Capacitance multiplier: which Darlington to choose?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Unfortunately, my EL156 pentode Regulator produces more than the regular noise. So first trying to bring this down.

I suggest you learn to use a simulator and read up on the Tube Regulator.

tech300B

Need this Low noise voltage for the phonostage Experiment. It does now hum Like hell.

I would use an LC Pre-filter and then work out the kinks. There is no reason for regulators to hum, but watch heater-cathode leakage and the parasitic heater-cathode diodes.

Compared with the EL34 triode Regulator of the Fairchild 670, which one will be better suited for an audio PSU, pentode or triode regulator?

I have only use Triodes as pass elements, generally speaking "high perveance" types.

Use a pentode as Error Amplifier.

If connecting the load as I show, the load will be bootstrapped by the pass element and gain is maximised, plus noise is minimised, as the error amplifier is supplied from the regulated output to lower noise.

Supply the screen and the reference from the regulated supply as well.

I thought a triode regulator has lower Output impedance, but might be wrong on this.
Both regulated PSU needs extra 200V, means the Transformers have to supply this extra voltage. If I would use my Siemens Tube E235L regulated PSU, it would only need 100V extra voltage in triode mode. Much more efficient.

Use 6AS7/6080 which were designed as regulator pass element, or why not go hybrid and use a MOSFET?

P. S. Nice amp! My F2A uses EL84 driver, too. Thats quite the Sakuma principle, beefy driver section for good sound.

It was one of the first 300B Amp'd I build I felt was good enough to share AND reliable and simple enough to build.
He liked to use cheap TV or Radio set tubes, so why Not cheap B+ regulators, too? LOL

Indeed. As said, sometimes excessive 2nd guessing is useless.

You think he destroyed the advantages of achieving a Low impedance B+ Output with heavy filtering after the IC?

Well, the chokes add a lot of DCR AND impedance. So the low output impedance will be swamped out by this.

Say a regulator has 100mOhm ouput impedance, we use a 10H/100Ohm Choke.

At DC the output impedance of the combination will be 100.1 Ohm (basically Choke DCR) at DCR and 6,283.3 Ohm at 1kHz.

And yes, changing the regulator one one with 10mOhm will not have a relevant change and having a 10 Ohm output impedance will still have minimal effects.

So any Choke Filters should be placed prior to the active component?

Depends on your goals.

I suggest that a first bulk filter first, solid state regulator and RC filter chain with relatively low value resistors and film capacitors preserves the classic "tube sound" without "the sound" of the regulator interfering, while getting good operating condition stability, minimal noise etc.

So hence my setup of:

CLC -> active Reg -> RC to tubes

for my stereophile Class A rated 12,000 Euro tube/Hybrid phono stage.

I really think he put passive filtering behind it, but its no safe bet.

I think so, because my experience is that even tube regulators with high feedback tend to cause what I'd call "overly tight" solid state sound to classic low/no feedback tube circuitry.

Prior to His IC are just Caps and a resistor, means the active Filter Starts very near after rectification.

Yes, have a first bulk capacitor to hold up line, use modern electronics for a bulk of noise rejection (this has penalties) and then a classic LCRC filter chain will give the most classic tube sound.

He could have applied it the wrong way, Just because to His ears it meant sounding better.

There is no "wrong ways", there are different goals and different ways of achieving them.

Will this, technically speaking, be the best serial Tube regulator? A pentode? Maybe I got this wrong? Please explain. Thanks.

What I showed is a high power tube rectified and regulated regulator with up to 2 X 100mA @ 450V output and avoiding any electrochemical capacitors (that's a hobby horse of mine).

It was one of my earlier attempt's when I was still overly influenced with a Pro-Audio and Industrial Electronics designer mindset and more of a Technocrat and given to achieving my design goals using orthodox "brute force" feedback based methods. I didn't know better.

Over the years I tended to avoid using ANY regulated supplies where this is possible. Yes, it needs more inductors and capacitors to get low noise, more size, weight, cost but to my ears I preferred the sound. You do need stable mains for that of course.

TANSTAAFL.


Thor
 
Last edited:
Walter66 said:
Need this Low noise voltage for the phonostage Experiment. It does now hum Like hell.

A properly designed and implemented phono tube preamp should not "hum like hell", even without a regulated PSU.
Look elsewhere for the cause. As Thor suggested, floating heaters, improperly locally-decoupled HT, hum loops...
 
Walter66 said:
Need this Low noise voltage for the phonostage Experiment. It does now hum Like hell.

