j-FET / Triode Phono front-end

What people are working on at the moment
Andrew
Eternally single
Posts: 4206
Joined: Thu May 24, 2007 2:18 pm

#226

Post by Andrew »

I'll happily have a listen if Mark is short on time, then pass it onto Mark, just reworking my LCR to (hopefully) get the noise down a bit more, so its an opportune moment.

Andrew
Analogue, the lost world that lies between 0 and 1.
JamesD
Old Hand
Posts: 997
Joined: Tue Oct 09, 2007 4:26 pm
Location: North Yorkshire

#227

Post by JamesD »

I'm away on holiday until 21st october so it will be after then before its finished - happy to send it to andrew and then maybe swap to mark at Owston???

So if we go for 100R in and 15 times gain?

J
Andrew
Eternally single
Posts: 4206
Joined: Thu May 24, 2007 2:18 pm

#228

Post by Andrew »

JamesD wrote:I'm away on holiday until 21st october so it will be after then before its finished - happy to send it to andrew and then maybe swap to mark at Owston???

So if we go for 100R in and 15 times gain?

J
That all works for me :)

Andrew
Analogue, the lost world that lies between 0 and 1.
JamesD
Old Hand
Posts: 997
Joined: Tue Oct 09, 2007 4:26 pm
Location: North Yorkshire

#229

Post by JamesD »

Now have a design with component values :D Attached.

For 300uV input its 100R input impedance and 4.5mV out with 1500R output impedance

I'm away until Friday this week walking in the Lake District - been great so far! So I'll knock it up this weekend and get it off to Andrew.

J
Attachments
MC gain block
MC gain block
JFET complimentary MC head gain block v2.png (10.82 KiB) Viewed 9245 times
JamesD
Old Hand
Posts: 997
Joined: Tue Oct 09, 2007 4:26 pm
Location: North Yorkshire

#230

Post by JamesD »

Just found this link on DIYAUDIO for someone doing almost exactly the same for a current output DAC!

http://www.diyaudio.com/forums/digital- ... erter.html

And its been published in Linear Audio - told you its been done before :D

I was aware of NP using it in his ZEN designs but this one is new to me - of course the CEN design uses it as a current input and I'm treating it as a voltage input - two sides of the same coin really...

J
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#231

Post by IslandPink »

Indeed, there it is !
Good effort, anyway . What will you use for the clean 9V supplies ? - batteries ?
"Once you find out ... the Circumstances ; then you can go out"
JamesD
Old Hand
Posts: 997
Joined: Tue Oct 09, 2007 4:26 pm
Location: North Yorkshire

#232

Post by JamesD »

Ideally Lithium PP3 cells. They are 9.6V new and stay the closest of any battery to there rated voltage as they age. The best lithium PP3 have 1200mA/hr capacity and the two FETs draw 6mA Idss (for the ones I choose) at 10V Vds. So thats a nominal 200hrs from one pair of cells - They are £9 each so not cheap but for a fully floating supply with very low noise they take some beating...

On the other hand I have two alkaline PP3s in the part box so I will probably start with them...

J
JamesD
Old Hand
Posts: 997
Joined: Tue Oct 09, 2007 4:26 pm
Location: North Yorkshire

#233

Post by JamesD »

I've been thinking about the folded cascode, both JC Morrison and John Curl have mentioned that this configuration reduces distortion compared to standard cascode and any of the normal connections and so the question remains as to how? JC suggested earlier in this thread that one devices transfer characteristic bends one way and the other devices characteristic curve bends the opposite way due to the use of P and N devices and the way they are used. This makes intuitive sense and is analogous to the way the Aikido operation of two stacked triodes operated as common cathode loaded with common anode (or is it common grid?) works...

But there is one thing about the folded cascode that is really not cascode like. :) A cascode is a voltage in current out device loaded with a current in voltage out device. A folded cascode with a relatively low value resistor in the middle is a voltage in current out device loaded with a voltage in voltage out device...

Except in the case where there is a constant current source in-between the two devices, then it is the same as the normal cascode. Interestingly this was the first use of folded cascodes in integrated circuit design.

So, to me, the folded cascode with resistor is not a true cascode but is a complimentary compound device that has the potential for very low distortion and high gain. We can also build it with normal input impedance or with low input impedance - Serendipity for many applications and particularly, in audio, for mc input stages, ribbon mic input stages and for current output DACS.

Reading John Curls Blowtorch pre-amp revisited thread on DIYAUDIO and, in particular, his release of some of his mc input designs is very interesting and strongly suggest that I should evolve the simple mc gain block design to incorporate folded cascodes. JC uses a hybrid but I can see using FET folded cascodes, cascoded into a pentode Gm amp. My big worry under these conditions would be controlling the gain to prevent nasty clipping under overload conditions but it does promise low noise and low distortion with predominantly low order distortion products... Maybe enough to pass the MJ test for tone!!!

