• 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.

Partial Feedback Amplifiers

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Re: RH88 (6550) in DC mode

Alex Kitic said:
I am going to try it as soon as I figure how to achieve 600V B+ (that is going to need some serious thinking on how to use the available transformers and caps...)

Alex,

If you are using a FWCT rectifier, you can simultaneously derive a negative supply for the driver, provided that you don't exceed the total VA rating of the transformer. That might even do away with that huge cathode resistor dissipating 28W continuous :hot:
 
serengetiplains said:
It mustn't have taken the world by storm, me assume.

Indeed, no. The author missed the point as well. Naming something the "93 Williamson" shows a complete misunderstanding of naming principles.

Who cares if the input/phase splitter/driver arrangement is of conventional Williamson design.... the zener diodes at the screens make it completely different and novel. :Pinoc:

If it was named the "Ultra-tron" or something it would have been much more interesting :D

Jason
 
serengetiplains said:
Aliwakil, did you notice any detriments adding the feedback?

Hi Serengeti,
From subjective listening, less of the 'bloom' normally associated with tubes and sounding leaner overall. I recall reading someplace that a high driver output impedance may be an issue with partial feedback - can someone shed more light on this ?

I tried global feedback before as well. although a small amount of nfb resulted in better cone contol other aspects were unchanged. higher levels of nfb led to oscillations. eventually i took out it out of the circuit. i guess my global nfb experience is much inline with your comments about passing the signal through phase changing passive components.
 
aliwakil:
I recall reading someplace that a high driver output impedance may be an issue with partial feedback - can someone shed more light on this ?

When using local feedback the driver stage ideally should have high output impedance. The plate to grid feedback resistor makes the output stage act like a I to V converter with low open loop gain.

The impedance looking into the output stage is quite low from the feedback action. The input impedance of the output stage is Rfb/gain+1. As an example, if the stage gain is 30 and the feedback resistor is 180K the input impedance would be 180,000/31 or 5.8K.

In my amp's case we have the input stage with an output impedance aproaching 100K driving the output stage with an input impedance of ~6K.

serengetiplains:
I glanced more carefully at GP's feedback method and saw a few items of interest. First, plate to grid feedback reduces output PSU noise. Whatever such noise appears at the output tube plate also appears, in same phase, at the output grid, where it is amplified in opposite phase to thus cancel, to some degree, the original noise signal. Cancellation is only effective, it seems, to the limited extent allowed by some small but necessary phase shift in the opposite-phase signal.

You are correct. The output stage has very high PSRR. In the early versions of the 47 amplifier the total power supply capacitance was 30uf. Ripple was IIRC 30 volts or more. The amp was very quiet and sounded good but it sounded better when I added 68uf to the supply. The 1624 amp has 200uf in the supply and is even cleaner sounding. I think this is a good example that even with high PSRR a quiet supply is best.

The second item of interest is that the output AC signal---that precious, sought-after thing---appears, at reduced level, at the driver plate to there modulate the driver signal, to some small but, in my mind, probably not inconsequential degree. Ouch.

I don't think the feedback signal modulates the driver stage. The driver stage acts like an AC current source. With the high output impedance of the driver stage you can swing the driver plates around a lot and not change the AC current flowing into the feedback resistors.

One comment about current driven local feedback output stages. If things go wrong it is a pain in the rear to trouble shoot. The voltage waveforms are almost useless for seeing what is going on. Gave me bad flasbacks to trouble shooting the current driven paraphase amplifiers in old Tektronix scopes.

Gary
 
GOSSIP COLUMN:
Well, I usually must think of the words I'll use before I use them in the context of close neighbours... Choky obviously does not need to.
Further, SY and the rest -- do you remember what was it like when the Ongaku came around? Everyone tought that it was something novel, while SE was actually older than PP... and it took the world by storm, eventually. I am not implying that RH or whatever you may wish to call it should take the world by storm -- but I will be very happy to know, once in my old age, that I did a small step for the audiophile part of mankind :)

Aliwakil...
Yes, there is a limit to classic nfb all around the amp -- if I recall correctly, 20dB is somewhere near the limit. Furthermore, since the loop is long, what serengetiplains said is very correct.
With the short loop of plate to plate feedback, things are much more serene towards oscillations, PSRR etc. When it comes to sound -- well, I leave it to those who listen. I think that some of the bloom associated with DHTs and similia gets lost; but, that is a small price to pay for the lost of boomy bass, noise... and other residuals of the no-nfb-dogma. Furthermore, the gustibus nihil discutandum est... if you like it, great! If you do not... you can always say it's not your cup of tea.

