Voltage DAC and grid choke

Design and use of Chokes for PS, anode and filament use.
Post Reply
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

Voltage DAC and grid choke

Post by oko »

Hello,
I would like to ask someone more experienced than me to have a look at attached schematics and ask if this circuit makes any sense.
Basically I’m using this line preamp with my Cambrige Audio Azur CD 640C. Because I could not find Wolfson DAC in Eagle library for drawing schematics I used random DAC. Current source is IXP10M45S. Others elements/values are on schematics.
I’m "stilling" signal directly from DAC legs bypassing whole analog part of circuit.
Till yesterday I was using grid resistor 100kR and that was pretty much common setup. I’ve decided to give a try grid choke and it plays even better. Perhaps I’m supersensitive but I have noticed that something strange going with bass response. Since I have no tools to measure it I can’t confirm my observations.
So, my concern is about using grid choke in this position in between DAC - coupling cap and tube grid.
If idea is right I wonder how to properly calculate coupling cap to avoid any problems with audio bandwidth.
Best Regards,
Oko
Attachments
12B4A_Wolfson.png
12B4A_Wolfson.png (8.11 KiB) Viewed 26384 times
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

From the values you listed, the CL resonance is at about 8hz and you don't have any damping so you could be having some issues.

I looked at the wolfson application notes and it has a differential out, are you simply taking 1/2 of that output for your SE analog out?

dave
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

Post by oko »

Hi Dave,
Thank you for answer.
Yes, I'm using only SE signal. The next thing what I'm going to do is use transformer for conversion balanced signal to SE. But this is another story and I will have some questions to you since I would like to order proper transformers.
But at this moment have to stay like that.
I'm not expert so could you possibly write more about damping thing?
Thanks
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hey Oko,

Cheap solution: If you for some unknown reason want to run your Wolfson single sided there is no need for neither cap nor choke, dump them! They can only worsen performance.

Just apply the signal directly to the grid and compensate for the added 2,5V bias by changing cathode-resistor.

Instead of a gridchoke an anodechoke on the 12B4 will give better results.

What do you need the high gain for? The half-Wolfson will give you 1Vrms which means ca 5Vrms at the output. If no gain is needed skip the tube and go directly with a cap from the DAC.

If you want to do it properly differential is a must.

A 1:1 transformer can be used as you mention. But it must be tailored to the secondary load.

If you want to go active with low Zout, the one below is a simple way to do it.
Last edited by reVintage on Sun Nov 15, 2009 6:26 pm, edited 1 time in total.
Brgds
Lars
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

I didn't see in the data sheet if the outputs were floating at 2.5Vdc and wondered if the DC were blocked inside the chip. Looking at Lars suggestion, I suspect they are at +2.5V dc and I'll second this as an approach.

Now if they would only design these things to work off -5 rails we could ground the cathode of our tubes :-)

dave
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Output DC is at Vdcc/2.
Brgds
Lars
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

Post by oko »

Hi Lars,
Thank you for suggestion but I’m not sure what do you man saying compensate voltage.
Should I increase cathode resistor to set grid voltage to 13V (11V + 2V) or should I decrease cathode resistor to set grid voltage to 9 V (11V - 2V).
Sorry for such dump question but as mentioned in first post I’m nob in this subject.
Also I was trying to skim whole tube amplification and connect DAC via coupling cap to amplifier.
Unfortunately for some reason this configuration sound not to good.
Regards,
Artur
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hey Oko,
Rk will be in the ballpark of 1 kohm. Don´t know about your direct-connection but maybe it has something to do with the single-mode?
Brgds
Lars
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Just looked at the wolfson sheet again and there appears to be a "mid rail" reference which should be the DC offset point and AC ground for the chip.

Can you just hand a SE bifilar off of one side or the other and be done with it? this would also give an effective phase reversal by choosing either the positive or the negative output. Do note that the caps from pins 11 and 18 to ground are in the signal path!

Hanging a 1:1 off of the entire balanced output will work and the impedances are low enough that the moving of the ground shouldn't be a problem.

dave
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

What impednace DAC sees if im using transformer..

Post by oko »

Hi Dave,
I got from you transformer 1:1 (custom made for Wolfson) for converting symetric sygnal to single ended. Could you explain to me what impednace DAC sees if im using this transformer and tube stage ? I'm sorry for perhaps stupid question but I'm novice in this matter.

As far as I techanical date of my transformers:

Ratio: 1:1
Bandwidth: 5Hz-70Khz
Core: 80% nickel.
Inductance: ~100H
PRI DCR - 63 ohm

Thank you in advance for answer ;-)
Regards,
Artur
Attachments
12B4_Wolfson.JPG
12B4_Wolfson.JPG (28.22 KiB) Viewed 25086 times
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

hello atur,

the dac basically sees the primary inductance which is frequency dependent.

at 100hy's the load is 2*pi*F*100 or 6280*F so at 20hz it sees about 12K and at 40hz it sees 24K. This value keeps doubling as you double frequency until midband where the shunt capacitance takes over and the value decreases at the same rate.

dave
Get Your Fix
www.hifiheroin.com
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

Post by oko »

Hi Dave,
Thank you for very fast answer..

Another question, transformer secondaries have low DCR, and relativly small inductance, can I add resistor in series or paraller with secondary to stabilize tube grid ?

Regards,
Artur
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hi Artur,
The grid is best stabilized with low DCR so no problem.

