|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
93.35.190.176
Hi
In attach the photo of LCR board, the LCR phono stage and the complete proto .
I wrote some article on LCR on udioreview magazine in Italy.
And I hve done four proto with Tango original stuff then I moved to develop a board following the Tango ideas
The indictors comes from Sowter. The tube are ECC81 and 6922
The results on test lab are fine.
The deenphasis error is within 0,1 dB.
Also the listening session was very good.
Walter
Follow Ups:
Walter,
Why did you build an LCR phono stage when everyone knows that it is very difficult (almost impossible) to make a coil that is perfect for such applications and the cost is also extreme. A good RC version is better and the 4-th filtering point (which was added later to the riaa) is easy to add.
Secondly, why a triode as a cathode follower, for this a pentode works better.
I forgot this one
Done years ago.
Useing pentode in gain stage
Walter
I did mean a penthode as cathode follower, for gain i prefer triode.
Ok
But a 6H30 will be a very good choice for that job
Walter
The 6h30 is a nice lineair triode.
For the cathode follower you can try a penthode or you change the circuit even more and use a ccs for the cathode follower.
Optional a White cathode follower.
I don't want to mix sand with tubes, so the CCS is not considered.
My opinion.
The only use of sand is in HT power supply because the goal is to kill the ripple and get low Zout
The reason to use in this circuit the ECC88/6N6/6H30 ( 12BH7 and ECC82 is a possible alternative) is due the very good results, test lab and sonic performances and the easier arrangement of pcb.
The layout is more simply with only two sockets.
I have in the lab lot of pentode able to do this job but I prefer to use in the gain stages.
I repeat the 6H30 is fine for CF, for me
Walter
Walge,
I was not specially meaning sand ccs. You could easely make a ccs design with tubes.
Hi
Yes but in my opinion there isn't a reason to use it when you have a simplest way to follow
Walter
well, if sound is important for you, it counts.
CCS for the cathode follower? Like the Allen Wright SLCF?
dave
Dave,
This is also an option. Allen Wright was a very capable man in designing tube amplifiers. I saw his work and also him at the ETF long time ago.
There are also other options if you like a less complicated design
in my view the achilles heel of all of this is finding a good sounding way to drive 600Ω.
Thus far I have not found a way to do it that I want to listen to for more than 15 minutes.
dave
Hello,
the only way to understand what happen in the real world is to test the circuit in the lab and not with simulation.
For this reason I confirm that the configuration of the LCR circuit is fine; the musical aspect is a different job.
I am not able to have a neutral critical opinion and, as I worte, I listen the opinion of the people involved in the listening test.
When they converge, in general aspect(each installation has a proper peculiarity) , I am sure that the circuit works well.
That's all.
And the presentation on magazine is always ( this is the same for all my project) after a good sonic results matched with a good test lab.
The measurement on LCR stuff on real circuit are rare ( I never seen) , mainly for Tango original stuff but also for a other clones.
Walter
As wrote before, I will publish the project only after an intesive listening tests.
Of course, the lab test must confirm the specs, as you can see from the results I published here.
Having more than a dozen of people available to play with these stuff I will have a right idea of what I am preparing.
The listening test are done with a huge and variable configurations so if the comments are in one direction ( with some differences, of course) I will have a good feddback and in case, I change something in the circuit.
It is a heavy but funny job.
I am not a good listener.
After that, I am preparing the last step, full trafo LCR. It will be a challenge.
But it will be a unique stuff not planned as kit.Walter
Edits: 04/29/21
What about a simple MosFet emitter follower?Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 04/28/21
Hey,
This enters the "slippery slope" area that often derails audio. A mofset follower absolutely will work and the measured results will be nearly "ideal". Since we all know "perfect" devices do not exist it brings up the old adage that just because you can do something doesn't mean you should. Way to many people in audio base their adver-speak on the engineering/technical size since that is unassailable. I tend to gravitate to and make note of those who have ears on experience that does not 100% align with the engineering status quo. Audio is one of those odd slivers of reality that is akin to the wild west and the guy with the best aim will win not the guy with the most expensive or most accurate gun. Audio systems should be curated and not engineered.
As a footnote, this is a Vacuum tube based audio forum so right there we are already against 80% of the audio world and 99% of the engineering world. Suddenly reverting back to the ideals of the 99% seems hypocritical to our fellow 1%er brethren.
dave
OK. So, in your opinion after building one and massaging it to bring out the most musicality, what is wrong with the sound of a MosFet emitter follower sitting between a tube gain stage and a LCR phono EQ followed by a tube gain stage?
