About VAS and drivers transistors

Status
Not open for further replies.
Jump to Latest
#61
Hello

Before buying any KSC3503 transistors for Vas, I have a question.

That transistor do have a Ic of 100 ma DC and 7 watt dissipation, is it enough to use it in the Vas for any amp up to 100 watt ?

I would not use it for the driver.

Thank

Bye

Gaetan
 
#63
> That transistor do have a Ic of 100 ma DC and 7 watt dissipation, is it enough to use it in the Vas for any amp up to 100 watt ?

It depends on the topology and VAS bias current, but it should be OK up to
about 20 mA bias current on +50v/-50v rails. You have to derate the Pd for temperatures higher than 25c and use a heatsink, but otherwise it's generally fine - however, you'll probably get better hFE, fT and Cob if you settle for Vceo around 120..200V, rather than the 300V that the KCS3503 is rated for.

> The Sanyo 2SA1477/2SC3787 look pretty good...
>Cob = 3.0 pF
>IC (MAX) = 140 mA

For a VAS, their siblings 2sa1478/2sc3788 are better. Similar,
but with Cob = 2.6/1.7 pF, Vceo=200v, Ic=100 mA, and Ic-Vce
curves that look like a Vernier caliper.
 
#66
Noise is not usually specified for medium-power devices like these that are used for CRT video drivers, audio VAS, predriver, driver, etc. These aren't exactly small-signal devices that are used at the LTP or front-end for low-level signal amplification, which is where noise is likely to be an issue.

For LTP or general-purpose low-noise pre-amplification, there are a very large number of Japanese small-signal transistors that have an outstanding combination of specifications. Presently, my favourites (for no particular reason other than local availability and price) are:

Toshiba 2sc3112 (fast, extremely low noise, super beta, low Cob)
Toshiba 2sa970/2sc2240
NEC 2sa992/2sc1845
 
#68
linuxguru said:
Noise is not usually specified for medium-power devices like these that are used for CRT video drivers, audio VAS, predriver, driver, etc. These aren't exactly small-signal devices that are used at the LTP or front-end for low-level signal amplification, which is where noise is likely to be an issue.

For LTP or general-purpose low-noise pre-amplification, there are a very large number of Japanese small-signal transistors that have an outstanding combination of specifications. Presently, my favourites (for no particular reason other than local availability and price) are:

Toshiba 2sc3112 (fast, extremely low noise, super beta, low Cob)
Toshiba 2sa970/2sc2240
NEC 2sa992/2sc1845
Click to expand...
The small-signal transistors like the 2SA970/2SC2240 get their dissipation limits exceeded if I use them in the I/V posted above with a DAC like the AD1955 which puts out a large current. So I need transistors that can handle more power but are still very linear, have similarly high fT, and are low noise enough for an I/V in a 24-bit DAC where I'm aiming for -130 dB THD @ 20 kHz. So, what transistors do I pick?
 
#69
> which shop in Bangalore do you get these?

Places on S.P. Road to try (I don't remember exactly where I bought
each of these, but these are my sources in decreasing order of
preference):

Chetan Electronics
Darshan Electronics
Swastik Electronics (?)
Om Electronics
Bangalore Electronics
Vishal Electronics

From memory, these are available from about Rs.5 to 7 each.

>So I need transistors that can handle more power but are still very linear, have similarly high fT, and are low noise enough for an I/V in a 24-bit DAC where I'm aiming for -130 dB THD @ 20 kHz. So, what transistors do I pick?

I don't know about the noise specs, but if hfe clustered at 60 is sufficient for your application, the Fairchild KSC3953 with an fT of 400 MHz and Ic=0.2A might work. It's linear like vernier calipers and cheap as chips.

Edit: added Darshan and changed order of preference.
 
#72
Hello

I have decide to ad a 11 pf phase lead capacitor to an amp that a friend ask me to modified, I was think to put it between the base of Q2 and the colector of Q6.

Just to be sure, is it the best place for a phase lead capacitor ?

Here is the amp input schematic.

Thank

Bye

Gaetan
 

Attachments

  • amp input schematic-sm.jpg
    amp input schematic-sm.jpg
    69.9 KB · Views: 1,493
#76
AndrewT said:
cap is in the correct place. Allow for a series resistor in case you need to adjust. Or alternatively a parallel R+C to trim the stability/margins.

C8 & 10 might need to be retained but at much lower values.
Click to expand...


Hello Andrew

I've found time to do those mods on the amp and a 15pf cap at that place do made the amp to sound better. The amp are a Sima 3050.

Thank

Bye

Gaetan
 
#77
gaetan8888 said:
Hello

My friend want me to modify this amp because the sound of it are a bit harsh and the sound stage not very good.

Thank

Bye

Gaetan
Click to expand...

R22 and R28 are center tapped to the output.
This sounds like a bad idea to me, in case the amp runs at rather low bias currents, because this kind of configuration tends to push the (pre-)drivers into Class B at rather low output currents.


Regarding VAS-Transistors in general:

Why should one use a VAS, which is not cascoded ?

In case it is cascoded the feedback capacitance of the VAS transistor turns from a feedback capacitance (often already the dominating pole) into a simple capacitive load for the VAS, just like the (nonlinear) capacitance of the (pre) drivers.

In case a cascode is used for VAS, why then should be a relaxed specification for the predriver's Cob ?? Their feedback capacitances are then a load very comparable to the one of VAS transistor's capacitance !

Conclusio:
For an optimal VAS, running in cascode mode, the predrivers should be the very same type of transistor as the VAS transistors !

On the other hand side:
Is there any proof, that the Cob of conventional VAS transistors does any harm to overall distortion ? I would say, that the load of the VAS always dominates, as long you operate in Class B or Class AB !
 
Status
Not open for further replies.
Top Bottom