2stageEF high performance class AB power amp / 200W8R / 400W4R

Member
Joined 2011
Paid Member
That is bad news. There is hard to find any drivers which are as good as 5171/1930...

Sajti

Other pre-drivers are only needed if the amplifier should drive 4 Ohm loads and to be able to use the current pcb board. Attached sim shows SOA using possible candidates MJE15032/MJE15033 driving 3 ohm load with max power and extreme clipping.

I have a big number of C5171/A1930 so the next pcb version may contain 4 parallel working pre-drivers. Need some time to simulate and build.
 

Attachments

  • soa_predriver_mje15032.png
    soa_predriver_mje15032.png
    130.4 KB · Views: 1,503
Member
Joined 2011
Paid Member
VAS overcurrent limiter may add distortion

The VAS protection is there to protect the VAS. There is no correlation between an output spike and the R22a value.are you sure?
what IC current is passing the protection transistor when 56mA of VAS current is flowing.
How low does the VAS current need to go to turn off the protection transistor.
I think very many builders get this wrong and then blame the protection for sounding wrong, when It is possibly early triggering that is affecting the sound.

max IC of Q37 is 4 mA when VAS protection is active...

Now I unterstand!
My findings and some pictures ...

  • switch off voltage on base of Q37 is 535 mV.
  • normal working mode 3R load near clipping reaches 580mV
so you are right. If we drive the amp near its limits the VAS current limiter may unnecessarily add more distortion.
Next amp revision will of course get a R22a/R22b divider.
Thank you for reporting this!

BR, Toni
 

Attachments

  • vas_overload_1.png
    vas_overload_1.png
    175.6 KB · Views: 1,405
  • vas_overload_2.png
    vas_overload_2.png
    176.1 KB · Views: 1,321
Member
Joined 2011
Paid Member
15032/33 is strong enough to drive the output stage. The penalty is, that they are slow, have large internal capacitances.
Did You try 2SC4793/2SA1837? They look not as strong as the 5171/1930, but check the much better SOA!

Sajti

15032/15033: first simulation tests are showing a good performance. Need to buy a bunch of them for real life testing. Maybe we have good news and of next week...

2SC4793/2SA1837 are already in my shopping cart. Need SPICE models...
 
Member
Joined 2011
Paid Member
No suckout capacitor help?

Sajti

current simulation: dual pre-drivers MJE15032 / MJE15033

with 150R and 0.47µ suckout cap: distortion 0.002%@200W8R
...rebiasing...
with 33R and 0.47µ suckout cap: distortion 0.002%@200W8R

nearly same. Hmm ... must compare with previous drivers.
Perhaps a biasing mistake during stepped simulation.
A very good hint. (Re)doing SOA checks with 150R ...

BR, Toni
 
Last edited:
15032/15033: first simulation tests are showing a good performance. Need to buy a bunch of them for real life testing. Maybe we have good news and of next week...

2SC4793/2SA1837 are already in my shopping cart. Need SPICE models...

No I'm working on cfp input buffer. Triple EF, with 2sc3423/2sa1360 predriver, and 2 pairs of NJW0281/0302 output. I tried some different drivers, and I got the following result:

2SC5171/2SA1930: VAS load capacitor:330pF to get stable working
2SC4793/2SA1837: 680pF
MJE15032/33: 1nF!!!!

The last means that 10kHz square wave results saturation in the VAS stage!

Sajti
 
2SC4793/2SA1837 are already in my shopping cart. Need SPICE models...

Not sure how good they are but here you go. Can't remember where I found them prob somewhere on this site.

.MODEL 2SC4793 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 2SA1837 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
 
Here is the standard.bjt file I use. It contains the best models I could find for most of the transistors in it. I made most of my effort removing the models which were found inadequate, so it is much less likely to contain duds. Some models I improved or created myself. All of the models at the top are the ones I use most often. The ones with suffixes I have almost certainly determined to be reasonably close to the real device. Below those are a collection of models I pasted from another model collection I found which generally seem trustworthy, but I haven't tested them all; I included them because there are common BJTs missing from the top group. I think you will find most of what you need in here. If not, there may not be any good models for them.

If you use this for your standard.bjt, be sure to back it up and restore it after any LTSpice update, since LTSpice overwrites all of the models, sometimes with inferior ones.
 

Attachments

  • standard.txt
    182.4 KB · Views: 385
Thanks for the file!
There are some different models for bjts. I have about 2 different for the 5200/1943, and 2 different for 4793/1837, same for mjl3281/1302. As I'm not able to find which is the better, I usually simulate my amplifier with both models. :D This looks foolprof solution for me.
But especially the models of NJW0xxx, 5171/1930, 3423/1360, 2240/970,bc550/560 looks OK, as my current amlifier works fine, and the measured results are quite consistent with the simulation.

Sajti
 
Member
Joined 2011
Paid Member
Here is the standard.bjt file I use. It contains the best models I could find for most of the transistors in it. I made most of my effort removing the models which were found inadequate, so it is much less likely to contain duds. Some models I improved or created myself. All of the models at the top are the ones I use most often. The ones with suffixes I have almost certainly determined to be reasonably close to the real device. Below those are a collection of models I pasted from another model collection I found which generally seem trustworthy, but I haven't tested them all; I included them because there are common BJTs missing from the top group. I think you will find most of what you need in here. If not, there may not be any good models for them.

If you use this for your standard.bjt, be sure to back it up and restore it after any LTSpice update, since LTSpice overwrites all of the models, sometimes with inferior ones.

Dear keantoken,

thank you for this collection of spice models!

BR, Toni
 
Member
Joined 2011
Paid Member
SOA test pre-drivers 2SC5171/2SA1930

Attached is a SOA simulation of the pre-drivers.
Simulation data: 3R load, input voltage 2V and 3V (~ max power 1%THD; strong clipping 15% THD). Frequency stepped 20Hz, 1kHz, 20kHz, 100kHz to see frequency dependend interactions with suckout capacitor.

It is now clarified that a missing re-biasing during stepping of R17 from 150R to 33R showed erroneously better distortion results.

Now I think the amplifier is safe providing power to a 4R load on using use the 2SC5171/2SA1930 pre-drivers with R17=150R. :)

If adequately cooled there should be enough SOA headroom. NOTE: output bjt's should have hfe minimum about 80 as this was the model setting!

To have more SOA safety margin I will test the alternatives as well.
So don't forget to visit this thread the next days. ;)

BR, Toni
 

Attachments

  • soa_predriver_2sc5171.png
    soa_predriver_2sc5171.png
    178.2 KB · Views: 1,490
  • soa_output_driver_2sc5200.png
    soa_output_driver_2sc5200.png
    169.9 KB · Views: 1,443
  • 2stageEF_class_AB_power_amplifier.asc
    37.2 KB · Views: 177