DIYHiFi.org

For the sake of audio
It is currently Tue Sep 26, 2017 12:46 pm

All times are UTC




Post new topic Reply to topic  [ 88 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6  Next
Author Message
PostPosted: Tue Aug 29, 2017 8:04 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1130
Hi,

Mikeg wrote:
I know someone with a good clock. It's amazing.


My German Clockmaker friend is restoring a classic "Regulator" Clock for me for my new place.

I suspect Jocko would find this Clock rather noisy.

:yahoo: :rotfl: :yahoo: :rotfl: :yahoo: :rotfl:

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Tue Aug 29, 2017 1:35 pm 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1130
Hi,

hanski wrote:
Quite close to T's circuit?


Yes.

Meanwhile I remembered a preamp design I came across yonks ago on the net and lo and behold, it's still there.

Simple. Solid State. Open Loop. Passive RIAA EQ (which works here because of the high voltages used).

It can give a "template" to substitute improved stages in each circuit block...

http://www.metaxas.com/pages/technology/diy.html

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Tue Aug 29, 2017 8:50 pm 
Offline
Cow

Joined: Fri May 30, 2014 6:24 am
Posts: 132
Kuei Yang Wang wrote:
Meanwhile I remembered a preamp design I came across yonks ago on the net and lo and behold, it's still there.

Simple. Solid State. Open Loop. Passive RIAA EQ (which works here because of the high voltages used).

It can give a "template" to substitute improved stages in each circuit block...

http://www.metaxas.com/pages/technology/diy.html

Ciao T

Thanks T!
Any chance on enlighting the basic idea of the article, the use of differential gain stage as a normal single ended gain stage?
I mean the other halve of differential stage is "wasted" there.

From a circuit point of view, the differential stage biasing is different vs. single ended stage biasing. (Casoding and CFP can be used with both circuits.)

But am I missing some main points when considering about subjective sonical quality?

-Hanski


Top
 Profile  
Reply with quote  
PostPosted: Wed Aug 30, 2017 3:39 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1130
Hi,

hanski wrote:
Any chance on enlighting the basic idea of the article, the use of differential gain stage as a normal single ended gain stage?


Well, there are a few things here.

First, a differential stage may use enhancement devices and can still be biased to have the input at ground level.

Second, transconductance (and thus gain) and biasing are no longer intrinsically linked as they are with SE stages, unless we add CCS tails to the negative supply and an RC combo with big value "C" to set the gain.

Third, a differential stage has almost no net signal current change on the PSU line (as long as it stays in Class A only any error current from the non-infinite tail impedance flows), this can be quite significant, especially in resistively loaded designs not using a folded cascode to reference the load to ground. In fact, I'd probably give the two Phono stage stages RC Decoupling from the PSU, even so.

And yes, cascoding, compound feedback pairs etc. remain applicable.

BTW, for a tone control, there is a trick.

If you use a correctly "tapped" potentiometer (very rare now) or a tap on a stepped attenuator, you can apply a tone control in parallel with the volume control. This can be found (for example) in the TA-S7 from Sony (service manuals at the usual places).

If (for arguments sake) we use 12-position switches for for the tone control and a 24 or 48 position switch for volume, we can make a passive Tone/Volume combo where the "tone" section is completely out of circuit when defeated, serious "hai ente" tone controls.

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Wed Aug 30, 2017 8:13 am 
Offline
Cow

Joined: Fri May 30, 2014 6:24 am
Posts: 132
Thanks T!
Quite convincing sales talk, or propaganda. :grin:
I think I'll buy it due to biasing reasons. (I was thinking of using a servo bias for the SE gain stage and AC CCS for active load in the phono amp.)
Differential stage biasing seems easier.

Sony Ta-S7 seems to have tone controls in parallel with the mentioned special volume pot.
When defeated, the tone stack is replaced with a fixed resistor. (If I got it correctly.)
Attachment:
Tones.PNG


You do not have the required permissions to view the files attached to this post.


Top
 Profile  
Reply with quote  
PostPosted: Thu Aug 31, 2017 5:35 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1130
Hi,

hanski wrote:
Quite convincing sales talk, or propaganda. :grin:


You did not ask about drawbacks (other than increased complexity).

Personally, I might do the Phono in SE (giving us something like the l'Pacific, Pass DIY Pearl etc.), but keep the line stage.

One trick is to make the first stage SE, the second P-Channel/PNP (perhaps with J-FET buffer) and direct couple. It produces a ground referenced output and removes the frontend PSU from the input loop of the second stage.