When I started the whole RH thinking, I was after gaining power, trying to have a "free meal" -- the power of the pentode (efficiency) and the sound of that pentode as if it were in triode mode (much better sound). There are no free meals, but sometimes the meal can be so cheap that it is like it was for free. Gaining from 1.5 meager watts (for those who own <94dB/W/m speakers) to almost 5W is a great deal, everyone would agree with that. On the other hand, doing it with a 300B (8-9W standard with 5% distortion) with some improvements (10+W with 1% distortion) in power and a different sound (the bass would be faster and tighter) is something completely different, maybe a matter of choice?

Finally, some might have noticed that I wrote about fine-tuning the circuit... while in the end, nothing has changed in the values. Well, fine tuning is finding the right make of the same type of tube for each position: i.e. the difference in sound between an ST shaped 5R4 RCA black anodes with brown base and a 5R4 Chatham (the one with large ceramic base with glass looking like it was a pot) -- to take it on the brighter side -- is enormously large (at least to my ears) and there are also lots of other choices to be done. Generally, there is little to gain in power (i.e. some rectifiers are a little more efficient) but lots to gain in the overall sound. Lost midrange bloom can be found again... just exchanging the same type both other model rectifier, or driver tube...

Regards to all,
Aleksandar
 
do you remember what was it like when the Ongaku came around? Everyone tought that it was something novel, while SE was actually older than PP... and it took the world by storm, eventually.

Yes, that's one of the lessons I learned over the years: technical merit has nothing whatever to do with success or failure when we're talking about fashion and entertainment.
 
Official Court Jester
Joined 2003
Paid Member
just for clarity sake - let's we all search for various implementation of local and another sorts of feedback but without "who was first" attitude.
with that approach we will end in all sorts of trouble :clown:
it's marconi inventor of radio or is it Tesla.
regarding electronics I always think in "Jazz mode" -"who cares who made this tune first , this version is just avesome!"

To Alex-CD will be in post Office on Monday ,for sure ;I was too bussy this few days .
another one -who is responsible for da brain and for da hands of your OPTs ?
cheers to all !
 
My PP version

This is my implementation of a PP configuration.
I don't claim for any novelty, just look at playing with the values of R3 and R12 to balance between local and global feedback.

Of course, a 12AX7 and a pair of 6BQ5 could be substitued by adjusting bias (e.g. R16)

Yves.
 

Attachments

  • b1.gif
    b1.gif
    13.3 KB · Views: 3,817
Nice amp, indeed -- I noticed a whole lot of SS devices in there, but I hope they do just help it do it's best :) Although I do not own the famous bumper-sticker (Use a transistor...) I must admit that I would like to share that idea, if I were somewhere else where sharing ideas thru bumper-stickers and T shirts is worthwhile or any good.

As I might have previously said, ANY RH amp that was published to this moment can be made into a PP amp (with feedback and all...) provided you make it double... and you have a phase splitter in front of the two symmetrical amps per channel. It would work great -- actually, as great as the phase splitter is.

And, the more you know about it, the more you feel you would like to do it without a phase splitter. I believe that part of PP amps is the main reason why SE amps sound... well, better.

Regards,
Aleksandar

When it comes to transformers, Choky... I'll write about it in a mail. Nothing particular. If I were to mention anyone, or give a link, that would be free publicity, would it not? On the other hand, I think you know what a reliable source of trannies is: take a look at the "wanna build an amp with 50's thread" and you will find my post on that issue as well.
 
Hi !

Alex Kitic said:
Nice amp, indeed -- I noticed a whole lot of SS devices in there, but I hope they do just help it do it's best :)

Thanks !
Mmmh..., say I'm too lazy to constrain myself in not using SS when I think than using tubes in place won't (IMHO) have any positive effect ;)

....
And, the more you know about it, the more you feel you would like to do it without a phase splitter. I believe that part of PP amps is the main reason why SE amps sound... well, better.

Sure, SE and PP have different behaviors, specially looking on how the OPT works ...
Both may be made lovely or terrible sounding :cool:

Regards,
Yves.
 
frugal-phile™
Joined 2001
Paid Member
Re: My PP version

Yvesm said:
This is my implementation of a PP configuration.

And i was just wondering what to do with the ECL86 PP donor amp pictured :) It's a Viking with tubes made in Holland.