First things first... Build the simple gain block and get it to Andrew, Owston and Mark (in that order) then build the more complex compound amplifier - in the mean time lots of thinking and design work...

Oh, John Curls solution for balanced input is to use two mc input stages together one side of the mc output each and have balanced output...

Ciao

James
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#234

Post by IslandPink »

Bravo James .
I did get some of the other jFets ( 147's ? ) needed for the folded cascode, and have them somewhere if you want to borrow .

ps. I don't even know if distortion (as such) was the problem with the jFet ideas I tried, as Andrew was unable to measure any :!: for some set-ups I was trying, but this should not stop anyone trying the folded cascode, as whatever was making the upper mids a little SS in my cascode, maybe something phase-related, might sound different anyway with the JC folded circuit for all we know .
"Once you find out ... the Circumstances ; then you can go out"
JamesD
Old Hand
Posts: 997
Joined: Tue Oct 09, 2007 4:26 pm
Location: North Yorkshire

#235

Post by JamesD »

Mark,

Thanks for the kind words and offer. I'm OK at the moment - we'll see how I progress!

Attached is my latest design with the folded cascodes - still got to calculate the values.

J

NOTE: This design is wrong and doesn't work as intended!!!! 19OCT2013.
Attachments
folded cascode mc gain block
folded cascode mc gain block
JFET complimentary folded cascode MC head gain block v2.png (9.06 KiB) Viewed 9138 times
Last edited by JamesD on Sat Oct 19, 2013 9:57 am, edited 2 times in total.
dreamth
New User
Posts: 1
Joined: Fri Oct 18, 2013 11:06 am

#236

Post by dreamth »

[quote="James D
Oh here is the bones of the circuit I'm playing with....

James[/quote]
Sir, I saw your folded cascode and I don't understand how it avoids miller effect if reflecting load is on collector.It looks like high frequencys are capacitively shorted to ground.
I have no electronics degree so please excuse me.
JamesD
Old Hand
Posts: 997
Joined: Tue Oct 09, 2007 4:26 pm
Location: North Yorkshire

#237

Post by JamesD »

Hello Dreamth,

Miller effect only applies to amplifiers that have the input signal taken to the grid, base or gate and the output from the anode, collector or drain. As this does not apply to the circuits here (input is to source) there is no miller effect.

The hf roll off the circuit will be the rf region... I haven't calculated it and haven't measured it as it isn't a concern.

regards

James
Andrew
Eternally single
Posts: 4206
Joined: Thu May 24, 2007 2:18 pm

#238

Post by Andrew »

IslandPink wrote:Bravo James .
I did get some of the other jFets ( 147's ? ) needed for the folded cascode, and have them somewhere if you want to borrow .

ps. I don't even know if distortion (as such) was the problem with the jFet ideas I tried, as Andrew was unable to measure any :!: for some set-ups I was trying, but this should not stop anyone trying the folded cascode, as whatever was making the upper mids a little SS in my cascode, maybe something phase-related, might sound different anyway with the JC folded circuit for all we know .
I need to have another go at that, I still think the jfet cascode is a great way to get the heavy lifting done to drive an LCR - still need to get happy with the LCR using the usual a pentode first, doing it on a PCB this time so it should be easy to swap out the gain stages.

Andrew
Analogue, the lost world that lies between 0 and 1.
JamesD
Old Hand
Posts: 997
Joined: Tue Oct 09, 2007 4:26 pm
Location: North Yorkshire

#239

Post by JamesD »

I'm thinking about this compound stage to drive the RIAA EQ. With the right choice of Pentode it could really fly :-)

How much current do you want in the driver? With E55Ls you could have 100mA of drive!!!

J.

NOTE: This design is wrong and doesn't work as intended!!!! 19OCT2013.
Attachments
Folded cascode-pentode mc gain block
Folded cascode-pentode mc gain block
JFET complimentary folded cascode Pentode Gm MC head gain block v1.png (9.68 KiB) Viewed 9075 times
Last edited by JamesD on Sat Oct 19, 2013 9:58 am, edited 1 time in total.
Andrew
Eternally single
Posts: 4206
Joined: Thu May 24, 2007 2:18 pm

#240

Post by Andrew »

JamesD wrote:I'm thinking about this compound stage to drive the RIAA EQ. With the right choice of Pentode it could really fly :-)

How much current do you want in the driver? With E55Ls you could have 100mA of drive!!!

J.
I use 50ma-ish but that's mostly to obtain the max gm from the high gm pentodes.
Analogue, the lost world that lies between 0 and 1.
Post Reply