• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

lytics and choke AFTER regulator

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Greetings,

I'd like to get some comments and opinion about using lytics immediately following the output of the B+ regulator *and* adding a choke before it gets to the final bypass caps (film or motor run) going to the anode.

As I understand the purpose of the regulator is to provide a constant B+ voltage no matter what the draw is on the load side, and to remove any ripples or noise that's not filtered by the CLC or CRC sections BEFORE the regulator.

I am aware that a few commercial designs that have chokes and loads of lytics AFTER the regulator, but doesnt that choke and cap combo make the regulator not able to see the voltage drop fast enough so it can compensate?

Whie measurng the B+ from this design usually gives me a perfectly flat DC, the overall sound becomes quite constricted, phasey, and slow. I'd like to see if fellow DiyAudio members share the same observations.

Cheers,
 
I don't see much point unless the regulator isn't very good. In that case, you'd be better off improving the regulator.

doesnt that choke and cap combo make the regulator not able to see the voltage drop fast enough so it can compensate?

Unless you're regulator remote senses, it won't see it at all. If it does remote sense, it could have stability issues.
 
Thanks for the response

Regarding the "remote sensing"

I am inclined to think that any well designed regulator circuit that uses a feedback mechanism would be able to
use the voltage sampled from the output side of the series pass and make adjustments to compensate for the voltage drop.

So if the regulator's output does go all the way to the anode (or at least anode resistor in most designs) with NO huge lytics or large choke in the way it should be working as it should be in compensating. But if the choke's in the way the regulator just wont see the drop unless it has some "sophisticated" remote sense where it probes the voltage a choke away from its output, right?

Please correct me if I am mistaken.

Thanks.
 
I am inclined to think that any well designed regulator circuit that uses a feedback mechanism would be able to use the voltage sampled from the output side of the series pass and make adjustments to compensate for the voltage drop.

If I understand you correctly, you're describing local sensing here.

So if the regulator's output does go all the way to the anode (or at least anode resistor in most designs) with NO huge lytics or large choke in the way it should be working as it should be in compensating.

Agree.

But if the choke's in the way the regulator just wont see the drop unless it has some "sophisticated" remote sense where it probes the voltage a choke away from its output, right?

Yep.

I just don't see any point in adding a choke to the equation. The regulator should be providing clean voltage. The choke increases output impedance. Don't we want an output impedance of 0?

Just a thought, are these commercial designs using a switching regulator by chance? I guess that I have been assuming that they are linear.
 
I guess it was a duct tape approach by the designer

I traced the circuit of the regulator to a textbook feedback amplifier (EF86) type regulator with a dual-triode (12BH7) as series pass element.

There is 20mV P-P or so AC ripple coming out of the regulator which I am not sure if it's
from residual ripple coming in from the CLC before the regulator or it's actually regulator tube filament induced (regulator tubes have AC filament with no raised potential).

So I am betting the designer of the preamp shoved in 200uF worth of lytics PLUS a choke before it gets to the audio board with 40uF of Sprague (also lytic :( ) bypass.

I boldly removed the choke and 200uF lytics and the ripple went from impressively less than 10mV to about 20mV (which I am not so happy about).

Readjusted the regulator to get B+ compensated to the right voltage
(choke was eating about 10V!), and the whole preamplifier transformed from having a very constricted tinny midrange to something that blooms.

Thanks for the confirmation of the goal of a low output impedance of the regulator.

Cheers,
 
Re: I guess it was a duct tape approach by the designer

phi70 said:
... There is 20mV P-P or so AC ripple coming out of the regulator which I am not sure if it's from residual ripple coming in from the CLC before the regulator or it's actually regulator tube filament induced (regulator tubes have AC filament with no raised potential).
I had similar experience with this regulator: http://www.enjoythemusic.com/magazine/sound_practices/3/one_tube_regulator.htm There was 240mV (!) ripple after the reg. The pass tube filement is lifted 150V which was the cause of the problem. By feeding it from a seperate supply ripple became nill.
I boldly removed the choke and 200uF lytics ... and the whole preamplifier transformed from having a very constricted tinny midrange to something that blooms.[/i]
Indeed, the absense of large capacitance looses coloration, especially in the basregion (in my case that was).
 