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Thu Aug 31, 2017 3:28 pm 
Offline
Cow

Joined: Fri May 30, 2014 6:24 am
Posts: 132
Wouldn't it be beautiful. All SE stages, DC coupled, inputs/outputs ground referenced, and all stage connections to PSUs with constant current sources.
Have to change the shirt, too much drooling. :cool:

Here redrawn the previous SE phono stage:
Attachment:
cascode CCS bias.png

Well, it almost fulfills the above requirements. Except that we have a large cap in the CFP cell source leg! :eek:
I hear someone has told to avoid such things..

It's this solution or the servo at Jfet gate (with input cap).
If possible, I would like to choose a solution without a cap. Then I would need to change my pants also. ;)


You do not have the required permissions to view the files attached to this post.


Top
 Profile  
Reply with quote  
PostPosted: Tue Sep 05, 2017 4:42 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1130
Hi,

hanski wrote:
Wouldn't it be beautiful. All SE stages, DC coupled, inputs/outputs ground referenced, and all stage connections to PSUs with constant current sources.
Have to change the shirt, too much drooling. :cool:

Here redrawn the previous SE phono stage:
Attachment:
cascode CCS bias.png

Well, it almost fulfills the above requirements. Except that we have a large cap in the CFP cell source leg! :eek:
I hear someone has told to avoid such things..


I would probably go two-stage. Asking one stage to do too much is always a bad idea.

Keep the RIAA, but IMNSHO using a folded cascode, without too much messing around.

Use parallel J-Fets for MC, low noise, say 8pcs J113 with 150 Ohm Source resistors giving around 4mA Iq per J-fet (theoretically around 0.2uV wideband noise as Ein, or -134dBV) giving 32mA/V Transconductance.

For MM just use a single fet (e.g. shut the remaining 7pcs off), for an Ein of around 0.6uV/-124dBV and 4mA/V transconductance.

With a 100K +10n * 4 + 8k2 RIAA gain will be around 30 (30dB) for MM & 240 (48dB) for MC. The second stage needs very little gain then.

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Fri Sep 08, 2017 3:22 pm 
Offline
Cow

Joined: Fri May 30, 2014 6:24 am
Posts: 132
Did Jocko say something about making things harder than they should be? :D

Here a “nothing fancy” version. Except that it has more gain than T. suggested (to be able to put power buffer after volume pot).
And the tone control is just "bass boost" parallel with volume pot.
Attachment:
Chain.PNG


Am I missing something obvious, because I think that the folded cascode current source (just inductor & resistor drawn in the schematic at the moment) should be servoed from RIAA stage output voltage?

The CCS should provide Jfet bias current and 5 volts across the 100k RIAA resistor to make the DC coupling work.
I think temperature drift causes the need for a servo?
Something is wrong here.. :scratch:


You do not have the required permissions to view the files attached to this post.


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 09, 2017 7:39 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1130
Hi,

hanski wrote:
Did Jocko say something about making things harder than they should be? :D


This is nothing like what I had in mind... But heck, so what.

Mine might be more like this:

Attachment:
simple_riaa.gif


On the Sim, 75dB unweighted SNR re. 0.5mV input, 50dB gain. So the following stage needs a gain of 10 - 20dB for a total gain of 60-70dB and high input impedance. And it should be inverting.

Ciao T


You do not have the required permissions to view the files attached to this post.

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 09, 2017 7:08 pm 
Offline
Cow

Joined: Fri May 30, 2014 6:24 am
Posts: 132
Inspiring T! Thanks!
I think a nice detail is that the folded cascode reference point is tied to +15V instead of ground.
Then you can use the 120R resistor to feed the folded cascode (+15V PSU ripple doesn't come through).

But if I read it correctly, the +15V PSU ripple goes directly to input Jfets drain node.
As an improvement, one could keep the servoed CCS parallel with RIAA load as it is your schematic but feed the folded cascode with CCS?
Then the folded cascode could be referenced to ground. Improved PSRR.

That way we could nicely avoid going into PSU discussion. ;)
Does this make any sense?


Top
 Profile  
Reply with quote  
PostPosted: Mon Sep 11, 2017 7:42 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1130
Hi,


hanski wrote:
Inspiring T! Thanks!
I think a nice detail is that the folded cascode reference point is tied to +15V instead of ground.
Then you can use the 120R resistor to feed the folded cascode (+15V PSU ripple doesn't come through).


The 120R just increases the impedance in the tail of the folded cascode.

At the around 6mA our transistor runs at the emitter impedance becomes around 4 Ohm. With the extra 120R resistor er get a tail impedance of 135 Ohm, or 33 times the emitter impedance.

hanski wrote:
But if I read it correctly, the +15V PSU ripple goes directly to input Jfets drain node.


Yes, a low noise PSU should be used, however, as the drain impedance of each J-Fet is quite high (around 11k for 8pcs in parallel) it is less a concern that it might seem.

