I'm asking this because in Google many articles seem to use current limiting and regulating interchangeably, so I'm confused.
First question... Are these two assumptions correct?
1. A current limiter sets a current target value, if the load resistance tries to draw more current it is stopped at that limit, but it can draw any current less and the limiter basically falls out of play.
2. A current regulator sets a current target value, if the load tries to draw more it is stopped at that limit just like a limiter. but ALSO if the load resistance tries to draw less current the current regulator forces the load to draw more current by raising the voltage.
Second question... When I wire up an LM317 with a resistor between ADJ and VO and put that in series as Vcc - Vi -Vo - load am I getting a current limiter or a current regulator?
I need a current limiter not a current regulator. Can a simple IC Regulator be used as a current limiter so that it only comes into action if the current gets too high and ignores a low current. (assuming the above LM317 isn't already a limiter)
First question... Are these two assumptions correct?
1. A current limiter sets a current target value, if the load resistance tries to draw more current it is stopped at that limit, but it can draw any current less and the limiter basically falls out of play.
2. A current regulator sets a current target value, if the load tries to draw more it is stopped at that limit just like a limiter. but ALSO if the load resistance tries to draw less current the current regulator forces the load to draw more current by raising the voltage.
Second question... When I wire up an LM317 with a resistor between ADJ and VO and put that in series as Vcc - Vi -Vo - load am I getting a current limiter or a current regulator?
I need a current limiter not a current regulator. Can a simple IC Regulator be used as a current limiter so that it only comes into action if the current gets too high and ignores a low current. (assuming the above LM317 isn't already a limiter)
Wow that was quick. So if I want a current limiter I should look for a discreet circuit, IC Regulators wouldn't work out of the box as current limiters?
1. True & False
Most of modern current limiters (called eFuse @ TI & Current limiter @ Maxim) can have 3 behaviors depending of Part number or configurable by a resistor:
- Latch : if current limit is reached then the device shut down and you have to do a OFF-ON cycle to reset the IC.
- Current limiter : as you say. When the limit is reached, the IC can't sink more current even if the load wants but it sinks the current limit.
- Auto-retry : when current limit is reached, then the IC go in off mode (teh FET in series go OFF) then the IC do a retry some ms after the limit to see if the current falls below the limit. If yes, it sink the current, if no it stays in auto-retry mode.
See MAX17608 as example. I use it on all of my design to protect DC line. It is an electronic Fuse. So no need for replacement. I always use the auto-retry mode. To do that, be sure to configure the current limit at a level that should never occur except if a short-circuit happens
So in your case, you can use a such IC with current limit behavior (it is not a constant current source except when limit is reached).
Most of modern current limiters (called eFuse @ TI & Current limiter @ Maxim) can have 3 behaviors depending of Part number or configurable by a resistor:
- Latch : if current limit is reached then the device shut down and you have to do a OFF-ON cycle to reset the IC.
- Current limiter : as you say. When the limit is reached, the IC can't sink more current even if the load wants but it sinks the current limit.
- Auto-retry : when current limit is reached, then the IC go in off mode (teh FET in series go OFF) then the IC do a retry some ms after the limit to see if the current falls below the limit. If yes, it sink the current, if no it stays in auto-retry mode.
See MAX17608 as example. I use it on all of my design to protect DC line. It is an electronic Fuse. So no need for replacement. I always use the auto-retry mode. To do that, be sure to configure the current limit at a level that should never occur except if a short-circuit happens
So in your case, you can use a such IC with current limit behavior (it is not a constant current source except when limit is reached).
Last edited:
Its for inrush so I dont want it to latch or turn off or do anything funky just hold a line in the sand for 30 seconds or so. I need it to simply limit the current during inrush then drop out of the picture after the inrush is over. For example a part that normally operates at 600ma after its all warmed up but it likes to pull 1000ma at start up. I want to limit the start up to lets say 675ma (so it will warm up slower but eventually come back down to 600ma). Basically just flatten out the warm up draw by letting it eat a little extra current but on a diet if it tries to eat 1000ma. When the part comes back down to only drawing 600ma (warmed up) then I want the current limiter to recuse itself, go away, it did its job.
A current limiting resistor approach is always present and wasting heat. So I'm thinking a programmable current limiter is a more modern way where it will just switch itself in then out.
eFuse is the solution. With a capacitor in the nF range you can configure the inrush current. Take a look at for example TPS2596 (Vin = 19V max, I=2A). I take it as example because it's a SO8 package so easy to solder. Of course, choice of Part Number depends on your V/I requirements. Some Maxim IC have an inrush current limiter too (not all).
1. True & False
Most of modern current limiters (called eFuse @ TI & Current limiter @ Maxim) can have 3 behaviors depending of Part number or configurable by a resistor:
- Latch : if current limit is reached then the device shut down and you have to do a OFF-ON cycle to reset the IC.
- Current limiter : as you say. When the limit is reached, the IC can't sink more current even if the load wants but it sinks the current limit.
