As far as the ones I could check on LTS models, none had KF or AF parameters. Maybe it's disguised?
Where's 1/f in the THAT models?
.MODEL QPNP_THAT_HF PNP (IS=2.330327E-15 BF=80 NF=0.998 BR=54.6051301 NR=1 ISE=6.471784E-14 NE=1.5256667 ISC=9.821992E-14 NC=1.4090231 VAF=50 VAR=8.5915034 IKF=0.1605012 IKR=2.997951E-3 RB=103.345 RBM=2.20352 IRB=1.3734E-4 RE=0.2 RC=17.9803001 CJE=5.921046E-12 VJE=0.7302802 MJE=0.49864 FC=0.9317095 CJC=3.374991E-12 VJC=0.6162124 MJC=0.497456 TF=2.416792E-10 XTF=5.9184348 ITF=0.0726275 VTF=1.2378779 QCO=1E-12 RCO=1 VO=1 GAMMA=1E-13 XTB=1.2 PTF=20)
.MODEL QNPN_THAT_HF NPN (IS=5.005475E-15 BF=150 NF=1 BR=119.4532856 NR=1 ISE=3.002016E-16 NE=1.3340565 ISC=3.173175E-14 NC=1.6525 VAF=60 VAR=10.8094728 IKF=0.2850543 IKR=0.0103675 RB=117.3192 RBM=2.035 IRB=1.750788E-4 RE=0.2 RC=12.4561714 CJE=5.086292E-12 VJE=0.7241957 MJE=0.49456 FC=0.97 CJC=2.190824E-12 VJC=0.5595858 MJC=0.498675 TF=7.063932E-11 XTF=17.846692 ITF=0.0862198 VTF=1.7447209 QCO=1E-12 RCO=1 VO=1 GAMMA=1E-13 XTB=1.1 PTF=20)
Where's 1/f in the THAT models?
.MODEL QPNP_THAT_HF PNP (IS=2.330327E-15 BF=80 NF=0.998 BR=54.6051301 NR=1 ISE=6.471784E-14 NE=1.5256667 ISC=9.821992E-14 NC=1.4090231 VAF=50 VAR=8.5915034 IKF=0.1605012 IKR=2.997951E-3 RB=103.345 RBM=2.20352 IRB=1.3734E-4 RE=0.2 RC=17.9803001 CJE=5.921046E-12 VJE=0.7302802 MJE=0.49864 FC=0.9317095 CJC=3.374991E-12 VJC=0.6162124 MJC=0.497456 TF=2.416792E-10 XTF=5.9184348 ITF=0.0726275 VTF=1.2378779 QCO=1E-12 RCO=1 VO=1 GAMMA=1E-13 XTB=1.2 PTF=20)
.MODEL QNPN_THAT_HF NPN (IS=5.005475E-15 BF=150 NF=1 BR=119.4532856 NR=1 ISE=3.002016E-16 NE=1.3340565 ISC=3.173175E-14 NC=1.6525 VAF=60 VAR=10.8094728 IKF=0.2850543 IKR=0.0103675 RB=117.3192 RBM=2.035 IRB=1.750788E-4 RE=0.2 RC=12.4561714 CJE=5.086292E-12 VJE=0.7241957 MJE=0.49456 FC=0.97 CJC=2.190824E-12 VJC=0.5595858 MJC=0.498675 TF=7.063932E-11 XTF=17.846692 ITF=0.0862198 VTF=1.7447209 QCO=1E-12 RCO=1 VO=1 GAMMA=1E-13 XTB=1.1 PTF=20)
I have these models for the THAT devices:
MODEL QPNP_THAT_NS PNP (
+ IS = 2.330327E-15 BF = 80
+ NF = 0.998 BR = 54.6051301
+ NR = 1 ISE = 6.471784E-14 NE = 1.5256667
+ ISC = 9.821992E-14 NC = 1.4090231 VAF = 50
+ VAR = 8.5915034 IKF = 0.1605012 IKR = 2.997951E-3
+ RB = 25 RBM = 1.604356 IRB = 1.3734E-4
+ RE = 0.2 RC = 17.9803001 CJE = 5.921046E-12
+ VJE = 0.7302802 MJE = 0.49864 FC = 0.9317095
+ CJC = 3.374991E-12 VJC = 0.6162124 MJC = 0.497456
