TSSA - The Simplest Symmetrical Amplifier

what is the benefit from the BIGBT...??

Hi MiiB ;)

Mhmm I simply love this BIGBT :hbeat:

Easy to drive (low input gate capacitance) output transistor (yes I consider BIGBT as one joint part only), very stabile output transistor as integrity (balanced conditions for NPN/PNP), IRF conduct only one base current for both outputs, so for one base current (and it is the same for both outputs) through IRF driver you get two parallel collector curents for the output (doubled hFE). Linear transconductance dIout/dVg, high speed ... you should try it once. ;)
 
Last edited:
TSSA BIGBT HP update

Update released

Changes:

- VAS cascode - two IR LED per cascode
- "ON" RED LED added - indicates bias-on state

Regards Andrej ;)
 

Attachments

  • TSSA BIGBT HP.jpg
    TSSA BIGBT HP.jpg
    854.1 KB · Views: 1,378
R6, R7 are voltage dumpers to CCS, necessary to use because 2N j-fets have too low Vds/power dissipation specs. Current through R6, R7 is constant as CCS-s are capable of, so ideally no modulation should be present on these resistors. Of course it would be better to use source/sink transistors without voltage dumpers (double cascoded CCS). CCS are not ideal and there's always tendency to have increasing level of distortions according to higher frequencies. Separate measurements of CCS shows their capability to maintain the current to be constant. If you suggest to replace the resistors with improved CCS-s you are invited to do so. ;)

Possible ripple suppression would be to add small parallel caps across R6, R7, just to prevent AC modulation (low level ripple) as voltage drop on them.
 
Last edited:
I couldn't even count it.

Question :rolleyes:
jeez, this is supposed to be the Simplest ? :eek: I think this thread has lost the plot along the way. Pitty, I like really simple.

Answer ;)
Hi Bigun ;)

I can not manage more than two threads simultaneously, so logically one would be full of stuff. :D

From simple to more complex version, all based on the same TSSA base core. It is the same like in school, gradually progressing the knowledge. :cool:

Regards Andrej

I hope you find it reasonable. :)
 
what if you reverse the positions.. so you place the resistors at the rails and the current sources where they belong. then you get both....reduced dissipation and no signal induced current modulation.

Changing the position of the serial part inside a current loop doesn't change the voltage drop on that part. I think CCS has to be connected to rail potential otherwise it would loose its reference. :rolleyes:

Think you should rearrange the driven VAS cascode... make the connection inside the Rail resistors for the input pair.

Can you be more specific or can you please draw it. :)

Thanks ;)
 
MiiB, maybe you think about something like this? :rolleyes:

In this way VAS gain transistor has input-differential current included in its local feedback. That can be helpful in extra reducing the distortions, since it doesn't influence back, 680 ohm resistor can be practically connected to any positive potential.
 

Attachments

  • TSSA VAS.jpg
    TSSA VAS.jpg
    190.2 KB · Views: 1,354
Last edited:
Sorry I was looking wrongly at the cascode.. connection is as it should be....

Since you have constant current you also have constant voltage drop.. so it creates it own voltage reference.. you can tie it to gnd with at good size cap...that way you make something similar to the zener set bridge you have in the SSA but with fewer parts...and lees loss.
 
Last edited:
Vgs from both types can be quite different in absolute values, lets say Vgs=4,2 V at Id=100 mA for IRF610 and Vgs=4,5V for IRF9610 at the same conditions. That is not a problem, offset zero setting equals that, more important is transconductance dI/dV ratio (step increase) curve, this has to be equal (very close) for both types. Also NPN/PNP types of one BIGBT has to have hFE within +/-5%. ;)

Regarding TSSA BIGBT HP PCB; is it possible to sueeze PCB more to eliminate empty spaces - quite large empty areas all over the PCB? PCB will be much smaller, more HF acceptable, shorter tracks, etc. What about VAS cascodes heatsinks? :rolleyes:
 
For BIGBT i have selected 4 MJL2381/1302 couple from a batch of 100u each after mesuring Vbe and HFE. I was just searching what conditions / mesure i do apply to select go mosfet candidat for BIGBT knowing I had good ones for output bjt. See attached document (pair 3/44; 94/19; 93/20; 84/92)
I take a look at NJW3281G/1302G datatsheet. They seems to good candidat for job as NPN/PNP gain matching is giving in 10%, that let good matching previsions between NPN and PNP after HFE/Vbe mesuring session.
For layout I need to rework to see if i can gain space et reducing traks length, but i am in single layer....that not easy. Pearhaps i can do something with TSSA core to bring it back, more closely to BIGBT module.

For VAS cascode i have let room around them, but i have any precise idea of the heatsink power requierement.

Marc
 

Attachments

  • 1108270122151246338648051.jpg
    1108270122151246338648051.jpg
    25.8 KB · Views: 1,125
Last edited:
I build the super simple version of post one. It worked right away. I made some additions and changes. I run it on plus-minus 35V with 500 Ohm resistors instead of trimmers. I runs quite hot that way but i can stilll toutch the cooling profile so it must be around 55° Celsius.
I added a low pass filter at 3MHz -3dB. This amp is very fast and with the filter the square wave is clean. I have cap multipliers in the PSU. That way the amp is dead quiet on my 95dB sensitive speakers.
The sound is a bit lighter then with my power buffer ( see MPP ). By adjusting the gain in the preamp and using a warmer sounding cable i got it going. I have to listen more but the amp is very transparent.