j-FET / Triode Phono front-end

What people are working on at the moment
User avatar
Joe Roberts
User
Posts: 16
Joined: Fri Jan 20, 2012 10:02 pm

#196

Post by Joe Roberts »

I'll await what Andrew can tell us on the test bench and Joe's impressions of the Silbatone trip
Sadly our listening impressions are inconclusive. It was a very busy trip and jc and Dr. Bae, our wunderkind layout engineer, were only able to finish the prototype preamp in time for us to give it a 30 minute listening at 11PM the night before we left.

Everything, every stage in this preamp is "new" so we cannot be sure what we were listening to. The unit had fantastic dynamics, an incredible 3-D effect that was kind of shocking, but, alas, a tiny bit of electronic glaze to go along with the radical detail. It was very encouraging but a project needing bench time and adjustments. A few days burn in would also help.

It might not be the FETs causing the coloration and jc thinks it isn't. We are using silver LCR chokes, a new kind of silver alloy wire in the 437A output/interstage transformer, and an output stage using a 6900 and a FET in a freaky new configuration. In other words, the whole project is a leap into the unknown at every solder connection.

To my ear, the negative effect we experienced could be RF. The prototype is still an open faced sandwich of PCBs on a metal plate, naked jumper wires and an outboard step up trans.

Then again, one grid stopper might fix the questionable glaze factor. The 437A in the post-EQ gain recovery stage is quite a hotrod, high gain x bandwidth product tube. Maybe more attention needs to be focused on implementation here.

While it was playing, jc and I just looked at each other puzzledly. The overall presentation was so very different from what we are used to that we haven't a clue which element contributes what to the sound.

I hope that the silver EQ chokes aren't the reason for the crazy 3-D because then that would be another stupid expensive part that we all have to buy to be "state of the art."

Here's a pic of the breadboard prototype. On the left is the silver RIAA module with relay switching for various phono EQs, then a pr of big 5k transformers for the 437A, audio board with D3a/437A/6900, and finally a complex power supply with lots of regs and whatnot.

I am sure it will turn out interesting but we don't know enough yet to predict the end result. That why they call it Research and Development... a.k.a. learn as you go.

Surely jc would have other things to say but that is my perspective. He might be flying back to Seoul to work out the details so we can play this box at the Munich show.
Attachments
newpre.jpg
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#197

Post by IslandPink »

Hey Joe ( as JC says ! )
This is great info, even if you aren't quite 'there' yet.
If we can contribute anything to this, as a sort of 'relay' effort, then I'll be really happy .
From my experiments in the last few weeks, I would say that the FET front end brings great grip and bass dynamics to the show, extension into the low bass ( because of the lack of low-end inductance issues ) and helps to give greater 'understanding' of the structure especially with bass guitar and percussion. I can see why Nick was a fan of this when he started building these phono units.

My experience of LCR always ( at ETF2005 and since ) is that it gives superior tone in the upper bass and mids, and a very organic feeling of harmonic and dynamic integration between the bass and mids . This is why I went ahead and bought some of Dave's 1.5k LCR units for 'the future' . The future is getting pretty close now !

My guess is the silver and 80% nickel will just build on the tone and low-level info. I would expect these to help extend the soundstage backwards ( depth ) .

I am frustrated that , so far, we have been unable to find any measurable sonic 'signature' from the FET that accounts for the shortcomings in my set-up . We hope to have a few more tests to help with that in due course.

I have now got some PCB's for DN2540 CCSs . I have in the past used Gary Pimm's Rev.5 and SB CCS's with some success, but not quite successful as anode loads in a phono ...
I will trial the DN2540's against Gary's units as loads for some 6N6P buffers into the LCR units in my new build .
Mark
"Once you find out ... the Circumstances ; then you can go out"
simon
Thermionic Monk Status
Posts: 5643
Joined: Thu May 24, 2007 11:22 am
Location: People's Republic of South Yorkshire

#198

Post by simon »

I must confess that there's much in this thread that I don't understand but it's fascinating nonetheless and I'm very grateful that so many have shared so much. Keep it up!
User avatar
Mike H
Amstrad Tower of Power
Posts: 20189
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#199

Post by Mike H »

It is that ~ I'm largely staying out of it as it's unknown territory for me as well :D
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8985
Joined: Mon May 21, 2007 9:42 pm

#200

Post by Paul Barker »

Mark there was a 9 missing at the start of that phone number. It begins 07989

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

#201

Post by JamesD »

Simon said
JamesD wrote:
so we have, in some ways, a bipolar bisexual able to bend both ways at once to end up looking straight to the norms and sexy kinky to those who look deeper and shoot from the hip - so to speak


I have no idea what that means but it sounds bloody great!
So what did I mean? What I think JC was referring too and what I am finding in my explorations is that the folded cascode allows one to balance the curvature of a transfer curve of a p-channel device against the curvature of the transfer curve of an n-channel device to generate a compound linear transfer curve that allows the cascode to generate very low amounts of distortion - and if one gets it right to be able to minimise the higher order distortion in particular. As JC states, this is very difficult stuff to optimise correctly and needs considerable tweaking to get it to work properly and, no doubt, is even more tricky to keep working as the devices age... oh well that is the joy of diy - we don't mind tweaking and rebuilding if necessary every six months :-)

I've been exploring this through sims and starting to see what JC means about following the trends through the sims - rather than finding exact values for the circuit - after enough playing with the circuit one develops a sort of feel for what happens as you move the operating points of the two devices around... honest!

