That transistor BSS223PW might work, but here is its Vgs parameter:
Notice that they specify it at 1.5 µA of drain current, whereas Rohm RZM001P02T2L specified it at 100µA--The Rohm part is more "turned on" than the Infineon part at about the same threshold. The threshold is not a well controlled parameter in either case, true of mosfets in general Furthermore, the BSS223PW does not show graphs below 2.2V, so I'd tend to prefer the RZM001P02T2L given a choice between the two. The BSS223PW is optimized for fast, low loss turn-on-off in DC converters, for example, the turnoff delay is 5ns versus 325ns, but that is a parameter of no consequence for what you want to do. The RZM001P02T2L is more "general purpose". There are so many different kinds of transistors available because of all the contradictory parameters that might need to be optimized for one application or another.
You said,
It shows source to gate of 2.5V and 2.2V though, so I can imagine where the 1.2V would lie and it seems it would still be over 100ma? A little confused.
I'm not sure where your confusion lies, but I do appreciate that it can be daunting. There was that 100 microamp figure for the Vgsth threshold. That isn't a current that will arise in your circuit with this transistor. It is only an indication that the transistor can support 100µA current at the low threshold, and could therefore also support the 1MΩ load without error.
Your schematic there looks good. Can you breadboard this before committing to a pcb?
That transistor BSS223PW might work, but here is its Vgs parameter:
Notice that they specify it at 1.5 µA of drain current, whereas Rohm RZM001P02T2L specified it at 100µA--The Rohm part is more "turned on" than the Infineon part at about the same threshold. The threshold is not a well controlled parameter in either case, true of mosfets in general Furthermore, the BSS223PW does not show graphs below 2.2V, so I'd tend to prefer the RZM001P02T2L given a choice between the two. The BSS223PW is optimized for fast, low loss turn-on-off in DC converters, for example, the turnoff delay is 5ns versus 325ns, but that is a parameter of no consequence for what you want to do. The RZM001P02T2L is more "general purpose". There are so many different kinds of transistors available because of all the contradictory parameters that might need to be optimized for one application or another.
You said,
It shows source to gate of 2.5V and 2.2V though, so I can imagine where the 1.2V would lie and it seems it would still be over 100ma? A little confused.
I'm not sure where your confusion lies, but I do appreciate that it can be daunting. There was that 100 microamp figure for the Vgsth threshold. That isn't a current that will arise in your circuit with this transistor. It is only an indication that the transistor can support 100µA current at the low threshold, and could therefore also support the 1MΩ load without error.
Your schematic there looks good. Can you breadboard this before committing to a pcb?
Thanks again for the info. This is making more sense. In that case I will switch back over to the Rohm RZM001P02T2L part.
I'm curious though with the BSS223PW, the mentioned 1.5uA is at the minimum threshold though right? How do you tell how much current afterwards though? This information is on the graph "5 Typ. output characteristic" right? Would I just estimate where 1.2V might be?
Unfortunately the circuit is not all that practical on a breadboard... I would have to solder small leads to all of these tiny parts and it might not be very reliable.
I tend to home fab little breakout boards and solder paste the components on them with pins for my breadboard work. I've yet to have a problem with the smallest of parts, even the ADCs are extremely stable.
Always enjoy Tracy's et al discussions of transistor characteristics.
So I gave this a shot and it works great! I used the RZM001P02T2L part, and I am getting practically no current drain when in shutdown mode.
The only downside is that the RZM001P02T2L part is VERY tiny and not the easiest to solder, but should be a piece of cake for a pick & place machine.
I did run into a bug that I was able to correct via firmware. If you have the device plugged in, then you unplug the device, there is enough stored energy in the capacitors so that the device continues to run even when the shutdown is enabled. This causes the device to behave as if its still plugged in (only for like 10 seconds) and the device continues to sample the battery level, which rapidly decreases because the mosfet is now turned off. After 10 seconds the device fully shuts down and its no longer a problem.
I'm wondering for in the future, if I should also connect the 12V power to the SHDN/LBO/mosfet gate (through a voltage divider to bring it down to 3.3v) that way the mosfet stays enabled when external power is plugged in?
In any case it works great so far, thanks again for the help!
Can anyone share a good method for creating an equivalent shutdown function using mosfets/transistors without having to source a voltage regulator with this functionality built into it?
These transistors seem to fail pretty easily, I'm wondering if I am not using an appropriate transistor? The one I'm currently using is RZM001P02T2L
That transistor does seem to be a bit anemic for a power supply switch. With a 3.8 ohm Rds you would be at 150mW when the circuit is drawing 39mA. Try the RSC002P03T316 (from Mouser) or something like it.
