If anyone wants to use
Exicon Lateral MOSFETs
you should download Christers submitted ZIP file
from this post:
http://www.diyaudio.com/forums/showthread.php?postid=362324#post362324
Exicon can be found with datasheets at http://www.profusionplc.com/
Profusion has got very good other transistor, too.
Like some excellent SANYO driver transistors in TO-126 and TO-92
with Low cob capacitance = 2 - 3 pF
Exicon Lateral MOSFETs
you should download Christers submitted ZIP file
from this post:
http://www.diyaudio.com/forums/showthread.php?postid=362324#post362324
Exicon can be found with datasheets at http://www.profusionplc.com/
Profusion has got very good other transistor, too.
Like some excellent SANYO driver transistors in TO-126 and TO-92
with Low cob capacitance = 2 - 3 pF
hey guys, I have a few newbie questions I hope you don't mind:
1) When simulating THD of an amplifier, do we simulate it at the close loop gain we normally use or do we simulate it at unity gain instead?
2) I'm not sure if I simulated correctly, but using TINA SPICE here's how I do it:
-Insert an input voltage of 1V Sin wave at 1000Hz
-then run fourier analysis.
3) Anyone here has commonly used LEDs meant to bias the folded cascode stage?
1) When simulating THD of an amplifier, do we simulate it at the close loop gain we normally use or do we simulate it at unity gain instead?
2) I'm not sure if I simulated correctly, but using TINA SPICE here's how I do it:
-Insert an input voltage of 1V Sin wave at 1000Hz
-then run fourier analysis.
3) Anyone here has commonly used LEDs meant to bias the folded cascode stage?
Many thanks TzeYang,
Much appreciated!
However, I can confirm it does not work well; I've found that if Ic and Vceo come up on these models as zero, then the model is bad. Certainly the results in my simmed amp are not appropriate; a 5401/5550 gives near perfect results notwithstanding poor current capacity, but these models give sub par gain overall, which is not good.
The usual turf wars between manufacturers, it seems.
Anyone have a model for 15030/31 that works well?
Hugh
Much appreciated!
However, I can confirm it does not work well; I've found that if Ic and Vceo come up on these models as zero, then the model is bad. Certainly the results in my simmed amp are not appropriate; a 5401/5550 gives near perfect results notwithstanding poor current capacity, but these models give sub par gain overall, which is not good.
The usual turf wars between manufacturers, it seems.
Anyone have a model for 15030/31 that works well?
Hugh
Hi Hugh,
Glen and OS have both seen problems with the MJE15030/15031 OnSemi models. Here is the part of the Frugalamp thread where it's discussed. I agree with jaycee that Fairchild's models tend to be pretty good.
You might try the 2SC4793/2SA1837. I created these. They shouldn't be used above 1A though.
Edit: Welcome to the SPICE world Hugh!
.MODEL Q2SC4793 NPN (
+ IS=1.8E-09
+ NF=1.43
+ BF=146.38
+ VAF=273
+ IKF=2.6
+ NK=0.95
+ ISE=6.286997E-10
+ NE=2.223629
+ BR=4
+ NR=1
+ VAR=20
+ IKR=1.05
+ RE=0
+ RB=1.7
+ RC=1.25
+ CJE=5.96964E-10
+ VJE=1.1
+ MJE=0.5
+ CJC=5.78E-11
+ VJC=0.3
+ MJC=0.3
+ TF=1.22678E-09
+ FC=0.5
+ ITF=10
+ XTF=99.52253015
+ TR=983N)
.MODEL Q2SA1837 PNP (
+ IS=2.39372559E-10
+ NF=1.304015937
+ BF=300
+ VAF=273
+ IKF=2.087725944
+ NK=0.94719458
+ ISE=1.46829699E-11
+ NE=1.526663542
+ BR=4
+ NR=1
+ VAR=20
+ IKR=1.05
+ RE=0
+ RB=1.8
+ RC=1.65
+ CJE=4.7407E-10
+ VJE=1.1
+ MJE=0.5
+ CJC=8.6700E-11
+ VJC=0.3
+ MJC=0.3
+ TF=1.642191E-09
+ FC=0.5
+ ITF=1.076260106
+ XTF=5.868994022
+ TR=1.38U)
Glen and OS have both seen problems with the MJE15030/15031 OnSemi models. Here is the part of the Frugalamp thread where it's discussed. I agree with jaycee that Fairchild's models tend to be pretty good.
You might try the 2SC4793/2SA1837. I created these. They shouldn't be used above 1A though.
Edit: Welcome to the SPICE world Hugh!