I've been investigating the folded cascode too and JCs version is, as you would expect, somewhat unconventional :-) It seems that the reason the folded cascode is so named is because the current variation output of the first device is reflected into the second devices input (cathode or source or emitter - a la cascode) off a very, very high output impedance current source hence the 'folded' term, and this allows the use of complimentary devices that in turn allow the folded cascode to operate from a much lower voltage than a normal cascode would require... useful for use inside integrated circuits and that is where this design originates from. What JC has figured out is that you can use this design to balance out the transfer curves for very low distortion performance and relatively high gain - just what we need for an MC input phono preamp!

Putting the feel for the folded cascode through sims and the genesis of the folded cascode together and my previous experience with J-FET cascode input stages and bipolar pnp input stages, I decided to try a little tweak of my own and I have been sim'ing using a couple of SMM2220 pnp transistors into a pentode as a folded cascode with a CCS stage to do the folding... Previously I found that the pnp input mc stage had better tone and bottom end compared to the J-FET cascode input stage so I thought it was worth trying the pnp input... It sims well and I seem to get a rather nice distortion null too so next step is to build it - groan.... no idea when I will get the time to build it but I'll have to try for the next Owsten so we can see how it compares to Nicks and Marks and Andrews designs :-) :-)

Any road up, that is where I am today and we can thank the delays of travelling via train for the free time to write this!

Oh here is the bones of the circuit I'm playing with....

James
Attachments
Folded Cascode - PNP into Pentode.png
Folded Cascode - PNP into Pentode.png (8.23 KiB) Viewed 8686 times
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#202

Post by IslandPink »

Now then James.
Nice work, glad you got some free time in-between the grind.
My problem in motivating myself is still that there is no measureable distortion, so the problems I heard still come into the category of 'that thing where solid-state doesn't work' .
Will your scheme just null the 2nd & 3rd-harmonic issues, or will it address other potential problems , higher-order unmeasureable (?) stuff, or level-dependant distortion, or those other nebulous forms of error ..

:?

MJ
"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

#203

Post by JamesD »

IslandPink asked
Will your scheme just null the 2nd & 3rd-harmonic issues, or will it address other potential problems , higher-order unmeasureable (?) stuff, or level-dependant distortion, or those other nebulous forms of error ..
Well that is the question! Given Joe's account of JC's beastie - it appears that the FET/triode combo nulls 2 and 3rd quite well but leaves the higher order stuff - this is speculation as there is plenty of sand in JCs design - generally not in the amplifying element though except the first stage cascode...

I'm trying to get the higher order stuff to cancel and sort of ignoring the 2nd and 3rd stuff - interestingly its relatively easy to find op points at which 2 and 3rd null out but much harder to effect 7th 9th etc. I'm using a horrible bastardised simulator that runs in Linux to get the spectrum of higher order stuff out... I know JRB does it in LTsim but that is windows and I try to avoid that!

Really I have to build it and try it - could be lots of pots so I can vary op points...

That means I need to get the Quasars mended too or I'm stuck with my Ruark Prologues! Oh - I can use the Alpair 12s once I have the cab made :-)

Bacl to the grind stone

J

[/quote]
Laurence
I am not a user, I am a free man
Posts: 73
Joined: Wed Oct 05, 2011 6:21 pm

#204

Post by Laurence »

Is anybody else looking at this thread and at the same time reading what steve is doing with the U/L cascode and wandering when the penny will drop?
JamesD
Old Hand
Posts: 997
Joined: Tue Oct 09, 2007 4:26 pm
Location: North Yorkshire

#205

Post by JamesD »

Not me!

I've done extensive experiments with triode cascode including UL cascodes and prefer meduim and high Gm pentodes everytime.

Its not a matter of right or wrong, just personal taste but I'll take the pentode everytime except for certain rf frontends...

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

#206

Post by JamesD »

Just had a thought - Laurence might be onto something...

If the Fet/pentode transfer curves curvature doesn't cancel for higher order harmonics then maybe a FET-U/L triode cascode can be tweaked to cancel higher order harmonics? And the compound stage might sound better than the alternatives....

Must be worth a shot...

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

#207

Post by IslandPink »

It would be good if we could actually find some higher-order distortion, because the JFET alone didn't seem to have any !
I wish I'd had a good oscilloscope to hand when I did my experiments, to see if there was any RF I failed to chase-out .
"Once you find out ... the Circumstances ; then you can go out"
User avatar
Nick
Site Admin
Posts: 15750
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#208

Post by Nick »

It would be good if we could actually find some higher-order distortion, because the JFET alone didn't seem to have any
Yep, that is where it all falls down :-)
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
Andrew
Eternally single
Posts: 4206
Joined: Thu May 24, 2007 2:18 pm

#209

Post by Andrew »

The new sound card has potentially a better noise floor, this is just a cable looping back, without the front end. I'll have to see what it's like when I add in the front end box, but I can see more generator harmonics than before, which is encouraging, they were lost in the noise before. THD+N figures aren't bad either. Next step is add in Pete Millet's ATEST box.
Attachments
loopbackx25_small.jpg
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

#210

Post by JamesD »

That looks really interesting! Look forward to the new measurements. :D :D :D
Post Reply