Still, a 7815 without added RC filtering would probably not do, PSRR on the positive rail is around 0dB, meaning noise there ends up at the output 1:1.

hanski wrote:
As an improvement, one could keep the servoed CCS parallel with RIAA load as it is your schematic but feed the folded cascode with CCS?
Then the folded cascode could be referenced to ground. Improved PSRR.


The problem with a CCS feed that you need to account for the variable Idss of the J-Fets. Complexity would also rise. Why not just add a pair of BJT Capacitance multipliers to kill noise flat? I could have included them, but prefer to leave such minor (if important) details out as copy protection for idiotic chinese who like to make a quick buck ripping off stuff they cannot understand.

hanski wrote:
That way we could nicely avoid going into PSU discussion. ;)
Does this make any sense?


I find making circuits excessively complex and sensitive to parameter variations just so we do not have to think about the power supply counterintuitive. It is an intellectual game much beloved by John Broskie, who never justified the need for high PSRR by any other reason - quite frankly that is bogus anyway, most of the time real circuits with normal parameter variatiosn have much less extra PSRR than in the Sim/Theory and you still have to make a decent PSU.

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Tue Sep 12, 2017 3:35 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1130
Hi,

Kuei Yang Wang wrote:
On the Sim, 75dB unweighted SNR re. 0.5mV input


Found my A-Weighting Sim Circuit and checked. A-weighted 80dB re 0.5mV, as LP as 0dB @ -15dB (max) we get 95dB usable dynamic range, if any LP is ever that quiet, in the midrange anyway.

Having done similar circuitry with 2SK170 before I think the SNR is realistic. Using 8pcs of "common garden" J113 (or SMT versions) delivers SNR within 1dB of a quad of 2SK170, just make sure to not by the NXP/Nexperia versions which are very noise, Fairchild/ON (now re-united) are quiet.

For the PSU's, simply use a Sziklai pair with a BC550C/560C (or similar) and a suitable bigger (BD139/140) added pass transistor. That knocks 40dB off any general PSU noise meaning normal noise from cheap 3-Pin regs is killed stone dead. The circuit does draw around 50mA on the positive line.

Being an annoying Tubeophile, I'd probably use a tube as second stage, a 5687WB SRPP looks good and can stand in as line stage with gain. Making the 100K resistor a stepped attenuator allows gain to be set.

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Wed Sep 13, 2017 7:46 am 
Offline
Cow

Joined: Fri May 30, 2014 6:24 am
Posts: 132
Ok T, I give up. I’ll use the given RIAA circuit as is.
It isn’t easy for me, I have the urge to put my fingerprint in everything. ;)

I also give up the idea that the following stage should have a high output voltage.
I understand it would be good to use 5687 tube there. But here the used voltages should be children level.

And the stepped attenuator is a natural solution. But…
How about using similar stage as in RIAA for line stage also?
Change the Jfet to a low transconductance type and put enough degeneration to limit gain.

Then for volume control, one could use (binary weighted) resistors to ground (7bit resistors would be about 30db attenuation range). Would perhaps be enough here.
I could then use my favorite MCU, Arduino for the switching task. :grin:

What would be the recommended switch type for resistor matrix?
Miniature relays or something sand?
In the beginning of this thread you mentioned Jfets for AC voltage switches.


Top
 Profile  
Reply with quote  
PostPosted: Thu Sep 14, 2017 9:31 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1130
Hi,

hanski wrote:
Ok T, I give up. I’ll use the given RIAA circuit as is.
It isn’t easy for me, I have the urge to put my fingerprint in everything. ;)


I was merely illustrating my suggestion, for something simple.

hanski wrote:
I also give up the idea that the following stage should have a high output voltage.
I understand it would be good to use 5687 tube there. But here the used voltages should be children level.


Well, the Metaxas Line stage would do great. And the +62V allow a lot of passive filtering for the PSU to the Phono stage.

You could even have switchable gain between line in & phono in, to have different gains (easy to do with a differential).

hanski wrote:
How about using similar stage as in RIAA for line stage also?


The stage shown is explicitly for low noise, low level signal.

I am not great fan of repeating identical topologies multiple times in series. It emphasises the relative weaknesses of the topology.

Using the shown Phono plus J-113 in the Metaxas Line stage with (or without) Fet/BJT Sziklai or even cascoded Fet/Sziklai.

hanski wrote:
What would be the recommended switch type for resistor matrix?


Mercury wetted reed relays.

hanski wrote:
In the beginning of this thread you mentioned Jfets for AC voltage switches.


Yes, they can be good. As with all things, it's more the how than the what that matters.

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 88 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6  Next

All times are UTC


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group