Or maybe I should ask, where does a MosFet emitter follower rank on you list of ways to drive a 600ohm load but "I [don't] want to listen to for more than 15 minutes."?
First? Last? Doesn't even make your list?
Thanks
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I haven't tried the a mosfet follower to specifically drive a 600Ω LCR but have heard several variations at the input of an autoformer volume control in an attempt to buffer it. The results were similar to a cathode follower in that the cure was worse than the disease. With a source that has a 3KΩ output Z I preferred the sound without the buffer... moving up to a 7KΩ source it was closer to a coin toss and to me the proper solution is to chose a suitable "low Z" source rather than buffering a high Z one. Maybe there is a magical symmetry out there between one specific Mosfet / 600Ω LCR pair and I haven't explored all the options but I do discount general statements in audio suggesting a generic "perfect" solution to a problem. Ultimately it becomes a balancing act between choices of compromise and the choices made are what leads to a "House Sound"
dave
"With a source that has a 3KΩ output Z I preferred the sound without the buffer"
This part is interesting to me. What is the reactance, @20Hz, of the inductance of the autoformer volume control?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
figure 200hy's so 25KΩ. I have had several people use them with tubes like the 26, 27, 56 or 76 and love the results but it is a bit to close to the "edge" for the left side of my brain to comprehend.dave
Edits: 04/30/21
This is the silk volumcontrole.
The measurement is made with a load of 10k. With the more traditional load of 100k the peak is +15dB.
Full review on https://audio-creative.nl/projecten/diy-passieve-voorversterker-met-tvc-deel-1/4/
I agree with you about this type of volume control.
Walter
That isn't necessarily representative of all magnetic volume controls.
This one isn't much better. And 5k output impedance is a very high output impedance.p.s where is the 10Hz-1000Hz part?
Edits: 04/30/21 04/30/21
5K output impedance? where did you get that number from? Did you read the source impedance of 500Ω in the plot title as 5000?
The full details including full bandwidth response against source impedance were in the referenced link titled "The details"
the LF plot from that link is below and aside from the settings with gain, the LF response doesn't change for a given source Z so repeatedly showing the 10-1000hz trace doesn't add any new info. The LF behavior is based on the inductance and once the full winding is used that number does not change. When a part of the winding is used to achieve gain then the sub 1kHz info can then be related to the source impedance.
dave
So I need to select at the best the source otherwise I will get a peak or a low pass?
This is a limit for me.
Walter
Oeps, sorry. Yes, the picture low resolution did make me read 5000 instead of 500
I wonder how it would measure as an AVC?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
The problem with ALL transformer based volume controle is bad coupling.
Excessive ringing. The ringing will be less if you have an outputimpedance which is "high". But a high output impedance dosn't help for a good frequency response either. Tribute for instance use a fluxring at the inside of the (toroid)core but that helps just a little. A snubber wont's work either.
Edits: 04/30/21
"The problem with ALL transformer based volume controle is bad coupling.
Excessive ringing"
I don't see that at all.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Just measure* it and you'll see. Any standard resistance** control does a better job.* measure with a low output impedance ofc and with a system that is not limited to 200kHz or less
**use a 10k to 50k value.
p.s the most measurements i see from manufactures are tuned for optimum results for just 1 setting
Edits: 04/30/21 04/30/21 04/30/21 04/30/21 04/30/21 04/30/21
Can you spell that out? I'm a little dumb.
10K to 50K source impedance?
low output impedance ofc?
Thanks
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I did mean any standard volume controle, a potentiometer, 10k-50k
But a 50k ohm pot will have an output impedance of 12.5k ohms at the -6db point of rotation and that will not drive cable capacitance very well.
The output impedance of a AVC at the -6db position will have an output impedance 1/4 of the source.
I've looked at my AVC using square waves and sine waves and I don't see anything to cause alarm. I've even adjusted the source impedance, up and down, and I still don't see any issue that should cause alarm.
Maybe you can show us the ringing that you think is problematic?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Why would you use a potentiometer to drive a cable directly? That is really silly.
But what about de capacity load for the tvc? Did you measured that too.
Edits: 04/30/21 04/30/21
Don't blame the cathode follower, blame the autoformer volume controle which is far more imperfect then the cathode follower. Secondly blame the guy who thought it was a good idea to use a cathode follower because it enlarge the faults of an autoformer volume controle
Edits: 04/30/21
the subject at hand was a mosfet follower driving a low impedance load and I listed my only experience with them. The load was not difficult and yet they still left me wanting and sounded better without the buffer.
