ThermalTrak+TMC amp - Page 14 - diyAudio
Go Back   Home > Forums > Amplifiers > Solid State

Solid State Talk all about solid state amplification.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 19th June 2012, 06:41 AM   #131
diyAudio Member
 
Join Date: Nov 2010
Quote:
Originally Posted by keantoken View Post
Dado, there are no good models for the 2SC2240/A970. I've tried all I could find.
What is the problem with the model(s)?
Have you tried Bob Cordell's book instructions to create a satisfactory model of your own?
This is a transistor I plan to use so I would be happy to help create a usable model.

Best wishes
David

BTW. What is the Toshiba PCT process claimed for these transistors?

Last edited by Dave Zan; 19th June 2012 at 06:47 AM.
  Reply With Quote
Old 19th June 2012, 07:39 AM   #132
diyAudio Member
 
keantoken's Avatar
 
Join Date: Aug 2006
Location: Texas
Blog Entries: 2
Firstly the models don't specify Hfe grade, which is somewhat important. Secondly, Cje is usually just a guessed value, way too low, causing it to be too fast at low currents. In the one model where Cje looked realistic, Tf was not defined, causing high-current Ft to be way too high. The KSC2690/A1220 models had the same problem despite excellent Hfe modeling, so I added Tf myself. So perhaps that one model could be made accurate, but there is no satisfactory basic model for the complement, and I haven't had the desire to make reasonable models for them.

spice models for 2sa970/2sc2240?

I've messed around a lot with BJT models. I have several models I've tweaked or created by hand. This was one such:

New 2SC5200 and 2SA1943 SPICE models

I also made models for the 2SC4883A/A1859A and 2N5769/5771. Here are the models I'm using for the 1845/992:

Code:
.MODEL KSC1845_k npn IS=1.075431E-13 BF=600.7 NF=1 BR=13.565 NR=1 ISE=1.98107E-13 NE=2 ISC=1.8378E-11 NC=1.5 VAF=82.803 VAR=20.6691 IKF=0.596 IKR=0.0190546 RB=157 RBM=12.092 IRB=1.258925E-6 RE=1.5 RC=0 CJE=2.057447E-11 VJE=0.7300286 MJE=0.3619943 FC=0.5 CJC=4.525739E-12 VJC=0.5 MJC=0.3659045 XTB=1.7281 EG=1.1809 XTI=3 TF=3.14E-11 Vceo=120 Icrating=50m mfg=Fairchild 
.MODEL KSA992 pnp ( IS=5.7544E-14 BF=348.1 NF=1 BR=3.62 NR=0.95 ISE=5.7544E-15 NE=1.5 ISC=1.8378E-14 NC=1 VAF=144 VAR=16.68 IKF=0.298 IKR=0.0525 RB=140 RBM=16.084 IRB=1.4125E-3 RE=0.38 RC=0.47 CJE=2.3093E-11 VJE=0.855 MJE=0.4104 FC=0.5 CJC=8.9251E-12 VJC=0.5 MJC=0.3497 CJS=0 VJS=0.8 MJS=0.33 XTB=1.2849 EG=1.1603 XTI=3 XCJC=0.3062 Vceo=120 Icrating=50m mfg=Fairchild)
They are good enough for me so far, though I don't use them much. I had to lower RC in the 1845 because the original model had the value way too high, causing bad behavior in current mirrors and CCS.

I don't know what the PCT process is. I don't know anything about processes or fabrication.

My questions is, if everyone now is able to just make their own models whenever they like, why hasn't anyone posted their custom models but me? Am I the only one who wants to share?

Granted, I tweak my models by hand. I never learned how to use spreadsheets. I know what each parameter does, I've done this a lot. It may not have the nth degree of accuracy, but I think most of my models have quorum, and transistors are not all the same anyways.
  Reply With Quote
Old 19th June 2012, 07:59 AM   #133
diyAudio Member
 
keantoken's Avatar
 
Join Date: Aug 2006
Location: Texas
Blog Entries: 2
I forgot to mention, I may have already created suitable models for the C2240/A970 if the datasheets included an Ft vs. Ic graph. This is necessary to find the right value of Cje in the case it is not included in the datasheet.

If a a few or more people were willing to work on a library of good transistor models to share with everyone, I would be happy to pitch in. Actually just today I have a good start on that. I have considered taking model commissions in return for money or parts through the mail; right now I'm trying to find the best seller of non-counterfeit 2SC3284/A1303 because I want to build a nice amp not unlike Dado's (although quite different).