+ TF = 2.416792E-10 XTF = 5.9184348 ITF = 0.0726275
+ VTF = 1.2378779 QCO = 1E-12 RCO = 1
+ VO = 1 GAMMA = 1E-13 XTB = 1.2
+ PTF = 20 AF = 1 KF = 1.4e-16)
*
and
.MODEL QNPN_THAT_NS NPN (
+ IS = 5.005475E-15 BF = 150
+ NF = 1 BR = 119.4532856
+ NR = 1 ISE = 3.002016E-16 NE = 1.3340565
+ ISC = 3.173175E-14 NC = 1.6525 VAF = 60
+ VAR = 10.8094728 IKF = 0.2850543 IKR = 0.0103675
+ RB = 30 RBM = 2.94235 IRB = 1.750788E-4
+ RE = 0.2 RC = 12.4561714 CJE = 5.086292E-12
+ VJE = 0.7241957 MJE = 0.49456 FC = 0.97
+ CJC = 2.190824E-12 VJC = 0.5595858 MJC = 0.498675
+ TF = 7.063932E-11 XTF = 17.846692 ITF = 0.0862198
+ VTF = 1.7447209 QCO = 1E-12 RCO = 1
+ VO = 1 GAMMA = 1E-13 XTB = 1.1
+ PTF = 20 AF = 1 KF = 7e-16)
*
Both of which have AF (unity) and KF ~10^-15
But I downloaded those models in 2017. The latest models dated 2018, and as you show above, seem to have mysteriously deleted 1/f noise. Weird,
MODEL QPNP_THAT_NS PNP (
+ IS = 2.330327E-15 BF = 80
+ NF = 0.998 BR = 54.6051301
+ NR = 1 ISE = 6.471784E-14 NE = 1.5256667
+ ISC = 9.821992E-14 NC = 1.4090231 VAF = 50
+ VAR = 8.5915034 IKF = 0.1605012 IKR = 2.997951E-3
+ RB = 25 RBM = 1.604356 IRB = 1.3734E-4
+ RE = 0.2 RC = 17.9803001 CJE = 5.921046E-12
+ VJE = 0.7302802 MJE = 0.49864 FC = 0.9317095
+ CJC = 3.374991E-12 VJC = 0.6162124 MJC = 0.497456
+ TF = 2.416792E-10 XTF = 5.9184348 ITF = 0.0726275
+ VTF = 1.2378779 QCO = 1E-12 RCO = 1
+ VO = 1 GAMMA = 1E-13 XTB = 1.2
+ PTF = 20 AF = 1 KF = 1.4e-16)
*
and
.MODEL QNPN_THAT_NS NPN (
+ IS = 5.005475E-15 BF = 150
+ NF = 1 BR = 119.4532856
+ NR = 1 ISE = 3.002016E-16 NE = 1.3340565
+ ISC = 3.173175E-14 NC = 1.6525 VAF = 60
+ VAR = 10.8094728 IKF = 0.2850543 IKR = 0.0103675
+ RB = 30 RBM = 2.94235 IRB = 1.750788E-4
+ RE = 0.2 RC = 12.4561714 CJE = 5.086292E-12
+ VJE = 0.7241957 MJE = 0.49456 FC = 0.97
+ CJC = 2.190824E-12 VJC = 0.5595858 MJC = 0.498675
+ TF = 7.063932E-11 XTF = 17.846692 ITF = 0.0862198
+ VTF = 1.7447209 QCO = 1E-12 RCO = 1
+ VO = 1 GAMMA = 1E-13 XTB = 1.1
+ PTF = 20 AF = 1 KF = 7e-16)
*
Both of which have AF (unity) and KF ~10^-15
But I downloaded those models in 2017. The latest models dated 2018, and as you show above, seem to have mysteriously deleted 1/f noise. Weird,
Last edited:
So I'm doing a bit of a random search through the spice database in Micro-Cap 12, and found MMBT6429LT1, which looked like the a real gem. 1/f included in model - check. Very low RB (the spice parameter for rbb), check.