Cathode bias would be nicer... even better Class A with a CCS on the output cathodes (i happen to have a pr of 210VAC trafos that would suit a set of monobloks)

Alex -- the differentiial phase splitter with the CCS works very well. When Allen Wright showed his amp at VSAC he had a lot of died-in-the-wool SE guys rethinking things.

dave
 

Attachments

  • viking-pp-ecl86.jpg
    viking-pp-ecl86.jpg
    40.3 KB · Views: 1,273
PERFECTION

Jason (and others)...

You see, point is that when I make an amp, I usually do the math first, then make a breadboard and start "developping it". I try to achieve levels of perfection in the sound that are eventually unthinkable. Using NOS tubes, NOS rectifiers, perfecting the PS up to details... until the sound is like I wanted, or the best possible.

Finally, I publish it... but most details remain as they were in the original schematics. The schematics is just math, the rest is "metaphisics" or should we say another story.

Point is, some amps (like RH84) can achieve levels of sonic delight... some others, can't -- like the ECL86 amp I made in an attempt to achieve sonic perfection on the cheap... when you use a "combined tube" there is not much that you can do about the sound after building it. That might be a good choice for people "on a strict budget" or people who do not have lots of tubes to try...

PP amps -- same story as what I stated above. On paper, it's great. It does perform, electrically. But, the sound... just lacks the refinement possible with SE amps in general (with the same tubes). I have tried several alternatives for splitter, and prefer the "long tailed versions" -- but I admit I never did it with a CS. To make a CS I would have had to adopt an additional tube... and that was just out of the question at the time. And, after trying SS in various ways in amps, I left it only at the "hubrid bridge stage" where it does not affect the sound at all.

In between, I started making quite good sounding SE amps... and I was beyond the point of no return towards PP amps. Maybe one day, in my old age, pensioned, without anything better to do :)

Of course, no pun intended towards PP builders and listeners! I actually advocate making a PP amp as "symmetrical SE amps" with a phase splitter before it... transformer, maybe?

Regards to all,
Aleksandar
 
diyAudio Senior Member
Joined 2002
Hi,

Of course, no pun intended towards PP builders and listeners! I actually advocate making a PP amp as "symmetrical SE amps" with a phase splitter before it... transformer, maybe?

Indeed, a PP amp should be no more than a SE*2 + splitter.

A good splitter isn't all that hard to make, the common mistake however is to make it do things it just can't.
As for adding a CCS to a splitter, well that has been done fifty years ago already....Nothing new under the sun.

The main problem with SE amps, as I see it, is that you need the speaker to go with it.
If you don't, forget it....

We've brought tens of SE amps to their knees with ordinary speakers during listening tests, much to the embarassment of their proud owners who were invariably convinced they had designed the amp of the century.

Not to mention the fact that more often than not SE amps are essentially the sum of their parts. You really hear every single component you put in there....
But then, anyone who's been serious about designing a good preamp would have discovered that already anyway....

Naturally diy is not about designing the universally best amplifier. It's all about finding that one design that is willing to let the magic through....
Not an easy task to accomplish...

Cheers, ;)
 
SE amps and loudspeakers

Hi, fdegrove...

I am in full agreement with you, except for one tiny bit:

The main problem with SE amps, as I see it, is that you need the speaker to go with it.

I personally own a pair of Heybrook Sextet's -- rewired, of course, but completely original otherwise. They should be 88dB/W/m and that is what got me going towards the RH amps series.

With the 4-5W class of amps (including my 6B4G) listening was quite "fabulous" (depending on the sum of the ingredients), while it might have been possible to "put the amp on it's knees" if I was to demonstrate it to a bunch of hard-audiophiles. Actually, that never happened, and who was not happy with listening levels (that will not make the amp clip, as well as keep the neighbours quite happy, and all of the plaster on the walls...) has had to be aware that the amp is only some 4-5W.

Now, with the new 10W classs of RH amps, there is no holding back. Plaster off the walls, neighbours rallying against loud music, whatever... while the listening remains "fabulous" just like the sum of the ingredients (quite high in my case).

What is it that I am trying to tell? That SE amps are quite adequate for people owning NORMAL loudspeakers. And, I have not yet added to the list the most powerful group of SE amps (over 15W, and not counting the PSE kind).

Furthermore, loudspeakers will go up to some level of loudness: after that, distortions of the loudspeaker itself will make you understand that you are no longer listening to music but mere screams. Real-world test: my speakers won't go any louder with my QUAD 405 (a remembrance of the past, dust collector these days) than with the latest 10W class RH amp (the RH88/6550, to be correct).

