| Electronics > Projects, Designs, and Technical Stuff |
| Saturating a MOSFET with a constant current sink driver |
| << < (3/7) > >> |
| Oleenick:
--- Quote from: mikerj on January 03, 2020, 12:38:37 pm --- --- Quote from: Oleenick on January 01, 2020, 11:29:55 am ---Would you have any idea if using a constant current sink would be an issue on the gate? --- End quote --- It will be fine, provided: 1) R15 is a high enough value that the current sink output saturates. No problem with a 30mA current set point. 2) The saturated voltage is sufficient to fully turn on the MOSFET. No problem, datasheet characteristics suggest you will be within ~100mV of 0V at 5mA. 3) The zero current gate voltage is close enough to Vcc to fully turn off the MOSFET. Leakage is only 1uA max so again no problem. --- End quote --- I'd just like to clarify my understanding of your help. 1) Do you mean that the constant current from CLVL0 will flow through R15 (30mA), which will make the gate see a low enough voltage to saturate the MOSFET? 2) Do you mean that the datasheet suggests that for the transistor to turn on it must be within less than + or - 0.100mV? 3) So when CLVL0 is high, the current through R15 to VCC will be low enough to turn off the MOSFET? I've revised the circuit but I'm considering switching to an N-channel MOSFET which has better thermal characteristics. Should I be using a resistor on the gate (between CLVL0 and gate) even though I'm using a constant current source? I'm not sure if thats how it works, my understanding is that it should reduce oscillation on the edges of the switching singal to the gate. Because of my current P-channel MOSFET RDSon value of 0.1ohm I'd like to find something less than 20mOhm RDSon. I don't really need it to be P-channel. I'm struggling to source a nice MOSFET from LCSC. Something close to the CSD18542KTT from TI would be good but I can't find anything like it on LCSC. EDIT: I think I found a good one from LCSC, here https://datasheet.lcsc.com/szlcsc/1808281537_Infineon-Technologies-IRLS3036TRLPBF_C73577.pdf It's an N-channel logic level device, which has a very low RDSon max at 2.4mOhm and a PD max of 3.75W while my load will dissipate only 2.4W |
| mikerj:
--- Quote from: Oleenick on January 06, 2020, 10:40:18 am --- --- Quote from: mikerj on January 03, 2020, 12:38:37 pm --- --- Quote from: Oleenick on January 01, 2020, 11:29:55 am ---Would you have any idea if using a constant current sink would be an issue on the gate? --- End quote --- It will be fine, provided: 1) R15 is a high enough value that the current sink output saturates. No problem with a 30mA current set point. 2) The saturated voltage is sufficient to fully turn on the MOSFET. No problem, datasheet characteristics suggest you will be within ~100mV of 0V at 5mA. 3) The zero current gate voltage is close enough to Vcc to fully turn off the MOSFET. Leakage is only 1uA max so again no problem. --- End quote --- I'd just like to clarify my understanding of your help. 1) Do you mean that the constant current from CLVL0 will flow through R15 (30mA), which will make the gate see a low enough voltage to saturate the MOSFET? --- End quote --- Correct. However the actual current flowing through R15 may be limited by the supply voltage rather than the TLC5927 target current, in which case it's output will saturate somewhere close to 0v. For some reason I thought that Vcc was 5v, but checking again I'm not sure you specified this value and it is important. Not only does Vgs need to be high enough to ensure the MOSFET switches on, it also has to be below the maximum Vgs permitted by the MOSFET or you will kill it almost instantly. --- Quote from: Oleenick on January 06, 2020, 10:40:18 am ---2) Do you mean that the datasheet suggests that for the transistor to turn on it must be within less than + or - 0.100mV? --- End quote --- No, I simply mean that Vgs must be high enough to fully saturate the transistor at the drain current you require. Vgs in this case is the difference between Vcc and the output of the TLC5927. The 100mV was based on my (faulty) assumption that Vcc was 5v giving a maximum current into the TLC5927 of 5mA (5v/1k), again you need to specify this value if it's something else. --- Quote from: Oleenick on January 06, 2020, 10:40:18 am ---3) So when CLVL0 is high, the current through R15 to VCC will be low enough to turn off the MOSFET? --- End quote --- Correct, with the output of the TLC5927 switched off, the voltage applied to the gate must be (significantly) lower than Vgs(thr) to ensure the MOSFET is fully turned off. --- Quote from: Oleenick on January 06, 2020, 10:40:18 am ---I've revised the circuit but I'm considering switching to an N-channel MOSFET which has better thermal characteristics. --- End quote --- Please turn the transistor around rather than the voltage rails! Most people expect the higher potential rail to be at the top of the schematic and this makes reading the schematic easier i.e. conventional current flow from top to bottom. [ Attachment Invalid Or Does Not Exist ] --- Quote from: Oleenick on January 06, 2020, 10:40:18 am ---Should I be using a resistor on the gate (between CLVL0 and gate) even though I'm using a constant current source? I'm not sure if thats how it works, my understanding is that it should reduce oscillation on the edges of the switching singal to the gate. --- End quote --- Shouldn't make any difference in this case as the impedance of the output during transition is determined by R15. If Vcc is high enough that the Vgs is close to or exceeding the maximum for the device chosen then you will have to add additional components to limit it. This could be as simple another resistor to form a potential divider with R15. |
| Oleenick:
