Low output impedance via MOSFET?

[oahsen]

I am supposed to design a power amplifier which has a MOSFET output stage. It should deliver as much power as possible to a 20 ohm load. Therefore, I am trying to make the output impedance as near to 20 ohm as possible.

Thus, obviously I should use a common drain MOSFET. However, the problem is output impedance of a common drain mosfet is approximately 1/gm with gm=2sqrt(KnId). With standart values of Kn we require Id in the order of Amperes to obtain 1/gm as 20 ohm. Typicially I can obtain around 1-2kOhm output impedance with this configuration. So, how can I solve this problem, ie how can I make the output impedance near 20 ohm using a MOSFET?

thanks in advance

 

[erikl]

... It should deliver as much power as possible to a 20 ohm load. Therefore, I am trying to make the output impedance as near to 20 ohm as possible.

No, to achieve this, Rout ≪ 20Ω !

Thus, obviously I should use a common drain MOSFET. However, the problem is output impedance of a common drain mosfet is approximately 1/gm with gm=2sqrt(KnId). ... So, how can I solve this problem, ie how can I make the output impedance near 20 ohm using a MOSFET?

Use a huge MOSFET (w/l = several 10k) as output stage of an opAmp in common source mode. Its inherent output impedance will be divided by the loop gain of the opAmp, s. e.g.

 

[oahsen]

... It should deliver as much power as possible to a 20 ohm load. Therefore, I am trying to make the output impedance as near to 20 ohm as possible.

No, to achieve this, Rout ≪ 20Ω !

Thus, obviously I should use a common drain MOSFET. However, the problem is output impedance of a common drain mosfet is approximately 1/gm with gm=2sqrt(KnId). ... So, how can I solve this problem, ie how can I make the output impedance near 20 ohm using a MOSFET?

Use a huge MOSFET (w/l = several 10k) as output stage of an opAmp in common source mode. Its inherent output impedance will be divided by the loop gain of the opAmp, s. e.g.

thank you, what you are saying is definitely worth trying. at this point I wanted to ask another quick question. How exactly are we going to determine the values of un,cox and W/L from the datasheet of a MOSFET. For instance one of the MOSFETs we can use has the datasheet which I attached.

For id=un*cox*W/L*(Vgs-Vtn)^2 where in the datasheet un,cox and W/L is written?

 

[erikl]

How exactly are we going to determine the values of un,cox and W/L from the datasheet of a MOSFET. For instance one of the MOSFETs we can use has the datasheet which I attached.

For id=un*cox*W/L*(Vgs-Vtn)^2 where in the datasheet un,cox and W/L is written?

You'll never find these values from the datasheet. The datasheet shows the tranfer and saturation characteristics instead. If you need SPICE parameters for hand calculation or simulation, you must try and get the appropriate SPICE model from the manufacturer. Also, simulation tools like PSPICE or LTSPICE possess models of generic devices.

BTW: With an nMOSFET in common source configuration you can only control the negative path (low side) of the power supply. You could, however, use an nMOS in common drain configuration (i.e. as source follower) to control the positive path (high side), but then it won't be an LDO (Low Drop Out) controller. For an LDO controlling the high side, you'd need a pMOS.

 

[FvM]

You didn't yet tell a word about current range and signal frequencies.

 

[oahsen]

You'll never find these values from the datasheet. The datasheet shows the tranfer and saturation characteristics instead. If you need SPICE parameters for hand calculation or simulation, you must try and get the appropriate SPICE model from the manufacturer. Also, simulation tools like PSPICE or LTSPICE possess models of generic devices.

BTW: With an nMOSFET in common source configuration you can only control the negative path (low side) of the power supply. You could, however, use an nMOS in common drain configuration (i.e. as source follower) to control the positive path (high side), but then it won't be an LDO (Low Drop Out) controller. For an LDO controlling the high side, you'd need a pMOS.