- Auto-retry : when current limit is reached, then the IC go in off mode (teh FET in series go OFF) then the IC do a retry some ms after the limit to see if the current falls below the limit. If yes, it sink the current, if no it stays in auto-retry mode.
See MAX17608 as example. I use it on all of my design to protect DC line. It is an electronic Fuse. So no need for replacement. I always use the auto-retry mode. To do that, be sure to configure the current limit at a level that should never occur except if a short-circuit happens
So in your case, you can use a such IC with current limit behavior (it is not a constant current source except when limit is reached).
Thanks! I'm going to look up those parts for the second mode, current limit.
I guess you want to limit DC current, not AC current.
B+ current, for example.
Not transformer primary current.
Correct?
What exact type of part normally draws 600mA when warmed up, but could draw 1000mA at start up?
A tube that has positive grid voltage at start up?
A filter capacitor?
Or what?
Please post a schematic, it might be helpful for finding the most elegant solution (and perhaps the most simple solution).
B+ current, for example.
Not transformer primary current.
Correct?
What exact type of part normally draws 600mA when warmed up, but could draw 1000mA at start up?
A tube that has positive grid voltage at start up?
A filter capacitor?
Or what?
Please post a schematic, it might be helpful for finding the most elegant solution (and perhaps the most simple solution).
Last edited:
It's about goals, and vocabulary.
Limiting is about protection: preventing unsafe current, without caring much about accuracy. For example, a fuse. It's very inaccurate, but it prevents fires when there is a short. Or the internal protection circuit for LDOs or switching regulators. The current at which it triggers is usually specified with low accuracy, because it is neither useful nor profitable to make it more accurate.
Regulation is about keeping a current well regulated, accurate, and constant. That's a different goal.
> When I wire up an LM317 with a resistor between ADJ and VO and put that in series as Vcc - Vi -Vo - load am I getting a current limiter or a current regulator?
It's pretty accurate and behaves as a current source, so current regulator.
> Can a simple IC Regulator be used as a current limiter so that it only comes into action if the current gets too high and ignores a low current.
Yes, most voltage regulators have overcurrent protection.
If you want to experiment, get a bench power supply. It's convenient to have two knobs.
Limiting is about protection: preventing unsafe current, without caring much about accuracy. For example, a fuse. It's very inaccurate, but it prevents fires when there is a short. Or the internal protection circuit for LDOs or switching regulators. The current at which it triggers is usually specified with low accuracy, because it is neither useful nor profitable to make it more accurate.
Regulation is about keeping a current well regulated, accurate, and constant. That's a different goal.
> When I wire up an LM317 with a resistor between ADJ and VO and put that in series as Vcc - Vi -Vo - load am I getting a current limiter or a current regulator?
It's pretty accurate and behaves as a current source, so current regulator.
> Can a simple IC Regulator be used as a current limiter so that it only comes into action if the current gets too high and ignores a low current.
Yes, most voltage regulators have overcurrent protection.
If you want to experiment, get a bench power supply. It's convenient to have two knobs.
1. True & False
Most of modern current limiters (called eFuse @ TI & Current limiter @ Maxim) can have 3 behaviors depending of Part number or configurable by a resistor:
- Latch : if current limit is reached then the device shut down and you have to do a OFF-ON cycle to reset the IC.
- Current limiter : as you say. When the limit is reached, the IC can't sink more current even if the load wants but it sinks the current limit.
- Auto-retry : when current limit is reached, then the IC go in off mode (teh FET in series go OFF) then the IC do a retry some ms after the limit to see if the current falls below the limit. If yes, it sink the current, if no it stays in auto-retry mode.
See MAX17608 as example. I use it on all of my design to protect DC line. It is an electronic Fuse. So no need for replacement. I always use the auto-retry mode. To do that, be sure to configure the current limit at a level that should never occur except if a short-circuit happens
So in your case, you can use a such IC with current limit behavior (it is not a constant current source except when limit is reached).
Ill admit the 600 to 1000 ratio I just pulled out of my butt
It could be 600 to 800 a 33% surge. I know when light bulbs burn out its always when you flip the switch. So I was thinking about what if you can just allow a load to eat a little extra current and not be a hog, in a way other than a resistor that has to waste energy all the time. Something that switches itself out and becomes a wire below the limit but let's the load eat just a little more to get going but no more than what you've set. Yes for filaments or B+ maybe other things too. My Rigol bench power supply has such a limiter, I can set the limit to 620ma fire up a 600ma tube cold see the draw ceiling at 620 then in 10 seconds drop to 600 never going over 620.
Last edited:
Well, designing a circuit to a spec you casually pull out of your butt is not a particularly effective practice.
Yes bulbs burn out when you flip the switch. They are essentially filaments. Have you ever seen a tube filament burn out when you flip the switch? And are you talking about limiting the filament current? Seems to me you were talking about plate current. Nothing in the plate to burn out when you flip the switch.
Yes bulbs burn out when you flip the switch. They are essentially filaments. Have you ever seen a tube filament burn out when you flip the switch? And are you talking about limiting the filament current? Seems to me you were talking about plate current. Nothing in the plate to burn out when you flip the switch.
