Yes, lovely, this diplomacy. My congrats to all - Syn08, Andy_C, Stinius, Jkeny, all is quiet on the western front.....
To me, here are the questions:
To what extent is demonstrated performance reliant upon:
1. Equality of base currents in the two drivers,
2. Parameter drift in semiconductors use, and
3. Variations in saturation voltage (Vce) of the two bias control transistors?
I really must sim this, but I'm concerned about the integrity of the models, which have profound effect, particularly in this case where saturation models are evidently not well described.
Cheers,
Hugh
To me, here are the questions:
To what extent is demonstrated performance reliant upon:
1. Equality of base currents in the two drivers,
2. Parameter drift in semiconductors use, and
3. Variations in saturation voltage (Vce) of the two bias control transistors?
I really must sim this, but I'm concerned about the integrity of the models, which have profound effect, particularly in this case where saturation models are evidently not well described.
Cheers,
Hugh
Models
Here are the models; they are actually available to download from the Internet.
Just a friendly warning: these models are, at best, a joke. The power devices that are supposed to be complementary are (almost) copy/paste with reversed (PNP/NPN) polarity. This provides kinda ideal matching which is, of course, far from reality.
I delivered what I promised, I'm going to hold any other comments for now.
Edit-1: as about saturation modelling, don't even think about.
Edit-2: it would be interesting to simulate an EF triple with the same models and compare the results.
Here are the models; they are actually available to download from the Internet.
Just a friendly warning: these models are, at best, a joke. The power devices that are supposed to be complementary are (almost) copy/paste with reversed (PNP/NPN) polarity. This provides kinda ideal matching which is, of course, far from reality.
I delivered what I promised, I'm going to hold any other comments for now.
Edit-1: as about saturation modelling, don't even think about.
Edit-2: it would be interesting to simulate an EF triple with the same models and compare the results.
Re: Models
As far as I remember, Steve wrote about the necessity to match devices. So, it seems that the models are correct for a situation when devices are actually matched.
syn08 said:
As far as I remember, Steve wrote about the necessity to match devices. So, it seems that the models are correct for a situation when devices are actually matched.
Re: Re: Re: Models
Quadrature is actually imaginary, in any case we are all in harmony at last.
syn08 said:
Quadrature is actually imaginary, in any case we are all in harmony at last.
Old models..
They seem to lack many of the parameters in newer
fairchild models (no Rc ,Mjc..etc , many others).
They might work to model a shunt regulator , but not a
krill..
OS
They seem to lack many of the parameters in newer
fairchild models (no Rc ,Mjc..etc , many others).
They might work to model a shunt regulator , but not a
krill..
OS
Re: Models
I'm confused. I thought the models you were looking for were the Intusoft models? Or are these the same as the device vendor's models? Are the Intusoft models downloadable from the internet?
syn08 said:
I'm confused. I thought the models you were looking for were the Intusoft models? Or are these the same as the device vendor's models? Are the Intusoft models downloadable from the internet?
as an example...
The old 1991 model:
.MODEL QPWR PNP (IS=1.63P NF=1 BF=130 VAF=254 IKF=11 ISE=1.34N NE=2
+ BR=4 NR=1 VAR=20 IKR=16.5 RE=12.1M RB=4 RBM=.4 IRB=5.556U RC=4.84M
+ CJE=1.09N VJE=.6 MJE=.3 CJC=708P VJC=.22 MJC=.2 TF=5.33N TR=204N)
+ XTB=1.5 PTF=120 XTF=1 ITF=9.6)
One version I have , I think have a 2001 version (I can't find it)
* created using Parts release 7.1p on 07/01/96 at 10:31
* missing storage time & output admittance
.MODEL Q2SA1302 PNP
+ IS=21.479E-12
+ BF=136.48
+ VAF=100
+ IKF=19.980
+ ISE=21.504E-12
+ NE=1.3784
+ BR=329.48
+ VAR=100
+ IKR=19.980
+ ISC=4.3670E-9
+ NC=1.4264
+ NK=.72845
+ RC=93.301E-3
+ CJE=755.31E-12
+ MJE=.33333
+ CJC=1.1417E-9
+ MJC=.33333
+ TF=1.2802E-9
+ XTF=10
+ VTF=10
+ ITF=1
+ TR=10.000E-9
quite a "spread" as a far as data goes. 😕
OS
The old 1991 model:
.MODEL QPWR PNP (IS=1.63P NF=1 BF=130 VAF=254 IKF=11 ISE=1.34N NE=2
+ BR=4 NR=1 VAR=20 IKR=16.5 RE=12.1M RB=4 RBM=.4 IRB=5.556U RC=4.84M
+ CJE=1.09N VJE=.6 MJE=.3 CJC=708P VJC=.22 MJC=.2 TF=5.33N TR=204N)
+ XTB=1.5 PTF=120 XTF=1 ITF=9.6)
One version I have , I think have a 2001 version (I can't find it)
* created using Parts release 7.1p on 07/01/96 at 10:31
* missing storage time & output admittance
.MODEL Q2SA1302 PNP
+ IS=21.479E-12
+ BF=136.48
+ VAF=100
+ IKF=19.980
+ ISE=21.504E-12
+ NE=1.3784
+ BR=329.48
+ VAR=100
+ IKR=19.980
+ ISC=4.3670E-9
+ NC=1.4264
+ NK=.72845
+ RC=93.301E-3
+ CJE=755.31E-12
+ MJE=.33333
+ CJC=1.1417E-9
+ MJC=.33333
+ TF=1.2802E-9
+ XTF=10
+ VTF=10
+ ITF=1
+ TR=10.000E-9
quite a "spread" as a far as data goes. 😕
OS
Hi All,
Deepest thanks for coming together. All I had to do was leave!