A properly designed and implemented phono tube preamp should not "hum like hell", even without a regulated PSU.
Look elsewhere for the cause. As Thor suggested, floating heaters, improperly locally-decoupled HT, hum loops...
OK, this was misunderstandable from my side. My actual Experiment Power supply does an uncomfortable hum, not the phonostage. Its a Wandel & Goltermann unit, you May have never Heard of that company, they produced the finest Professional measurements Equipment in Germany Post war era.
To bring it up to specs, it needs some work. After thats done, the preamp can be constructed. It will work in Combination with the discussed and to be designed PSU in this thread, Not the Wago unit.
 
Last edited:
Depends on your goals.
I think so, because my experience is that even tube regulators with high feedback tend to cause what I'd call "overly tight" solid state sound to classic low/no feedback tube circuitry.
Understood. It shows the pro's and con's of using those circuits clearly.
Yes, have a first bulk capacitor to hold up line, use modern electronics for a bulk of noise rejection (this has penalties) and then a classic LCRC filter chain will give the most classic tube sound.
Interesting, will try.
What I showed is a high power tube rectified and regulated regulator with up to 2 X 100mA @ 450V output and avoiding any electrochemical capacitors (that's a hobby horse of mine).
Still using them, my gear is mostly in the Tradition of golden age tube amps. I Like this sound most.
It was one of my earlier attempt's when I was still overly influenced with a Pro-Audio and Industrial Electronics designer mindset and more of a Technocrat and given to achieving my design goals using orthodox "brute force" feedback based methods. I didn't know better.
My mindset is influenced by all great amp Designs of the past, mostly Studio Gear, as this had the highest Standards of design and construction.
Over the years I tended to avoid using ANY regulated supplies where this is possible. Yes, it needs more inductors and capacitors to get low noise, more size, weight, cost but to my ears I preferred the sound. You do need stable mains for that of course.

Thor
Me too until now, but I have'nt designed a perfect sound and measurements PSU yet. That's not as easy with just using passive parts so I need a little help from IC's, too. Space is limited in the box.
We shouldn't think too dogmatic about the subject. If it sounds good, everything is fine.
 
Last edited:
We shouldn't think too dogmatic about the subject. If it Sounds good, everything is fine. Pragmatismus is the way to Go For me.

I don't. I first did s lot of work to figure out what sounded most real to me using low NFB tube Gear.

This led to the AMR products.

And with a suitable reference and understanding, the next stop was to find ways to get the most of that sound in solid state and hybrid circuitry that would be "commodity" based and priced.

That led to iFi and onwards.

I have no problems using anything available that gets me what I am looking for in sound.

Thor
 
Non linearity is a non issue in this case, because there is significant NFB.
As to the relevance of Gm as the dominating factor of performance in this application, hmmm...
Thor,
do you think that, because of significant NFB, the Mosfet or BJT cap multiplier causes the same negative effect on the Sound which you described as being "transistor like" or do they less harm compared to tube voltage Regulator circuits?

For my own experience, the pentode EL156 regulated PSU do sound good compared to my unregulated PSU.

The Fairchild 670 compressor is a legend in Studio Sound, often used with the highest selling records Like Beatles or The Who. I can't audition anything Like a transistor sound on those records. It used EL34 triode mode as a series pass element for it's regulated PSU.

One could put this attribute even to a TAB V72 (Beatles console for mic amping) or TAB V76. Would you put this Label For both, too? They Sound Crystal clear, but much improved to Transistor amps at the Same time, to me.

If so, one could better understand what you meant. In the end, it's much subjective what we Like or dislike in Sound. But all Details should be Audible, thats opposite to steam Tube Radio Sound IMHO. If that's Kind of Like "Transistor Sound", I can Accept it.
 
Last edited:
do you think that, because of significant NFB, the Mosfet or BJT cap multiplier causes the same negative effect on the Sound which you described as being "transistor like" or do they less harm compared to tube voltage Regulator circuits?

I find looped feedback across multiple stages often problematic, especially "early" circuits.

For my own experience, the pentode EL156 regulated PSU do sound good compared to my unregulated PSU.

Now, the next question is "what is your unregulated PSU like?"

For example, if we look at Kondo Baransu:

1687418142266.png

(this unit belongs to a friend in HK)

We find lowish impedance power transformer, single 5U4/GZ34 rectifier, 4 X 100uF+100uF 500V Black Gate Power supply capacitors and a ~ 5H Choke.

Seven of the 8 sections are connected as main power supply reservoir in parallel, the PSU looks like this:

1687418509725.png

With 6.5mV P-P and almost pure sine we get 2mV of 100Hz ripple.

The output stage is fixed bias 300B with a higher than common primary.