As I understood from the Id vs. Vds curve in the datasheet, the gm value for this MOSFET could be quite higher than I thought(even in the order of 10 S). So, how can I verify this with PSPICE. In the datasheet the following "PSPICE Electrical Model" is given:

.SUBCKT FDP8880  2 1 3 
Ca 12 8 9.5e-10
Cb 15 14 9.5e-10
Cin 6 8 1.15e-9
Dbody 7 5 DbodyMOD
Dbreak 5 11 DbreakMOD
Dplcap 10 5 DplcapMOD
Ebreak 11 7 17 18 32.88
Eds 14 8 5 8 1
Egs 13 8 6 8 1
Esg 6 10 6 8 1
Evthres 6 21 19 8 1
Evtemp 20 6 18 22 1
It 8 17 1
Lgate 1 9 5.3e-9
Ldrain 2 5 1.0e-9
Lsource 3 7 1.7e-9
RLgate 1 9 53
RLdrain 2 5 10
RLsource 3 7 17
Mmed 16 6 8 8 MmedMOD
Mstro 16 6 8 8 MstroMOD 
Mweak 16 21 8 8 MweakMOD
Rbreak 17 18 RbreakMOD 1
Rdrain 50 16 RdrainMOD 1.0e-3
Rgate 9 20 2.2
RSLC1 5 51 RSLCMOD 1e-6
RSLC2 5 50 1e3
Rsource 8 7 RsourceMOD 6.8e-3
Rvthres 22 8 RvthresMOD 1
Rvtemp 18 19 RvtempMOD 1
S1a 6 12 13 8 S1AMOD
S1b 13 12 13 8 S1BMOD
S2a 6 15 14 13 S2AMOD
S2b 13 15 14 13 S2BMOD
Vbat 22 19 DC 1
ESLC 51 50  VALUE={(V(5,51)/ABS(V(5,51)))*(PWR(V(5,51)/(1e-6*170),5))}
.MODEL DbodyMOD D (IS=3E-12 IKF=10 N=1.01 RS=5e-3 TRS1=8e-4 TRS2=2e-7
+ CJO=4.8e-10 M=0.55 TT=1e-11 XTI=2)
.MODEL DbreakMOD D (RS=0.2 TRS1=1e-3 TRS2=-8.8e-6)
.MODEL DplcapMOD D (CJO=5.5e-10 IS=1e-30 N=10 M=0.45)
.MODEL MstroMOD NMOS (VTO=2.10 KP=170 IS=1e-30 N=10 TOX=1 L=1u W=1u)
.MODEL MmedMOD NMOS (VTO=1.75 KP=10 IS=1e-30 N=10 TOX=1 L=1u W=1u RG=2.2)
.MODEL MweakMOD NMOS (VTO=1.39 KP=0.05 IS=1e-30 N=10 TOX=1 L=1u W=1u RG=22 RS=0.1) 
.MODEL RbreakMOD RES (TC1=8.0e-4 TC2=-8e-7)
.MODEL RdrainMOD RES (TC1=-12e-3 TC2=.35e-4)
.MODEL RSLCMOD RES (TC1=9e-4 TC2=1e-6)
.MODEL RsourceMOD RES (TC1=5e-3 TC2=1e-6)
.MODEL RvtempMOD RES (TC1=-2.78e-3 TC2=1.5e-6)
.MODEL RvthresMOD RES (TC1=-1e-3 TC2=-8.2e-6)
 MODEL S1AMOD VSWITCH (RON=1e-5 ROFF=0.1 VON=-4 VOFF=-3.5)
.MODEL S1BMOD VSWITCH (RON=1e-5 ROFF=0.1 VON=-3.5 VOFF=-4)
.MODEL S2AMOD VSWITCH (RON=1e-5 ROFF=0.1 VON=-1.3 VOFF=-0.8)
.MODEL S2BMOD VSWITCH (RON=1e-5 ROFF=0.1 VON=-0.8 VOFF=-1.3)
.ENDS

I tried to put a generic n-MOSFET, then "Edit-Model-Edit Instance Model" and change the text with the previous code. However, since the given model seems to be a subcircuit, obviously my method is not working . So how can I add this MOSFET to my design?

Added after 3 minutes:

You didn't yet tell a word about current range and signal frequencies.

I just need to amplify a 1v-pp, 30kHz signal coming from an ADC circuit, so that it will dissipate around 1W on a 20ohm load.

 

[FvM]

A single transistor means class A operation. If the load is resistive, the DC bias current causes a dissipation twice the signal power
(assuming a sine signal). Is this what you intend? Otherwise, you should think about a push-pull amplifier.

 

[oahsen]

A single transistor means class A operation. If the load is resistive, the DC bias current causes a dissipation twice the signal power
(assuming a sine signal). Is this what you intend? Otherwise, you should think about a push-pull amplifier.

indeed, today someone has told me that I could use push-pull amplifier. However, currently the main problem is whatever my design is, I can not simulate it since I do not know how to add the MOSFET model to the PSPICE as I explained in my previous post.

Any suggestions on this problem?