Thanks for the confirmation

My situation is a bit different.
If I read it correctly
your regulator tube had a raised filament but mine is just plain old AC filament.
240mV is indeed quite a bit. Interesting to see the tube type you use.
Saves a lot of space I guess.
Also I notice the ripple is pronounced when I use a ECC99 instead of a 12BH7 as the pass element. The ECC99 gives me a more dynamic sound at the expense of about 20mV ripple.
using the 12BH7 I get about 12-15mV. No choke no lytics at regulator output.
Cheers,
 
Re: Thanks for the confirmation

phi70 said:
My situation is a bit different. If I read it correctly your regulator tube had a raised filament but mine is just plain old AC filament.
Having both is possible but one needs a seperate filement supply.
240mV is indeed quite a bit. Interesting to see the tube type you use. Saves a lot of space I guess. Also I notice the ripple is pronounced when I use a ECC99 instead of a 12BH7 as the pass element. The ECC99 gives me a more dynamic sound at the expense of about 20mV ripple. using the 12BH7 I get about 12-15mV. No choke no lytics at regulator output.
How do you calculate the voltages for the error amplifier? Maybe of interest: there's rumour about a new SS super regulator, The Swanson. This thread is located at: http://www.diyaudio.com/forums/showthread.php?threadid=82641&goto=newpost
 
AX tech editor
Joined 2002
Paid Member
I don't think I should post it here, as that would make the people at AX probably a bit unhappy...

For the 6528 you need Va-k to make it work, of course. depending on the load, but surely more than 50V IIRC. Should be clear from the data sheet. OTOH, using a FET for the pass device, that runs with much less Vd-s, much cooler.

Jan Didden
 
Re: Re: Thanks for the confirmation

disco said:

How do you calculate the voltages for the error amplifier?

No I didnt calculate the voltages, it was not a real DIY effort but an investigative effort to FIX bad sound of a commercial product.
The original circuitry of the audio section was a close model of a famous Wada Shigeru SRPP follower added to a Marantz 7 gain stage. The power supply as far as I know is just textbook stuff, the final output B+ is tuned using a trimpot (with resistors on each side) connected to the cathode of the 12BH7 that adjusts the grid bias to the EF86.
 
janneman said:
20mV ripple from a regulator is just a bad regulator. And then to add injury to insult and add an L-C after it....

My tube HV reg has <500uV over audio band (<50uV over most of it). Will be published in AudioXpress in March or April issue.

Jan Didden

Interestingly enough this manufacturer of my preamp keeps adding more and more post regulation chokes to his newer offerings!
I am not sure where the 20mV ripple came from, Didnt see it in the raw B+ going into the regulator so I am inclined to think it came from the regulator filaments or bad tube choice.

Will definitely look at your design @ AudioExpress. Thanks and best regards.
 
Tube regulator circuit on AudioExpress

I came across an article on Audio Express on tube based regulators published earlier:

www.audioxpress.com/magsdirx/ax/addenda/media/bicknell2890.pdf


I tried to model the basic regulator circuit (figure 1) which is quite similar to the one that is in the preamp I am working on. I discovered that the g2 of the EF86 has a noticeably higher voltage than the control grid g1 and g3 being tied to cathode is referenced by the zener stack.

As a result the plate voltage of the EF86 is in the order of more than 100V below the
output tube's cathode. This leads me to suspect that the output tube is not really turned on.

May I ask if anyone has had experience building a similar regulator with an EF86 control
and if so whether the voltage on the EF86s screen grid (g2 between R5 and R6 in the diagram) really should be higher than the anode?!

Any pointers appreciated. Cheers,
 

Attachments

  • pastedgraphic.jpg
    pastedgraphic.jpg
    48.4 KB · Views: 197
never mind... the screen grid G2 of the EF86 was wired wrong

It was wired to the regulated B+ which is substantially lower than raw B+,
rendering current starvation on the EF86 which makes the plate output too low
and as Morgan Jones describes in his valve amp book I was getting the exact
"sulking effect" when wiring the screen grid to regulated B+ instead of unregulated B+
(which is the proper way as depicted in the AudioExpress diagram).
Sorry for the confusion.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.