Hi Steve,
I wish you well. I hope your appointments go better than mine! 🙂
Any improvements are worth while, so may you have good news and non-invasive procedures.
For posting your results, my deepest thanks. I think that everyone can have some meaningful discussions now. You may have started some investigations into your design with more possibly going forward. I am hoping you find this worthwhile for you to be involved in.
Regards, Chris
Deepest thanks for coming together. All I had to do was leave!
Hi Steve,
I wish you well. I hope your appointments go better than mine! 🙂
Any improvements are worth while, so may you have good news and non-invasive procedures.
For posting your results, my deepest thanks. I think that everyone can have some meaningful discussions now. You may have started some investigations into your design with more possibly going forward. I am hoping you find this worthwhile for you to be involved in.
Regards, Chris
Hi Joshua,
Your earlier comments were not required as you have posted them frequently here.
There is no need for you to take up sword and shield to defend anyone. This type of posting will generally only prolong "hostilities".
The responses you see from your post are very mild compared to what I have seen in similar situations in the past. Almost polite really.
-Chris
I really think you are reading far too much into things. I know that syn08 never questioned Steve's integrity. It was the design, not the man and that is fair within limits.
Nope. It wasn't that bad. Underdog complex?
Yes. In fact, that was posted in the clear and everyone, including myself, saw and understood this. Nothing wrong with that in my book, as long as measurements eventually do. They did.
Your earlier comments were not required as you have posted them frequently here.
There is no need for you to take up sword and shield to defend anyone. This type of posting will generally only prolong "hostilities".
I guess you don't understand how semiconductors or simulators work. I am a huge proponent for matching certain stages in amplifiers (and other stuff). But this is also accepting that NPN and PNP transistors are different. They are not true compliments to each other. They are designed to be as close as can reasonably be, but you have to accept the basic reality that they are not. Just closer in ratings. Knowing this, have a look at real spice models for these parts. They will be very different, so when you see something that syn08 very correctly pointed out, he is actively assisting others. This is of great importance to know if you plan to use those models.
The responses you see from your post are very mild compared to what I have seen in similar situations in the past. Almost polite really.
-Chris
ostripper said:
OS
It’s the same model that I’m using, but I don’t think its “perfect”
As I have tried to say in another thread I would like to have all the “h” parameters on the data sheet, and thereby be able to include some interesting parameters in the model.
Re: Re: Models
You can go to their website and read, most of their models are compiled from the freely available vendor sites. I know of no one characterising discrete devices in order to sell better models. There are literally millions of dollars of equipment involved for little or no return. You can also read their description of the differences between their program and Berkeley Spice. The differences are in the user interface and I/O options, the math engine is virtually the same.
andy_c said:
You can go to their website and read, most of their models are compiled from the freely available vendor sites. I know of no one characterising discrete devices in order to sell better models. There are literally millions of dollars of equipment involved for little or no return. You can also read their description of the differences between their program and Berkeley Spice. The differences are in the user interface and I/O options, the math engine is virtually the same.
Re: Re: Models
These are the Intusoft models. Don't have the foggiest idea where are they originally coming from (I can ask though) but I've seen these (with slight variations, like skipping the SUBCKT part) all over the Internet.
EDIT: For example http://www.diyaudio.com/forums/showthread.php?postid=174227#post174227
andy_c said:
These are the Intusoft models. Don't have the foggiest idea where are they originally coming from (I can ask though) but I've seen these (with slight variations, like skipping the SUBCKT part) all over the Internet.
EDIT: For example http://www.diyaudio.com/forums/showthread.php?postid=174227#post174227
Okay, sorry - I should have done my homework before posting. I didn't realize that any SPICE vendors were actually making their models available.
None are perfect.. But some are better. After my 2sa1381/ksa1381 "dilemma", I check the actual data
on a model . I have tried the 1991 model , the 97,
and finally andy c.'s mjl1302_x model.
What I found is that the model with less data will
exhibit ideal performance over a much wider range
of circuit conditions (wrong Ic , at clipping , etc) , but the
more advanced one with more data(newer) will only
perform well after the circuit is "fine tuned".
The MJE340 model is a good example of this. It's
model will simulate a working circuit , but when you build it..
So, I now use those devices for shuntsand CCS's.
- Status
- Not open for further replies.