--- Quote from: mikerj on January 06, 2020, 11:47:02 am --- --- Quote from: Oleenick on January 06, 2020, 10:40:18 am --- --- Quote from: mikerj on January 03, 2020, 12:38:37 pm --- --- Quote from: Oleenick on January 01, 2020, 11:29:55 am ---Would you have any idea if using a constant current sink would be an issue on the gate? --- End quote --- It will be fine, provided: 1) R15 is a high enough value that the current sink output saturates. No problem with a 30mA current set point. 2) The saturated voltage is sufficient to fully turn on the MOSFET. No problem, datasheet characteristics suggest you will be within ~100mV of 0V at 5mA. 3) The zero current gate voltage is close enough to Vcc to fully turn off the MOSFET. Leakage is only 1uA max so again no problem. --- End quote --- I'd just like to clarify my understanding of your help. 1) Do you mean that the constant current from CLVL0 will flow through R15 (30mA), which will make the gate see a low enough voltage to saturate the MOSFET? --- End quote --- Correct. However the actual current flowing through R15 may be limited by the supply voltage rather than the TLC5927 target current, in which case it's output will saturate somewhere close to 0v. For some reason I thought that Vcc was 5v, but checking again I'm not sure you specified this value and it is important. Not only does Vgs need to be high enough to ensure the MOSFET switches on, it also has to be below the maximum Vgs permitted by the MOSFET or you will kill it almost instantly. --- Quote from: Oleenick on January 06, 2020, 10:40:18 am ---2) Do you mean that the datasheet suggests that for the transistor to turn on it must be within less than + or - 0.100mV? --- End quote --- No, I simply mean that Vgs must be high enough to fully saturate the transistor at the drain current you require. Vgs in this case is the difference between Vcc and the output of the TLC5927. The 100mV was based on my (faulty) assumption that Vcc was 5v giving a maximum current into the TLC5927 of 5mA (5v/1k), again you need to specify this value if it's something else. --- Quote from: Oleenick on January 06, 2020, 10:40:18 am ---3) So when CLVL0 is high, the current through R15 to VCC will be low enough to turn off the MOSFET? --- End quote --- Correct, with the output of the TLC5927 switched off, the voltage applied to the gate must be (significantly) lower than Vgs(thr) to ensure the MOSFET is fully turned off. --- Quote from: Oleenick on January 06, 2020, 10:40:18 am ---I've revised the circuit but I'm considering switching to an N-channel MOSFET which has better thermal characteristics. --- End quote --- Please turn the transistor around rather than the voltage rails! Most people expect the higher potential rail to be at the top of the schematic and this makes reading the schematic easier i.e. conventional current flow from top to bottom. --- Quote from: Oleenick on January 06, 2020, 10:40:18 am ---Should I be using a resistor on the gate (between CLVL0 and gate) even though I'm using a constant current source? I'm not sure if thats how it works, my understanding is that it should reduce oscillation on the edges of the switching singal to the gate. --- End quote --- Shouldn't make any difference in this case as the impedance of the output during transition is determined by R15. If Vcc is high enough that the Vgs is close to or exceeding the maximum for the device chosen then you will have to add additional components to limit it. This could be as simple another resistor to form a potential divider with R15. --- End quote --- Thank you thats very valuable help. I am using +5V as VCC, I should have stated that earlier. Also I think I will be going with the IRLS3036TRLPBF N-channel MOSFET now anyways because of its very low RDSon. That will fix my thermal issues as I'd like it to be capable of a continuous current of each layer just incase the program hangs during switching between the 8 levels. There's the schematic that I would follow with the new MOSFET. The resistor to ground would be 620ohms and the capacitor 100uF. In this case should I use a resistor on the gate? |
| Wimberleytech:
--- Quote from: Oleenick on January 06, 2020, 11:54:02 am ---I am using +5V as VCC, I should have stated that earlier. Also I think I will be going with the IRLS3036TRLPBF N-channel MOSFET now anyways because of its very low RDSon. That will fix my thermal issues as I'd like it to be capable of a continuous current of each layer just incase the program hangs during switching between the 8 levels. (Attachment Link) There's the schematic that I would follow with the new MOSFET. The resistor to ground would be 620ohms and the capacitor 100uF. In this case should I use a resistor on the gate? --- End quote --- The threshold voltage for that transistor is min of 1v and max of 2.5 volts. Therefore, your output voltage will be a maximum of about 4 volts and a minimum of about 2.5 volts. A gate resistor is not necessary. |
| Oleenick:
--- Quote from: Wimberleytech on January 06, 2020, 02:18:30 pm --- --- Quote from: Oleenick on January 06, 2020, 11:54:02 am ---I am using +5V as VCC, I should have stated that earlier. Also I think I will be going with the IRLS3036TRLPBF N-channel MOSFET now anyways because of its very low RDSon. That will fix my thermal issues as I'd like it to be capable of a continuous current of each layer just incase the program hangs during switching between the 8 levels. (Attachment Link) There's the schematic that I would follow with the new MOSFET. The resistor to ground would be 620ohms and the capacitor 100uF. In this case should I use a resistor on the gate? --- End quote --- The threshold voltage for that transistor is min of 1v and max of 2.5 volts. Therefore, your output voltage will be a maximum of about 4 volts and a minimum of about 2.5 volts. A gate resistor is not necessary. --- End quote --- That is potentially problematic. Do you mean that Vds (output voltage) can only be 2.5-4V? Where did you find that on the datasheet? Or by output voltage do you mean gate voltage? This figure from the datasheet shows that at a voltage of ~3V Id can be 10A, which would be enough for my application. Do you mean that the gate voltage needs to be between 2.5 and 4V to achieve the current I need? |
| Navigation |
| Message Index |
| Next page |
| Previous page |