So, I thought I'd buy a bunch and test them.
Recently obsoleted by our friend/enemy On-Semi. No stock easily available.
So, I thought I'd buy a bunch and test them.
Recently obsoleted by our friend/enemy On-Semi. No stock easily available.
What's the use of trying a part which is obsolete?
How can we that missing data from those KSA/KSC parts?
Sorry, which is the complementary PNP for that transistor? We need both.
Why so many? How many preamps do you need?
I'm really thankful for this info, but would you please do something for me?
Given the fact that you have a curve tracer, would you, please, make some measurements to see if the quasi saturation effect that Mark saw in mpsa18 is still there?
https://www.diyaudio.com/forums/sol...asi-saturation-measured-data.html#post4755597
I'm involved in a project that has nothing to do with this thread, that involves low voltage noise transistors with a high hFE. So I needed enough with spares. And they are cheap. 500 cost UKP24 including tax and shipping. so 4.8p each - say 7 cents.
The TO92 version from Central semiconductor (the only manufacturer of the MPSA18) is nearly 10x more expensive.
And yes - they are NPN only. So what? Make a long tail pair with decent gain and shove the output into whatever opamp you like. Then use active or passive RIAA, your choice.
Anyhow, I'm trying to he helpful is all. If you guys are going to bust my buns, I'll shut up right now.
That is interesting! Thanks for the link. I have some SM to leaded adapters, so when they arrive I'll certainly run them through the 576.
Just looked through the curves - now I see what we're talking about.
Last edited:
OK - my phone has got no capability of adjusting exposure time. And my DSLR battery is on charge.
However verbally the saturation effect of the MMBT6429LT1G, although present, is nothing like as bad as the MPSA18 in the link above. Set up as close to Mark's scaling, the saturation shoulder starts at 7.5mA IC (the MPA18 was 5mA) and has totally finished by 1.5V (compared with 1.85V)
And the shoulder is much more linear, as compared with the Mark's MPSA18 result, which is much more rounded in appearance.
Of course that is not ideal. But provided the bias point device is used at all points with Vce>a few volts this behaviour should not be a problem.
The beta is huge - about 900.
A pic will be forthcoming once the camera battery is charged.
I've done a few more devices, and I'll run them again. Tentative conclusion (which I might revise when I re-run them!) is that the pnp complement has better performance regarding this saturation effect that their npn counterpart.
However verbally the saturation effect of the MMBT6429LT1G, although present, is nothing like as bad as the MPSA18 in the link above. Set up as close to Mark's scaling, the saturation shoulder starts at 7.5mA IC (the MPA18 was 5mA) and has totally finished by 1.5V (compared with 1.85V)
And the shoulder is much more linear, as compared with the Mark's MPSA18 result, which is much more rounded in appearance.
Of course that is not ideal. But provided the bias point device is used at all points with Vce>a few volts this behaviour should not be a problem.
The beta is huge - about 900.
A pic will be forthcoming once the camera battery is charged.
I've done a few more devices, and I'll run them again. Tentative conclusion (which I might revise when I re-run them!) is that the pnp complement has better performance regarding this saturation effect that their npn counterpart.
What I meant is explained in more detail in the attachment, I hope it will be of use to you and to whoever may read this.
Thank you Marcel - that makes a lot of sense.
So calculating the open loop gain (as you'd normally do to check for gain and phase margins) is actually misleading, because it overestimates the input capacitance. You only get the first stage Miller capacitance modification/reduction once you close the loop, provided the second stage gain is high enough.
Am I reading that right?
So calculating the open loop gain (as you'd normally do to check for gain and phase margins) is actually misleading, because it overestimates the input capacitance. You only get the first stage Miller capacitance modification/reduction once you close the loop, provided the second stage gain is high enough.
Am I reading that right?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.