I thought the transistor was there only to break the connection from the battery supply to the ADC ch0 input, microamps of current at most, to prevent leakage via the ADC pin when powered down. Not a real power supply switch. The boost converter has it's own shutdown input. It seems odd that the RZM001P02T2L are failing. How so?
That transistor does seem to be a bit anemic for a power supply switch. With a 3.8 ohm Rds you would be at 150mW when the circuit is drawing 39mA. Try the RSC002P03T316 (from Mouser) or something like it.
I thought the transistor was there only to break the connection from the battery supply to the ADC ch0 input, microamps of current at most, to prevent leakage via the ADC pin when powered down. Not a real power supply switch. The boost converter has it's own shutdown input. It seems odd that the RZM001P02T2L are failing. How so?
Thanks for the responses!
We are hand assembling a few batches of these, and when we test them out, half of the mosfets are bad. Sometimes we have to change them only once, or a couple of times, then they finally work.
I'm wondering if heat is damaging them? These mosfets are super tiny (package SOT-723-3), maybe the assembler is putting too much heat into the part?
I figured I would check back in on this. For some reason, while we are hand assembling these, we always get 50% with bad mosfets and so we have to replace them. Are we really damaging them somehow? Or is it possible there is a design flaw associated with the mosfet?
After we swap them out everything seems to be fine.
MOSFETs are very static sensitive. So you need to be careful how you handle them -- especially in the winter when the inside air is dry. Antistatic mats connected to wrist cuffs may be required in extreme conditions at every workstation where they are handled.
Comments
Notice that they specify it at 1.5 µA of drain current, whereas Rohm RZM001P02T2L specified it at 100µA--The Rohm part is more "turned on" than the Infineon part at about the same threshold. The threshold is not a well controlled parameter in either case, true of mosfets in general Furthermore, the BSS223PW does not show graphs below 2.2V, so I'd tend to prefer the RZM001P02T2L given a choice between the two. The BSS223PW is optimized for fast, low loss turn-on-off in DC converters, for example, the turnoff delay is 5ns versus 325ns, but that is a parameter of no consequence for what you want to do. The RZM001P02T2L is more "general purpose". There are so many different kinds of transistors available because of all the contradictory parameters that might need to be optimized for one application or another.
You said, I'm not sure where your confusion lies, but I do appreciate that it can be daunting. There was that 100 microamp figure for the Vgsth threshold. That isn't a current that will arise in your circuit with this transistor. It is only an indication that the transistor can support 100µA current at the low threshold, and could therefore also support the 1MΩ load without error.
Your schematic there looks good. Can you breadboard this before committing to a pcb?
Thanks again for the info. This is making more sense. In that case I will switch back over to the Rohm RZM001P02T2L part.
I'm curious though with the BSS223PW, the mentioned 1.5uA is at the minimum threshold though right? How do you tell how much current afterwards though? This information is on the graph "5 Typ. output characteristic" right? Would I just estimate where 1.2V might be?
Unfortunately the circuit is not all that practical on a breadboard... I would have to solder small leads to all of these tiny parts and it might not be very reliable.
Always enjoy Tracy's et al discussions of transistor characteristics.
The only downside is that the RZM001P02T2L part is VERY tiny and not the easiest to solder, but should be a piece of cake for a pick & place machine.
I did run into a bug that I was able to correct via firmware. If you have the device plugged in, then you unplug the device, there is enough stored energy in the capacitors so that the device continues to run even when the shutdown is enabled. This causes the device to behave as if its still plugged in (only for like 10 seconds) and the device continues to sample the battery level, which rapidly decreases because the mosfet is now turned off. After 10 seconds the device fully shuts down and its no longer a problem.
I'm wondering for in the future, if I should also connect the 12V power to the SHDN/LBO/mosfet gate (through a voltage divider to bring it down to 3.3v) that way the mosfet stays enabled when external power is plugged in?
In any case it works great so far, thanks again for the help!
These transistors seem to fail pretty easily, I'm wondering if I am not using an appropriate transistor? The one I'm currently using is RZM001P02T2L
Thanks for the responses!
We are hand assembling a few batches of these, and when we test them out, half of the mosfets are bad. Sometimes we have to change them only once, or a couple of times, then they finally work.
I'm wondering if heat is damaging them? These mosfets are super tiny (package SOT-723-3), maybe the assembler is putting too much heat into the part?
Its always open - e.g. providing voltage to the gate does not allow current to flow.
After we swap them out everything seems to be fine.
-Phil