I will assume 3k - 400V/80mA/3k will produce > 10W with -26dB H2 (5% HD).

Noise across the transformer primary will be ~ 1.8mV and signal levels at ~180V for max out and 54V for 1W, so unweighted 1W SNR would be 89dB improved by nearly 20dB by A-weighting.

The impedance of the power supply at 30Hz will be ~ 7.5 Ohm and the AC load impedance for a mono signal ~ 1,850 Ohm.

So the ratio between load impedance and source impedance is 0.004, commonly 0.1 is considered adequate, so here we have a very solid "simple PSU", low enough noise and impedance.

Black Gate "Electron Transfer" are among the most "clean" sounding electrochemical capacitors that used to be available (long gone) so other than costing now 1,000 USD per capacitor and costing 200 USD each even when they where available, giving ~1,000 USD PSU cost, they do fine.

Modern Oil Filled "Motor Run" MKP's are a good alternative, but 800uF of these will take some serious space.

My answer of a modified Verdier regulator with only film capacitors achieved a lot of sound quality of the Kondo PSU at a more modest cost.

The Fairchild 670 compressor is a legend in Studio Sound, often used with the highest selling records Like Beatles or The Who. I can't audition anything Like a transistor sound on those records. It used EL34 triode mode as a series pass element for it's regulated PSU.

I think attributing "Beatles Sound" and "The Who" sound to "EL34 Triode pass element in the PSU of the Compressor is a bit adventurous.
One could put this attribute even to a TAB V72 (Beatles console for mic amping) or TAB V76. Would you put this Label For both, too? They Sound Crystal clear, but much improved to Transistor amps at the Same time, to me.

Or the use of TAB Amplifier cassettes.

Talking about 60's-70's Transistor gear is not useful. We see people gasping for ways to make dem newfangled 3-legged fuses work well, while tube gear was very mature.

I do not know how much you know about recording music, I can recommend this documentary (though it covers after Beatles time, Macca is in it and a lot of my favourite musicians).

Sound City (2013) | WatchDocumentaries.com

In the old days, recorded sound starts with instruments and players (or vocalist), the room, then a microphone which will overload early at LF as it is an old Tubed Neumann, the transformer input Mic-Pre will overload early at LF. Then we send the Signal to Compressor and a Tape machine with input and output transformers each and more early LF overload.

Tape will have minimum distortion and maximum output levels at ~ 300Hz. If, like me and everyone I knew you run metters "pegging" tape distortion was substantial.

I had made a my own peak meter pre-emphasised and calibrated to light green at 0.5% HD, yellow at 1% HD and red at 3% HD, naturally frequency dependent. I always tried to limit reds to short flashes and to see as much yellow as possible.

This itself gives a kind of "Aphex" effect for both HF & LF with soft peak compression that crushes the tops.

That is just tracking. Does it sound distorted or clean? You guess.

Nowadays it looks like this of course:

1687420208702.png

If so, one could better understand what you meant. In the end, it's much subjective what we Like or dislike in Sound. But all Details should be Audible, thats opposite to steam Tube Radio Sound IMHO. If that's Kind of Like "Transistor Sound", I can Accept it.

Ok, the "Grosssuper" in the 50's & 60's is very much a german phenomena, much of the rest of the developed world had moved to separates.

The distinct sound has a number of causes and sources. It exceeds the format of this forum, involves (naturally) secret Nazi Science and MK Ultra level projects as well as as the soviets (really NOT making this up).

But Speaker drivers (high Qt open baffle) and arrangements (raumklang), undersized output transformers, typical passive components including nonlinear ceramic capacitors used for coupling and undersized electrolytic capacitors in power supplies all contribute.

And yes, the dampfradio sound does "homogenise" instruments yet, vocals are absolutely clear.

To my experience the biggest contributors to the homogenising characteristics are caused by transformers with fairly thick steel laminations and electrolytic capacitors in power supplies for tube gear.

You can compensate by using parts, tubes and circuits that have a "bright sound" that allows the mids and highs to cut through some of the "elko schlamm" and "stahl nebel" but you start using balancing complementary colorations, not removing colorations.

Doing this "balancing act" is challenging, like cooking with heavy spices and balancing everything just so. Get it right, it's great, be off just a small bit and all you get is overwhelmingly spicy without great taste.

Solve those instead, you end up with something like the crystal clear like a pure bell Kondo Sound which nevertheless has warmth and is never abrasive.

