
Eg, press the button , controller wakes up, performs it's task and automatically goes back to sleep, all in software.
I hope thie helped!
Depending on the Uc family, some of the sleep/low-power current draws are on the order of microamps.
Eg, press the button , controller wakes up, performs it's task and automatically goes back to sleep, all in software.
I hope thie helped!
Who is online
Users browsing this forum: No registered users and 94 guests