I made some more experiments with the input stage of the Cro Magnon.
I was in search if the J-Fet BJT cascode limits the gain but i found it is clearly the CCS that sets the limit. To test that i made a J-Fet J-Fet cascode that should have a very high output impedance. I am now starting with Frans simple design and when that does not work i will cascode the MosFet. If that does not help i have an idea for a two stage design.
I was in search if the J-Fet BJT cascode limits the gain but i found it is clearly the CCS that sets the limit. To test that i made a J-Fet J-Fet cascode that should have a very high output impedance. I am now starting with Frans simple design and when that does not work i will cascode the MosFet. If that does not help i have an idea for a two stage design.
Joachim,
like I anticipated before gain is limited mostly by the CCS which confirms your experiment.
I am attaching a circuit I have simulated carefully that offers minimum noise and maximum RIAA precision.
It took a while to get proper parameters dialed in.
However I tried to build this one and I was unable to get a proper results.
Maybe it had to do with my built and maybe you are more lucky that I was.
I also notices that changing CCS thus output impedance, strongly affects low frequency response and requires significant adjustments on the RIAA.
This looks like a tricky circuit hard to get it right, but IMHO if we could get things done right could be a pleasent surprise.
Here you go...
Ok, that looks decent. What is the problem ?
DC drift at the output node of the input stage ?
Yes, this is a tricky circuit and to my knowledge it has being never tried ( successfully) to make a one stage single ended RIAA. With tubes i got it working up to a gain of 500. That is a bit meager so i added a mu stage at the output. On the other hand with the Lyra Atlas that has 1mV output it works.
If we can trust Samplers simulation we could get distortion down to -80dB.
After what i have learned this year, simpler can sound better.
That is my main interest, how good it will sound.
DC drift at the output node of the input stage ?
Yes, this is a tricky circuit and to my knowledge it has being never tried ( successfully) to make a one stage single ended RIAA. With tubes i got it working up to a gain of 500. That is a bit meager so i added a mu stage at the output. On the other hand with the Lyra Atlas that has 1mV output it works.
If we can trust Samplers simulation we could get distortion down to -80dB.
After what i have learned this year, simpler can sound better.
That is my main interest, how good it will sound.
Agree but it will also lower gain significantly.
Be aware though that output noise strongly depends on the quality of the CCS.
The arrangement and the values here offer 3db or less noise compared to sampler's arrangement.
That could also be another concern and even though noise is not the most important thing it is still very important!
Be aware though that output noise strongly depends on the quality of the CCS.
The arrangement and the values here offer 3db or less noise compared to sampler's arrangement.
That could also be another concern and even though noise is not the most important thing it is still very important!
It is possble if you count a cascode as single stage as it seems to be here.
Of course an additional follower stage or so is also required. I built it some
years ago in order to compare with similar push pull stages. Gain is sufficient
for a power amp.
yeah Gain is plenty for a Power Amp.
My only concern would be the capability to drive a capacitive load which could cause distortion and osillation and that you have limited headroom.
How did it hold up against other PP topologies?
as, was that really single ended or parallel symmetric ?
I showed one stage folded cascodes a long time ago that had enough gain.
with single ended you have the problem that the high current in the folded cascode gives much too much DC over the transimpedance bass resistor. 50mA over 100kOhm is a lot.
I showed one stage folded cascodes a long time ago that had enough gain.
with single ended you have the problem that the high current in the folded cascode gives much too much DC over the transimpedance bass resistor. 50mA over 100kOhm is a lot.
The load drive capability as far as the follower output is concerned is of no matter here,
I think you talk about the drive capabiltity of the gain stage (driving the capacitive network),
and yes, it is different from that of push-pull stages.
If you call a cascode single ended a folded cascode is single ended also, isn't it?
Since the output of the folded cascode is at or near zero DCV I don't see the point here.
I don't choose a higher current in a folded casc. than I would in a regular one (of course
as seen for a single transistor).
No, the collector current of the pnp transistor in your drawing has to be sinked
to the negative side in a current source (gain is too low without).