Windcrest77,
Caution! Do not fully believe in the current limiter of your power supply.
Two things almost every commercial Test and Measurement Power Supply has:
1. A settable current limit, and . . .
2. An electrolytic capacitor inside the power supply, directly across the output terminals.
If you set the voltage to 6.3V, and the current limit to 600mA,
and then connect the filament across the power supply, you will get a transient.
A tube, say a 6SN7, has a 600mA filament @ 6.3V. When warm, the filament is 10.5 ohms . . .
But when the 6SN7 filament is cold, it is less than 2 Ohms.
6.3V/2 Ohms = 3.15 Amps.
Yes, your power supply, that has an electrolytic capacitor across the inside of the output terminals, will deliver a short transient of 3.15 Amps to a 6SN7 filament.
(even though you set the current limit to 600mA).
Shocking but true (pun intended).
Know thy Test Equipment!
Caution! Do not fully believe in the current limiter of your power supply.
Two things almost every commercial Test and Measurement Power Supply has:
1. A settable current limit, and . . .
2. An electrolytic capacitor inside the power supply, directly across the output terminals.
If you set the voltage to 6.3V, and the current limit to 600mA,
and then connect the filament across the power supply, you will get a transient.
A tube, say a 6SN7, has a 600mA filament @ 6.3V. When warm, the filament is 10.5 ohms . . .
But when the 6SN7 filament is cold, it is less than 2 Ohms.
6.3V/2 Ohms = 3.15 Amps.
Yes, your power supply, that has an electrolytic capacitor across the inside of the output terminals, will deliver a short transient of 3.15 Amps to a 6SN7 filament.
(even though you set the current limit to 600mA).
Shocking but true (pun intended).
Know thy Test Equipment!
Still thinking / learning about this subject, Elliott Sound has a nice article covering inrush, very interesting.
Inrush Current
Inrush Current
If your power supply is a good enough design, you should be able to do the following:
With no load, set the current limit to very low, say 10mA.
Set the voltage to 6.3V. There is no current with the open circuit.
Connect the 600mA filament, and notice the voltage is at or near zero volts (10mA into the filament is current limited).
Now, set the current limit to 700mA, and notice that the voltage goes up to 6.3V when the filament warms up.
Now, set the current limit back down to 10mA.
Disconnect the filament, and the voltage should go up to 6.3V (open circuit).
Now, connect a 100mA fuse (0.1A) across the output terminals. The electrolytic capacitor inside the power supply across the output terminals will blow the fuse before the 10mA current limit takes over.
Just sayin'
With no load, set the current limit to very low, say 10mA.
Set the voltage to 6.3V. There is no current with the open circuit.
Connect the 600mA filament, and notice the voltage is at or near zero volts (10mA into the filament is current limited).
Now, set the current limit to 700mA, and notice that the voltage goes up to 6.3V when the filament warms up.
Now, set the current limit back down to 10mA.
Disconnect the filament, and the voltage should go up to 6.3V (open circuit).
Now, connect a 100mA fuse (0.1A) across the output terminals. The electrolytic capacitor inside the power supply across the output terminals will blow the fuse before the 10mA current limit takes over.
Just sayin'
It is one of Rigols largest 30-0-30 v linear supplies, weighs a ton and is a good 20 inches deep so it must have some serious iron. Right now I'm liking Circuits that put a shorting relay across an NTC, the NTC does its thing then after a time period gets shorted out by a relay so its not an energy waster, it becomes a wire.
mmm it doesn't even have to be an NTC a 10 ohm resistor in the torroid primary, a relay shunt across that, relay coil can be hooked to a "Delay on make" HVAC module set to 5 or 10 seconds, 120 volt AC relay. Everything is on the primary side so it protects everything in the amp especially with torroids feeding large filter caps.
Typical "Delay on make" module to delay-energize the resistor bypass relay or to bypass the resistor directly maybe, this $8 piece is rated 1.5 Amp contacts:
Icm ICM102 $7.66 Relay, Time Delay | Zoro.com
general idea except I'd use 120v relay with the delay on make module so no need for a low voltage source:
Windcrest77,
Get a schematic of your Rigol power supply.
I bet you will find there is an electrolytic cap across the output terminals (one from + to Ground; and one from - to Ground).
Take a 100uF electrolytic cap that has 6.3VDC across it. Short it with a 28 gauge solid wire, and see if it tries to weld the wire where you connect it.
Example: 6.3V / 0.1 Ohm Cap ESR = 63 Amp transient.
Brief, but enough to cause a metallization exchange.
Get a schematic of your Rigol power supply.
I bet you will find there is an electrolytic cap across the output terminals (one from + to Ground; and one from - to Ground).
Take a 100uF electrolytic cap that has 6.3VDC across it. Short it with a 28 gauge solid wire, and see if it tries to weld the wire where you connect it.
Example: 6.3V / 0.1 Ohm Cap ESR = 63 Amp transient.
Brief, but enough to cause a metallization exchange.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.