Class A Design in progress

[m_buzzi]

Hi all,

I'm designing my own 30W class A amplifier or at least trying to do it....

At the moment I decided to use Class A CFP as reported on Douglas Self book.

At first I tought to use a current mirror topology and a cascode buffered vas. I tried to implement it but my need of low closed loop gain leads me to have an amplifier with more than 100dB of feedback. This cause me to great problems in compensation and looking for stability.

Now I decided, at opposite, that I will try to have a very linear open loop amplifier and then apply feedback. At the moment I'm using a differential VAS and in this way I can get a closed loop thd20 of about 3e-3 prior to clipping and this quite satisfies me. Do you have some advices on differential VAS since its the first time I use it?

Thanks,
Michele

[Wavebourn]

If it is already very linear before applying GNFB why do you need so much gain?

[m_buzzi]

Hi,

The second version I posted has 70dB open loop gain as opposed to the first version, not reported here that has 130dB open loop gain so I have reduced OLgain of a great amount. I was only looking for an advice about differential vas or better topologies if anyone knows.

Thanks,
Michele

[AndrewT]

Hi,
the cascode diodes are upside down. They pass in the forward direction to generate Vf.
What current does the j503 pass?

[m_buzzi]

Hi,

Unfortunately my own version of spice does not have differences in symbol for normal diodes and zener ones. The diodes you see are zener so they're inverse biased to generate Vz. J503 is a CR diode for about 0.55-0.6mA.

Thanks,
Michele

[AndrewT]

Hi,
you might be better using red LED or two for the diode cascode voltage, rather than a Zener.

The CCS to the ltp is sourcing about 6mA. Take off the 0.6mA of the Zener and 2.7mA is passing each collector load.
Is ~4V across these loads what you intended?

[m_buzzi]

Hi,

Thanks, I will try to use LED instead of zeners. I planned to leave 4 volts over the load resistor since the voltage swing there is not great and also because in origin I planned to use a cascode CCS also under the differential VAS. Is it worth?

Michele

[AndrewT]

Originally posted by m_buzzi
I planned to leave 4 volts over the load resistor

have you checked the currents in the VAS?

[m_buzzi]

Hi,

4.962mA in the side connected to output stage opposed to 4.815mA in the other side. This seems to me quite reasonable. Isn't it? Note that the output stage has a predriver and so current requirements to VAS are really low.

Michele

[AndrewT]

I think VAS currents are over 6mA.
But even 5mA is a lot of dissipation for 550/560 To92 devices.