On the other hand, a lot of SE amps were brought to their knees by EFFICIENT loudspeakers, showing their hum and other technical failings... i.e. 300B amps with AC on the heaters listened thru 98+dB/W/m loudspeakers. When you take into account that at least to me, DHT sound at their best with AC on the heaters, that is quite a huge drawback -- actually, larger drawback than having speakers that are >90dB efficiency.

Any opinions and personal experiences? Maybe someone lives in a Chateau and will like to share his experience with listening to SE and similia in ballrooms? My living room (and listening room) is 4m x 5m with a height of 2.8m... many people actually listen in rooms smaller than that... and, since it is quite "full" of books, carpets etc. it is quite "absorbing" as well.

Regards,
Aleksandar
 
On the other hand, a lot of SE amps were brought to their knees by EFFICIENT loudspeakers, showing their hum and other technical failings... i.e. 300B amps with AC on the heaters listened thru 98+dB/W/m loudspeakers. When you take into account that at least to me, DHT sound at their best with AC on the heaters, that is quite a huge drawback -- actually, larger drawback than having speakers that are >90dB efficiency.

Alexe,
we have to be very carefull when choosing AC heating for a SE DHT amp not only for the output tube but for the driver also( assuming a circuit with only two stages).Although I agree with you and with many others who vote for AC on heaters(definitelly the curves are not identical using AC or DC),it is very critical and dependable WHICH tube we decide to use for a driver .Personally I had very bad experience with 5687 (an excellent sounding tube BTW) regarding powering the heaters with AC.I did not manage to silence that tube using all techniques I know of(twisting,using hum pot,lifting 30V above ground,bypassing with cap...)So the hum originated more in the precedent(which contributed mostly to the final result) than in the output stage and thus was simply proceeded to the loudspeaker.Finally I concluded that it is the nature IMHO of that particular tube (maybe the heater winding of that tube is not suited for such application,as we already know that it was primarilly intended for computer usage)which was not tha case with other tubes I used instead of 5687(ECC... or 6SL/N7).

Regards,
Yugovitz
 
EAGLE FALLEN... :)

Hi to all!

I am still down with the flu, but I had to see whether there was something going on (after the first notice, they stop coming if you do not log in).

Well, let me start with the 5687 issue and AC heaters...

I own only two 5687 tubes, and have used both in a little project that spawned many variants, but never reached full boxed completion - with 6AS7/6080 output tubes, SE of course. ALL my amps are AC heated... I feel that even amps that use IH tubes benefit from AC on the heaters. Well, anyway, both 5687 tubes (black anode, Tungsol production, one marked "Los Gatos") showed the same behaviour as any other tube, meaning that there was no hum or buzz problem.

Therefore, if you have experienced such problems, that might be because of a faulty tube (or, eventually, if the tube was used in a phono preamp... and I did not understand that was the case).

When it comes to the 805 amp... did I miss something, I mean, where is the partial feedback, or anything we are discussing about? Anyway, nice and complicated schematics... only I must say that I shake my head every time I see some SS concoction in tube amps :) If you want to delay the heaters (for example) you can always use a second switch... unless you imagine that you are a manufacturer :)

Regards to all,
Aleksandar
 
Aleksandar,


If you want to delay the heaters (for example) you can always use a second switch... unless you imagine that you are a manufacturer
For your information the delay is applied to the high tension B+, so the tubes can warm up first with the heaters.

.
where is the partial feedback,
The partial feedback is the small secondary winding of the transformer that goes back to the 805 tube`s cathode/filaments. If that`s not partial feedback i do`nt know what is.

Anyway, nice and complicated schematics
I use ldr passive pre amp (no gain) you tell me what tube section i can delete and still have enough gain.

George
 
Alexsander,
Please help me out here by explaining how the output tube connections work.
I see where R5, the (1k) is connected between the anode and G2 and R4 (2.5K) is connected between the B+ and the junction of the anode and top of R5 so in my thinking the tube is working in some sort of a pseudo triode connection because of the connection between the anode and the junction of R4 / R5.

Also, is the main intent of this design to get a lot of power out of the tube? I ask because if I can read the voltage test points correctly (they are small and hard to read) the cathode is running @ ~ 110mA and the screen is at 577Vdc and iirc the maximum on the KT88 screen is 600Vdc. I have 8 Golden Dragon KT88 from a PPP amp I built around 15 years ago but have my doubts about them standing up to the "pressure" of this circuit.

I have all the parts for the RH 807 besides the tubes, the guy who was supplying me with 10 has not gotten back to me so I will have to use 6L6GC or KT 66 until I order some 807's from my usual supplier.

The RH 84 is still sounding great :D

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