Thor
 
Last edited:
I find looped feedback across multiple stages often problematic, especially "early" circuits.
So you mean audio amp stages and Not PSU-NFB in this case?
I had No Problems with Feedback loops until they act stable. Every pro-tube audio company did in the golden days. The engineers knew their stuff.
Now, the next question is "what is your unregulated PSU like?"

For example, if we look at Kondo Baransu:

View attachment 110520

(this unit belongs to a friend in HK)

We find lowish impedance power transformer, single 5U4/GZ34 rectifier, 4 X 100uF+100uF 500V Black Gate Power supply capacitors and a ~ 5H Choke.

Seven of the 8 sections are connected as main power supply reservoir in parallel, the PSU looks like this:

View attachment 110526

With 6.5mV P-P and almost pure sine we get 2mV of 100Hz ripple.

The output stage is fixed bias 300B with a higher than common primary.

I will assume 3k - 400V/80mA/3k will produce > 10W with -26dB H2 (5% HD).

Noise across the transformer primary will be ~ 1.8mV and signal levels at ~180V for max out and 54V for 1W, so unweighted 1W SNR would be 89dB improved by nearly 20dB by A-weighting.

The impedance of the power supply at 30Hz will be ~ 7.5 Ohm and the AC load impedance for a mono signal ~ 1,850 Ohm.

So the ratio between load impedance and source impedance is 0.004, commonly 0.1 is considered adequate, so here we have a very solid "simple PSU", low enough noise and impedance.
Thats a good rule of thumb, unknown to me.
Black Gate "Electron Transfer" are among the most "clean" sounding electrochemical capacitors that used to be available (long gone) so other than costing now 1,000 USD per capacitor and costing 200 USD each even when they where available, giving ~1,000 USD PSU cost, they do fine.
Sorry, never Heard. I prefer multisection Twist lock from Sprague (don't be vague, ASK For Sprague 😀)
Modern Oil Filled "Motor Run" MKP's are a good alternative, but 800uF of these will take some serious space.
Thats what I use, the Military stuff the best Tube amps we're build with. Indestructible.
My answer of a modified Verdier regulator with only film capacitors achieved a lot of sound quality of the Kondo PSU at a more modest cost.
Sorry, Not my Kind of pathology. I have No fear to use good electrolytics when necessary. But they don't come cheap. Thats what Most people have Problems with.
I think attributing "Beatles Sound" and "The Who" sound to "EL34 Triode pass element in the PSU of the Compressor is a bit adventurous.
Thats correct. But you can hear For yourself, Not the faintest shimmer of Transistor Sound.

Or the use of TAB Amplifier cassettes.
Thats what I referred to. TAB casette preamps V72&V76, the Rolls Royce of Tube preamps. Dozens of coils and Transformers, Sounds Like a clear mountain Lake still.
I do not know how much you know about recording music, I can recommend this documentary (though it covers after Beatles time, Macca is in it and a lot of my favourite musicians).

Sound City (2013) | WatchDocumentaries.com
Thanks. Excellent Studio Sound.
"Grosssuper" in the 50's & 60's is very much a german phenomena, much of the rest of the developed world had moved to separates.
Yes, the Reichsminister of Propaganda wanted it that way.
The distinct sound has a number of causes and sources. It exceeds the format of this forum, involves (naturally) secret Nazi Science and MK Ultra level projects as well as as the soviets (really NOT making this up).
Yes, known to me. I studied the sound of some electric parts decades ago.
But Speaker drivers (high Qt open baffle) and arrangements (raumklang), undersized output transformers, typical passive components including nonlinear ceramic capacitors used for coupling and undersized electrolytic capacitors in power supplies all contribute.
Correct. I hate undersized component. Thats why I Design it myself. All oversized LOL
And yes, the dampfradio sound does "homogenise" instruments yet, vocals are absolutely clear.
Vocals often are good.
To my experience the biggest contributors to the homogenising characteristics are caused by transformers with fairly thick steel laminations and electrolytic capacitors in power supplies for tube gear.
Good audio Signal Transformers are rare and expensive. Thats right. Good Transformers don't make For Bad Sound, that's what many people think. But its Untrue. The Fairchild is one example for this. Do you think it sounds Bad?
You can compensate by using parts, tubes and circuits that have a "bright sound" that allows the mids and highs to cut through some of the "elko schlamm" and "stahl nebel" but you start using balancing complementary colorations, not removing colorations.
Did you read "In Stahlgewittern"?
I Like Just the right amount of colorations, Today its Most often Missing unfortunately. Black and White Sound mostly. It's a shame.
Doing this "balancing act" is challenging, like cooking with heavy spices and balancing everything just so. Get it right, it's great, be off just a small bit and all you get is overwhelmingly spicy without great taste.
Yes, to become a Chief Cook is hard but a thrilling adventure.
Solve those instead, you end up with something like the crystal clear like a pure bell Kondo Sound which nevertheless has warmth and is never abrasive.