Here is an example of the buffer bootstrap thingy. R37 needs to be chosen so Q15 has a comfortable Vce. Notice the drive current through Cbe returns to the buffer collector in a very short loop. Otherwise it goes all the way around the entire perimeter of the amp through several decoupling caps, and who knows what happens in between. Maybe it is negligible, maybe not.
Attached Images
File Type: png AudioCube_BetaEnhancerBootstrap.png (80.9 KB, 301 views)

Last edited by keantoken; 19th June 2012 at 08:03 AM.
  Reply With Quote
Old 19th June 2012, 08:47 AM   #134
dadod is offline dadod  Croatia
diyAudio Member
 
dadod's Avatar
 
Join Date: Apr 2006
Location: Zagreb
Quote:
Originally Posted by keantoken View Post
I forgot to mention, I may have already created suitable models for the C2240/A970 if the datasheets included an Ft vs. Ic graph. This is necessary to find the right value of Cje in the case it is not included in the datasheet.

If a a few or more people were willing to work on a library of good transistor models to share with everyone, I would be happy to pitch in. Actually just today I have a good start on that. I have considered taking model commissions in return for money or parts through the mail; right now I'm trying to find the best seller of non-counterfeit 2SC3284/A1303 because I want to build a nice amp not unlike Dado's (although quite different).

Here is an example of the buffer bootstrap thingy. R37 needs to be chosen so Q15 has a comfortable Vce. Notice the drive current through Cbe returns to the buffer collector in a very short loop. Otherwise it goes all the way around the entire perimeter of the amp through several decoupling caps, and who knows what happens in between. Maybe it is negligible, maybe not.
Thanks keantoken, I will simulate it to see what I could get.
Could you show your amp schematic, I am very interested to see it, even if I do not intend to build new amp in near future.
dado
  Reply With Quote
Old 19th June 2012, 09:03 AM   #135
diyAudio Member
 
Join Date: Nov 2010
Quote:
Originally Posted by keantoken View Post
Firstly the models don't specify Hfe grade, which is somewhat important.
OK. Easily done. Just two models with different BF

Quote:
Secondly, Cje is usually just a guessed value, way too low, causing it to be too fast at low currents. In the one model where Cje looked realistic, Tf was not defined, causing high-current Ft to be way too high.
OK. Put plausible value of CJE and TF from datasheet Ft

Yes, some of them don't inspire confidence.


Quote:
I've messed around a lot with BJT models. I have several models I've tweaked or created by hand...
My questions is, if everyone now is able to just make their own models whenever they like, why hasn't anyone posted their custom models but me? Am I the only one who wants to share?
Well, Andy and Bob share their models. And other people try to share the factory models they have found - even those that don't inspire confidence.

Quote:
Granted, I tweak my models by hand. I never learned how to use spreadsheets. I know what each parameter does, I've done this a lot. It may not have the nth degree of accuracy, but I think most of my models have quorum, and transistors are not all the same anyways.
Andy's site has an impressive amount of work on the use of spreadsheets to try to best fit the data. Not sure how much better than a hand fit in practise.

Quote:
Originally Posted by keantoken View Post
I forgot to mention, I may have already created suitable models for the C2240/A970 if the datasheets included an Ft vs. Ic graph. This is necessary to find the right value of Cje in the case it is not included in the datasheet.
I don't have the equipment to measure Ft as a function of collector current so it will have to be estimated.

Quote:
I want to build a nice amp not unlike Dado's (although quite different).
I want to build a nice amp like Dado's but quite different too
  Reply With Quote
Old 19th June 2012, 09:16 AM   #136
dadod is offline dadod  Croatia
diyAudio Member
 
dadod's Avatar
 
Join Date: Apr 2006
Location: Zagreb
Quote:
Originally Posted by Dave Zan View Post

I want to build a nice amp like Dado's but quite different too
Dave, please show your schematic too.
dado
  Reply With Quote
Old 19th June 2012, 09:41 AM   #137
diyAudio Member
 
keantoken's Avatar
 
Join Date: Aug 2006
Location: Texas
Blog Entries: 2
I have not changed your schematic any. As your amp inspired me to mess with EF3 topology, instead I went ahead and added a beta enhancer and EF3 onto my current so-far-so-secret power amp design. It is this I plan to build using the remains of a broken HK AVR45 amp and the 2SC3284 outputs. I would really appreciate it if someone knows of a good source of 2SC3284Y. I've found several options for lower Hfe grades, few of which are genuine for sure. After I've explored my design in more detail I think I will share it on the forum, but that is some time from now.

Dave,
I haven't messed with the 2SC2240/A970 models because I haven't felt the need to use them. A friend has just mailed me a package of the BL grade though, so maybe I should. One thing I don't know much about is noise. I would not know how to model noise correctly, and that is most of the spec of this pair. I can get Vbe, Hfe and Ft curves set up, and with good references I might be able to quickly learn how to do noise, but you may have to do that part.

Cje could be directly measured, if you have something that can measure around 25pF with a test signal as small as 100mV. Then there will be no guessing Ft. I don't have any kind of measuring bridge though I seem to need one. Maybe now that I think about it I really could measure this...

