vsat of EF output stage

[phoenix87]

hi Guys ,
what should be the typical vsat for EF type output stage say for a 200W amp running on 4ohms , and it is possible to reduce the vsat by increasing the base current ?
I am using type a configuration in my amp and i generally get a vsat of around 5V for 200W at 4ohms which increases to around 8-9V for 400W at 4ohms , is it possible to reduce the vsat ?

 

[PRR]

Bigger transistors.

 

[Rick PA Stadel]

.. and higher gain driver transistors.

5V and 8 - 9V is a lot of power / waste heat to discard -- plus lost headroom.

Maybe consider VAS improvements, too.

Cheers

 

[BesPav]

is it possible to reduce the vsat ?

Run drivers (or +predrivers) from higher supply rails.

 

[Mark Tillotson]

Higher rails for the predrivers and drivers is good, but also picking better output devices might help (if the existing ones have poor saturation performance)
Which transistors are you current using?

 

[phoenix87]

Ksa3503 /1381 for vas stage , MJE15033/32 for drivers and 2Sc5200/1943 for main output devices

 

[john_ellis]

Data sheet for 2SC5200/A1943 suggests suitable for 100W amp. Gain rolls off at 4A so I'd think that meant into 8 ohms. You would need to use bigger transistors as PRR says or parallel 2 devices for 200W or 4 for 400W etc.

 

[PRR]

> Data sheet for 2SC5200/A1943 suggests suitable for 100W amp.

In the lab, maybe. But I do not know why the makers try to fool us about how many devices we need. All things considered, anything over 50 Watts per pair is reckless, IMHO. And at today's prices, more utterly affordable. With savings due to more efficient heatsink. And higher sales numbers and less unhappy buyers.

Use 4 pair for 200W, have 1A of base current available, and I think your Vsat will be very satisfactory, sub-Volt.

 

[phoenix87]

I am using 6pairs for 400W at 4ohms , how do you generate 1A base current , it would great if you share some insight or some schematic..

 

[sajti]

You can share your schematic as well. It can be useful to check, if there is any design problem.

 

[phoenix87]

This is the base schematic the drivers are MJe15032/33 vas is 3503 /1381/ output transistors are 6pairs of 5200/1943..

 

[sajti]

I woud remove the ripple eaters, and put some small resistors, say 10ohms series. Increase C9 and C10 as much as possible (depending by the footprint size on pcb).
It will save some loss.
For 400W I would use output triplets, but that means redesign for the pcb.

 

[phoenix87]

I would change the PCB , how should i implement the triplets , and anyother changes ?

 

[sajti]

First, remove Q30 and Q31. Replace D23, and D24 with 5-10ohm resistor.
Check if the amplifier works without noise after this change.
Another question: What size of the transformer You have, and how much capacitor in the psu?

 

[phoenix87]

Around 1100Va , 53-0-53V and 40000uf per rail ...

 

[sajti]

That transformer looks OK.
Anyway 400W/4ohm means 56.5Vpeak, so +/-65V PSU would be enough.

 

[phoenix87]

My noload DC voltage is 72VDc and loaded DC is around 67VDc when output voltage is 40vrms -56.6Vpk I am thinking if it possible to reduce the Loaded DC voltage to 61V there by reducing dissipation.

 

[john_ellis]

PRR- Especially for the C5200/A1943. These are only 150W devices, though perhaps the datasheet is - like a lot of marketing BS - "music power" rating. (For those who don't know this means an instantaneous 200W peak power from an unloaded/lightly loaded PSU that then collapses to a lower voltage and can only drive maybe 70-80W continuously).
Transistors like the MJ21193/4 at 250W may just be able to do 100w as a single pair, but the 200W MJL3821/1302 need to be in pairs for 100W, and the higher the output power the more devices are needed as this encroaches more into the second breakdown territory.

 

[phoenix87]

John i am using 6pairs to counteract marketing bs......

 

[john_ellis]

Yep, that stands a chance!
I recall the old RCA 300W design using 10 pairs. Nine each for the output and because the gain was so low another output device as a driver to the 9.