Thor
I Hope to! Thanks very much,Thor!
 
Last edited:
Bonjour,
pour un multiplicateur de capacité qui alimente une section phono du préampli du tube audio, quelqu'un peut-il vous conseiller de mettre en œuvre un gain élevé TO 220 Darlington Transistor dans ce circuit? Il devrait fournir une suppression du bruit élevé et un excellent son. Je vous remercie.
View attachment 109449
Hello Walter Does this post always active for you ? you can find an article at Elliot sound 'mutiplier capacitance. It use sziklai transistor . Sorry for this bad English
Hello,
for a capacitance multiplier which powers a audio tube preamp phono section, can someone please advise for a high gain TO 220 Darlington Transistor to be implemented in this circuit? It should provide high noise supression and excellent sound. Thank you.
View attachment 109449
 
Hello,
after a longer relocation process, I intend to finish this PSU cap multiplier project.
Did found a nice chip with high current amplification factor and high voltage which should be used.
How to dimensionate the R1 resistor in that given circuit for a current pull of 50mA @320V, please?
Can I go as low as 190 Ohm (saw that in another circuit)? What's the optimum value for the resistor?
 

Attachments

  • 2SD1409.jpg
    2SD1409.jpg
    569.4 KB
  • IMG_20231116_215048.jpg
    IMG_20231116_215048.jpg
    197.8 KB
Last edited:
That darlington connection is basically an emitter follower. The output voltage depends on the input voltage minus the sag from R1. The base-emitter resistors connected across each base-emitter will define an almost constant current draw based on Vbe/R. With two in series the lesser current will dominate. So this current + the base current feeding the darlington times R1 will definer the voltage drop.

This might be a good project for circuit simulation...
===
Note you can breadboard that darlington up run from a lower voltage supply just to figure the drop across R1. How many volts is your unregulated?
JR
 
R1 and the 100uF capacitor determine how much the ripple is reduced so from that perspective the bigger R1 is the better. However, it also has to provide base current to the Darlington. This has a current gain of about 600 so your 150mA load translates into 150/600 = 0.25mA of base current. So if R1 was 10K you might expect about 2.5V drop across it so the output voltage would be about 2.5 + 1.3 = 3.8 volts below the input voltage. R1 at 10K should give you over 50dB of ripple reduction.

Cheers

IAn
 
Does adding zeners to transistor base have any real advantage for class a circuits, except having stabilized voltage? I'm frequently using very similar circuit to Walter's (zeners sets the voltage) with only one more capacitor and resistor. As designed with TIP50 it works at 250-300V/50mA max, components are just a little warm up to 25mA making it very cheap, easy to p2p.
 
As John mentioned, R1 has to supply both the base current for the Darlington and the current that circulates in the base-to-base resistor (R1 in teh spec sheet), which value is unknown.
The latter is likely to be comparable to the base current, particularly considering using the Darlington at about 1/100th of its rating.
You may try to measure the base-to-emitter DC resistance; since R1 is likely to be significantly higher than R2, that would give a ballpark figure.
Then you may try to calculate the constant current drawn by the Darlington, using I=0.7V/Rbe.
This current will come in addition to Iout/Beta, in your case 50mA/600 or a little less than 100uA.
Then you would use that to determine the voltage drop the total current produces.
Knowing the non-regumated voltage and the desired output voltage, it's easy to dtermine your R1.
 
How to dimensionate the R1 resistor in that given circuit for a current pull of 50mA @320V, please?

Is it 50mA or 150mA?
Can I go as low as 190 Ohm (saw that in another circuit)?
Because the calculation is sbject to too much unknowns, I suggest you experiment with it.
190 ohms is certainly way too low, but not likely to result in failure of teh circuit. Il may result in failure of teh circuit is is connected to, so I would suggest caution and using a dummy load.
320V at 50mA is 16W, 48W at 150mA, so this dummy load should be commensurate.
Actually, with such a low value of R1, the output voltage would be close to the unregulated voltage, which you haven't specified.
 
As John mentioned, R1 has to supply both the base current for the Darlington and the current that circulates in the base-to-base resistor (R1 in teh spec sheet), which value is unknown.
The image file name seems to imply the part is the Toshiba 2SD1409 in which case the two resistor values are 2K5 and 200R respectively:

2SD1409 Datasheet PDF Download - Toshiba

Cheers

Ian
 

Latest posts

Back
Top