I think the closest approximation will be to establish Cje and Cjc and then adjust Tf to get the Ft to match the datasheet. I could be missing something if one of the h parameters has to do with Ft, that's another thing I haven't studied.
  Reply With Quote
Old 19th June 2012, 10:05 AM   #138
diyAudio Member
 
Join Date: Nov 2010
Quote:
Originally Posted by dadod View Post
Dave, please show your schematic too.
dado
Thank you but I have only started to define. No schematic yet. But I want to build at about the level of complexity that you and Andrew (Bonsai) have achieved. Not as complicated as Edmond and Ovidu's masterpiece but better performance than the typical Blameless clone.
So far it looks like.
Complementary symmetry front end - low noise is important for compression drivers.
A two transistor "VAS" -
A triple BJT ThermalTrak OPS - probably as CFP + EF.
TMC/TPC and Input Inclusive compensation.
I would like to use current mirrors but the problems with that + Symmetrical IPS are well known.
So I currently research the options
1. Bob Cordell's resistor load - Reasonably simple
2. Common Mode Control Loop (CMCL) - can this be done reasonably simply?
3. Edmond Stuart's SuperTIS - Very clever but untested, so potential development problems.
There is a lot of old discussions on DIYAudio to read about this and I don't want to write until I understand it - so I don't waste anyone's time. But when I have a circuit I will certainly show it for criticism and review.
  Reply With Quote
Old 19th June 2012, 10:24 AM   #139
dadod is offline dadod  Croatia
diyAudio Member
 
dadod's Avatar
 
Join Date: Apr 2006
Location: Zagreb
Quote:
Originally Posted by Dave Zan View Post
Thank you but I have only started to define. No schematic yet. But I want to build at about the level of complexity that you and Andrew (Bonsai) have achieved. Not as complicated as Edmond and Ovidu's masterpiece but better performance than the typical Blameless clone.
So far it looks like.
Complementary symmetry front end - low noise is important for compression drivers.
A two transistor "VAS" -
A triple BJT ThermalTrak OPS - probably as CFP + EF.
TMC/TPC and Input Inclusive compensation.
I would like to use current mirrors but the problems with that + Symmetrical IPS are well known.
So I currently research the options
1. Bob Cordell's resistor load - Reasonably simple
2. Common Mode Control Loop (CMCL) - can this be done reasonably simply?
3. Edmond Stuart's SuperTIS - Very clever but untested, so potential development problems.
There is a lot of old discussions on DIYAudio to read about this and I don't want to write until I understand it - so I don't waste anyone's time. But when I have a circuit I will certainly show it for criticism and review.
You have put almost all what was discussed here in this forum. Do not over engineered it, keep it "as simple as possible, but not simpler".
I am impresed with Stuart's amps and I learned a lot from him, but one thing keep me from implementation, to complicated PCB layout and I don't think I am qualified to develop good enough one. Still with simpler design you can get very close to his level of distortion, simulated only, but Stuart's amps were simulated only too.
Complementary symmetry front end is not important for me, I am more to D. Self reasoning.
What is Bob Cordell's resistor load, I bought his book but don't remember that?
dado
  Reply With Quote
Old 19th June 2012, 10:27 AM   #140
diyAudio Member
 
Join Date: Nov 2010
Quote:
Originally Posted by keantoken View Post
One thing I don't know much about is noise. I would not know how to model noise correctly, and that is most of the spec of this pair. I can get Vbe, Hfe and Ft curves set up, and with good references I might be able to quickly learn how to do noise, but you may have to do that part.
Yes, Bob does not discuss noise models much and the local library had their SPICE/Semiconductor texts on loan when I last checked. As I understand it Rbb is a primary factor for white noise so RB is important. Theoretically this won't conflict with the requirement to use RB to adjust Vbe at different currents.
I will check how 1/f noise is modelled in SPICE, if at all. I don't recall I have seen it discussed.

Quote:
Cje could be directly measured, if you have something that can measure around 25pF with a test signal as small as 100mV. Then there will be no guessing Ft.
If you measure Cje then Ft is still an independent variable. That's why there's CJE and TF of course.

Quote:
I think the closest approximation will be to establish Cje and Cjc and then adjust Tf to get the Ft to match the datasheet. I could be missing something if one of the h parameters has to do with Ft, that's another thing I haven't studied.
That sounds the way to me too.

Best wishes
David
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
MEMDIST POWER AMP + JC + TMC - help to set up padamiecki Solid State 8 10th January 2014 09:38 AM
Compact Sized LME49810/11 +ThermalTrak Amp panson_hk Chip Amps 302 25th September 2013 10:23 AM
New 200 watt Kit amp features ultra low distortion and Thermaltrak devices snoopy Solid State 56 3rd May 2010 08:42 AM
Mini-Sized 49811 ThermalTrak Amp panson_hk Chip Amps 3 1st April 2010 08:06 AM
What to do with TMC tube radio gear? diamondsouled Tubes / Valves 4 31st March 2009 01:30 PM


New To Site? Need Help?

All times are GMT. The time now is 02:38 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2