.MODEL Q2SC4793 NPN (
+ IS=1.8E-09
+ NF=1.43
+ BF=146.38
+ VAF=273
+ IKF=2.6
+ NK=0.95
+ ISE=6.286997E-10
+ NE=2.223629
+ BR=4
+ NR=1
+ VAR=20
+ IKR=1.05
+ RE=0
+ RB=1.7
+ RC=1.25
+ CJE=5.96964E-10
+ VJE=1.1
+ MJE=0.5
+ CJC=5.78E-11
+ VJC=0.3
+ MJC=0.3
+ TF=1.22678E-09
+ FC=0.5
+ ITF=10
+ XTF=99.52253015
+ TR=983N)
.MODEL Q2SA1837 PNP (
+ IS=2.39372559E-10
+ NF=1.304015937
+ BF=300
+ VAF=273
+ IKF=2.087725944
+ NK=0.94719458
+ ISE=1.46829699E-11
+ NE=1.526663542
+ BR=4
+ NR=1
+ VAR=20
+ IKR=1.05
+ RE=0
+ RB=1.8
+ RC=1.65
+ CJE=4.7407E-10
+ VJE=1.1
+ MJE=0.5
+ CJC=8.6700E-11
+ VJC=0.3
+ MJC=0.3
+ TF=1.642191E-09
+ FC=0.5
+ ITF=1.076260106
+ XTF=5.868994022
+ TR=1.38U)
Thank you Andy,
Much appreciated!
I have found it useful already; it's compressed development time considerably, enabling me to home in on stability issues very quickly. The open loop techniques have been very useful establishing accurate levels of feedback, nested and global, and this too has been a great advance.
It tells me nothing of the sound quality, of course, but I can get through the engineering issues much more quickly.
Thanks again, I have your 4793/1837, find them very good indeed, and use these as drivers in some of my other amps.
Hugh
Much appreciated!
I have found it useful already; it's compressed development time considerably, enabling me to home in on stability issues very quickly. The open loop techniques have been very useful establishing accurate levels of feedback, nested and global, and this too has been a great advance.
It tells me nothing of the sound quality, of course, but I can get through the engineering issues much more quickly.
Thanks again, I have your 4793/1837, find them very good indeed, and use these as drivers in some of my other amps.
Hugh
Not always true , hugh... The ( ) header is what gives LT
the data to place device info in the listing. All the generic
ones that come with LT have it , but some (almost all)
of the 3rd party .models omit it.
The fairchild models have extra data and show more
realistic conditions within a design..
As a test I designed a new amp with MPSA 94/44's
using the old models .I had a easy time getting it to be
stable, but when I switched to the new fairchild KSP94/44
(2007 models) ,my triple OPS oscillated and saturated.
(in a quite realistic way
)Thank you andy for new 4793/1837's ,old japanese models
I have are much less verbose (data wise).
OS
Hope they catch the SCURVY B@$$tards that set some of them..
If you want to correct this small inconvenience ,here is an example...
.MODEL KSP44 NPN
+ IS=2.4446E-13 BF=184.79 VAF=100 IKF=9.6380E-2
+ ISE=2.8775E-12 BR=0.17129 VAR=100 IKR=2.2351
+ ISC=6.0845E-10 NC=1.5000 RE=0.1 RB=4.7095
+ RC=2.7922 CJE=1.0100E-10 CJC=1.9010E-11 VJC=0.66207
+ MJC=0.40238 TF=1.1000E-8 XTF=10 VTF=10
+ ITF=1 TR=3.86493E-5 FC=0.5 XTI=3
+ EG=1.11
+ Vceo=400 Icrating=.3 mfg=fairchild
The last 3 entries is the data the LT device listing reflects
(+ Vceo=400 ,Icrating=.3 , mfg=fairchild)
It works good ..as reflected here: (attached)
Attachments
Good advice.. turn off all compression..
I was just doing some FFT's and noticed this thread..
Andy C. has a good way to do it.
Add this as a general LT comment:
.param timestep 1m/65535
.tran 0 10m 0 {timestep}
.options plotwinsize=0
.four 1k
Notice the {timestep}....
add this to the AMPLITUDE of your signal source;
{sqrt(2)}
The "2" will drive the amp to about 30 db , near full power
depending on the amp. (A '1" in the equation will give the 20db
below)
simulated by itself the signal source has noise level of -200db
, Much better than a "strait up" generator.
here is what you might see (-110db .0001%)
os
I was just doing some FFT's and noticed this thread..
Andy C. has a good way to do it.
Add this as a general LT comment:
.param timestep 1m/65535
.tran 0 10m 0 {timestep}
.options plotwinsize=0
.four 1k
Notice the {timestep}....
add this to the AMPLITUDE of your signal source;
{sqrt(2)}
The "2" will drive the amp to about 30 db , near full power
depending on the amp. (A '1" in the equation will give the 20db
below)
simulated by itself the signal source has noise level of -200db
, Much better than a "strait up" generator.
here is what you might see (-110db .0001%)
os
Attachments
Thanks for information
Experienced both forms, the first simulation, Total Harmonic Distortion: 2.12%
then enable ASCII data files Total Harmonic Distortion: 0.77%
I be much distortion for a signal "pure" sine
Netlist:
V1 N001 0 SINE(0 1 1K)
R1 N001 0 1K
.four 1K 9 V(N001)
.tran 0 1 0.01
.option plotwinsize=0
.backanno
.end
Experienced both forms, the first simulation, Total Harmonic Distortion: 2.12%
then enable ASCII data files
Total Harmonic Distortion: 0.77%I be much distortion for a signal "pure" sine
Netlist:
V1 N001 0 SINE(0 1 1K)
R1 N001 0 1K
.four 1K 9 V(N001)
.tran 0 1 0.01
.option plotwinsize=0
.backanno
.end