It seems that all of the praise with the followers stems around the need for a near infinite load in the cathode and that infinite load becomes superfluous when you parallel it with a low impedance. The only place I see a CF as a viable option is driving an A2 grid.
dave
I can't undestand why there are a posts around trafo volume control.
Then the use of CF to drive in A2 is a bad solution due the non linear current in a grid of power tube because the impedance falls quickly with the swing (a lot of volts) in the region of positive Vg; just to look on the data sheet.
And it is not a case that on curves of output tubes for Audio the positive region of Vg aren't specified. They stop at 0 volt
In same case with big triodes but there is also the figure of the grid current that is always not linear
In the LCR phono the swing is limited and the correct set point for this job allow a very goodo performances.
Walter
Hi
excuse me but I can't understand what are you sayng
Walter
Just measured a PEC potentiometer 50K with a output source of the amplifier of 500 Ohm. Capacitive load 100pF.
Correction: the 100 pF is wrong, i measured with my 1:10 probe which is 15-17pF
This is the worse setting at 3 o'clock
Measured at 12 o clock setting
And this is measured at the 9 o'clock setting
Looks "bad" but look at the scale, flat within 0,3dB.
Edits: 05/01/21 05/01/21 05/01/21 05/01/21 05/01/21 05/01/21 05/01/21 05/01/21
The problem with measurements and sims for that matter is they only show you what the presenter wants you to see. You were trying to show how perfect a 50K pot was but fed it into an unrealistic 15pf load. Using the 1X probe and its ~100-150pf capacitance is much more realistic of what may happen in the real world. In the plot below a 50Ω source is driving a 50kΩ pot set to -6dB. The green trace is 15pF, blue 30, red 60pf cyan 125pf Violet 250pF and gold 500pF.
Granted this represents the absolute worst case setting scenario (-6dB) and explains why many opt for a 10K pot for attenuation which makes things appear much better and 1K would be nearly perfect but then we are back to the question of how to properly drive a 1K volume control and the answer is obviously a cathode follower loaded with a CCS and we are back on topic. :-)
10K pot set to -6dB into varying capacitances
The approach the the tests are done just to show what the presentr want is out of the reality.
When we test the equipmente on Audioreview magazine these tests follows ( in part) the intl indication then there are other test but are well documented and replicable.
Otherewise the reputation will be lost.
Also when I present on some forum the circuit and measurement.
The REAL test are replicable and there aren't nothing to manipulate.
And, in my opinion, the simulation are the main problem that we can met in forums; there are thousand of "designer" that presents lot of circuit tested only with simulations.
And is is almost impossible to see the proto of that strange things.
Some of them are a strange mix between tubes and sand.
After that I will return on topic
Walter
Nice plots Dave, and you are right about the 15pF for a normal tube circuit although if we use a ecc88 the total capacity is less then 60pF and if it is a cathode follower even a lot less.
but what about the cable capacitance that is added to the Miller of the tube?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 05/03/21
If you know how to design the cable capacity is not a big issue.
If you are using a pot for a passive preamp (and some people do) and you have a cable between it and a power amp, that cable will have at least 20pf per foot. That can be a big issue.
That what I was referring to.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I already posted that that is a stupid thing to do and the plots from Dave shows what actually happens when you add to much capacity to a volumcontrole.
Also transformer coupled volume controle have issues with to much capacity.And if people insists on a direct connection potentiometer to the next amplifier, keep de value of the pot low, LOW + the length of the cable short.
You can calculate (or measure) the hf roll off
Edits: 05/04/21
"Also transformer coupled volume controle have issues with to much capacity."Mine don't.
Edit, I should have said anything circuit or device has issues with too much capacitance.
With the normal capacitance of a tube input plus some cable capacitance my AVC doesn't have issues.Let me ask you a question. Does a signal level step down output transformer have trouble with a normal amount of capacitance? What is a TVC if not a signal level step down output transformer?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 05/04/21 05/04/21 05/04/21
Well, a normal transformer has not the high L value as a inductive volume controle and also the winding capacity is totally different, actually the whole design is different. So why you compare them that way?
The small signal output transformers that I am familiar with in vintage tubed pro studio gear are not gapped like an output transformer for a SE power amplifier. They are used in what one might call a "parafeed" arrangement (no DC in the primary) and I see no reason that they wouldn't have relatively high inductance. Also I see no reason that the winding capacitance would necessarily be high.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Correct and the opposite is the normal line output transformer with low induction and high capacity.