About inductance/reactance in the low band, you don´t need to worry as the Wolfson with it´s low Zout will not be sensitive to load at all. Anyway when load is of 12k magnitude.
Brgds
Lars
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

Post by oko »

Hi,
So seems that posted circuit is 100% correct and I should have no issues with driving from that setup integrated amp with 100kR input impedance ?
Am asking that because, this setup cusses serious distortions even at very low volume ;-(. At firs i thought that my problems are connected with not appropriate load of Wolfson.
Input tube in amplifier is 6n6p with -12V on grid. So 6n6p should not clip and before tube is potentiometer so signal from DAC is under control..
Yesterday in act of desperacy I added in series with secondary resistor 27kR (in between trafo secondary and bias battery). Problem gone, (coincident ?)

Artur
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Probably selfoscillation. Were is the gridstopper? Must be soldered directly at the input pin of the tubesocket. No other explanion for me. Down to 100ohm should usually work. Keep seriesresistor as low as possible as it usually causes loss of dynamics.

Your workingpoint of the 6N6P seems somewhat wacko. Should be somewhere in the ballpark of 120V/5mA/-4V.
Brgds
Lars
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

Post by oko »

12B4 is bitch to work with, because easy oscillate or simple injecting hiss to circuit. I'm using two 1kR resistors edirectly soldered do socket legs for each grid pins.
And this keep me away from tube oscillations. I'm using this tube at list two years in different configurations so I think I know how to handle all related problems.
Operating point for 6n6p is unusual but as far as I know is correct.
I do not cross plate dissipation, anode voltage etc. People tend to use 6n6p with low voltage.
I'm driving PSE EL34 and i prefer to keep "fast" power supply i decided to do not drop voltage for 6n6p to "standard" regions.

Operating point for 6n6p
Ua =224V
Ug = 12,0V
Rk = 1 kR
Ua = 12mA
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Why use 6N6P in a fairly unlinear region? 10-12mA is Ok but move Ua lower to better linearity. Shouldn´t have to be a problem at all to drive your EL34PSE. But if you have listened and found it best ignore this.

About your 12B4 remove both the 1k and 27k and instead bit by bit add a gridstopper that stops oscillation.
Brgds
Lars
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

Post by oko »

Hi Lars,
Thank you for your comments.
All my issues were caused by strange connection of 6n6p (I connected both halfs triodes in PSE) and operating point. I've change setup for SRPP and all issues are gone.
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hey Oko,
How could you use something like SRPP? In my point view this is awful compared to a well designed SE. IF you want to have low Zout the only thing you can do this to use a HighGm device on top of the triode in SRPP-style. It´s easy done with a depletion MOSFET like DN2540. This sounds OK!
Brgds
Lars
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

Post by oko »

reVintage wrote:Hey Oko,
How could you use something like SRPP? In my point view this is awful compared to a well designed SE. IF you want to have low Zout the only thing you can do this to use a HighGm device on top of the triode in SRPP-style. It´s easy done with a depletion MOSFET like DN2540. This sounds OK!
Hi Lars,
Yes, SRPP as well as cathode follower is on the very end of my list faworit tube circuits , and I'm trying to use CCS as much as possible (cascode based on DN2540N5 and IXYS IXTP01N100D from K&K Audio) this circut souds very good. But, in this case I'm having poroblem with high B+ for driver. Please hava look at attached simulations from PSUD. Corect me if I'm wrong, IMO fast PSU tend to sound better. Better transients reproduction, faster bass, sound is more open and detailed etc.
So in that case I decided to use SRPP instead of common cathode to keep lover voltage on anode without droping voltage using larg resistor. I would like to use CCS but I have no clue how to drop more that 100V/20mA without using big resistor in PSU. Another thing I don't know how to calculate voltage drop on CCS with given B+ and current.

I'm novice in tube circuits, so any input from you is highly appreciated !
Attachments
FAST PSU
FAST PSU
FAST_PSU.JPG (172.04 KiB) Viewed 24992 times
SLOW PSU
SLOW PSU
slow_PSU.JPG (182.28 KiB) Viewed 24992 times
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hey Oko,
You can have as high B+ as you like using a CCS. If you want low Zout take the signal at the source instead of at the anode.

Cathode resstor sets Ua and CCS sets current. Decide Ia and then adjust cathode resistor to get the right Ua.

If using a high B+ and low Ua just take care not to overheat the FET. So use heatsinks.
Brgds
Lars
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

Post by oko »

I'm afraid that your sugestion is not obvious for me, sorry for being knowledge resistant ;-). Could you calculate value of Rk for given input parameters ?
Lest say (for easy calculations) that we have:

Ia=10mA,
Ug=-2V,
Ua=100V

and B+ is 300V

Regards,
Artur
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Rk=Ug/Ia
Rk=2V/0,010A=200ohm
Brgds
Lars
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

Post by oko »

;-).

At first I thought that you suggested to drop 200V on cathode resisistor to keep anode-kathode voltage at 100V ;-), and that was odd to me..
That is why i decided to ask ;-)
Now every thing is clear. Today I'm going to test that..
Thanks Lars..
oko
Posts: 12
Joined: Sat Nov 14, 2009 10:56 pm

Post by oko »

For some reason I thought that using CCS with high voltage is more complicated.. Larger heatsink will do all job. ;-).
Once again thank you Lars for clarifying things.

One more question. I would like to try "Triode Trick" http://www.t-linespeakers.org/tubes/triode-trick.html. I'm having problem with getting 1N821 diodes. Can I buy instead of 1N821 any "voltage reference diode" at 6,2V ?.
Post Reply