Totally different designs.
But i would be glad to see the measurements so we can see how the transformer in real life behaves with a capacitive load.
Edits: 05/05/21
Please send the test at - 6dB from the max position of volume.
Which brand is the pot?
Walter
Edits: 05/01/21
Btw, the test was not done with 100pF load but with my 1:10 probe which is about 15- 18pF. So the 3 o'clock setting is about what you asked for (-4dB, almost -6dB )Walge, the capacitance load of the potentiometer largely determines the frequency response. So the capacitance load of the next stage is very important.
The great advantage of a normal potentiometer is the absence of all kinds of resonances at higher frequencies. At the same time, the impedance of a normal potentiometer is still high, that's is why professional equipment with tubes usually use lower values and also compensation circuits.
Edits: 05/01/21 05/01/21 05/02/21
Hi
I know well the problem and, when possible I use a low value of pot.
In the last numbero of Audioreview magazine there is a presentaion of a potential kit of pot with 6 relè and a (almost) costant impedance ( done by a collaborator).
Will be the possibility to change the value desired.
Walter
What is there not to understand? It has been my contention that benevolently driving a 600Ω load is not an easy thing to do independent of it being resistive or reactive. Since this is about 600Ω LCR units, Tre asked about mosfet followers. Since I hadn't tried the exact combo he asked about but have experience with a similar situation of using a buffer to drive a difficult load I offered that example. As for the grid of an A2 tube, that is also a difficult load but unlike the benign flat impedance of the 600Ω LCR, it varies substantially. The key point here is with an A2 tube, you are stuck with the load the grid provides and if you choose A2 you must find a way to deal with it. With a 600Ω LCR, within limits you have the option of changing that value to something that is easier to drive without compromise. Thats my opinion and I'm sticking to it :-)
dave
I still think a lot people use not so good designs. It dosn't have to be sofisticated but it should be good designed, designed for the whole audio system, not just one stage.
Maybe some people here should read a little bit more in the books about cathodefollowers.
Here is a start.
http://valveradio.net/audio/low-distortion-cathode-follower.html.And this is for the Brosky fans
https://www.tubecad.com/2011/08/blog0212.htm
Edits: 04/30/21
Hi
personally, as said, I don't like to mix sand with tubes.
Walter
Hi
I always done the RC phone like this
It is a Junior phono, the LCR stage took something from this project.
The use of Tango style LCR is a challenge for me and the results are very fine.
The inductors from Sowter are very fine as the test said and aren't so expensive.
The use of CF with the right triode when properly set is fine for this job.
We aren't speaking about great swing; the goal is to get right results with the test lab
For the LCR stage I made around 200 test lab included the protos with original Tango stuff
I spoke only with test lab not simulation what I always see about LCR stages.
In addition, as said, I test the project also on the way with the help form my firends; they give me the indication about sonic performances.
When there is a convergency I got the results expected.
About the 4th time costant I think it is a shadow.
In next future with a help of my collegue I will show a LCR stage with several different curves about main historic label.
It takes some time.
Walter
What drive impedance does the LCR module require?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
The Pultec type LCR is a 600Ω network that either needs to be driven from 600Ω or terminated with 600Ω to have the desired response. Since it is a type of T network, it has a nearly constant 600 impedance independent of the source and load impedances. If the unit is terminated by 600Ω, that appears in parallel with the 600Ω network Z for a 300Ω output impedance. If a 600Ω source is used this presents the situation where Rsource=Rload and a 6dB loss of signal occurs.
dave
If we look at impedance only with a source of 600 ohm and closed with same Z we got the real costance.
As my post where I put the freq answer with diffferent Zsource now I put the resposùnse with 600 as Zsource and different Zload, the lowest trace is 600 ohm Zload, nominal for Tango spec
The response aren't so bad but not as expected.
I have used my two Audio precision for test; the enphasis RIAA curve test are made from AP software, in attach the table
measured results in the frequency domain don't tell the whole story and they can easily be adjusted. The issue I see is the 600Ω impedance and driving it with a sonically benevolent circuit. The options are limited to cathode followers and high transconductance tubes loaded with step down transformers if you want to remain true to the forum title. The step downs have frequency domain issues and reduce the precious gain and I do not find cathode followers to be the panacea that many make them to be. Don't get me started on the silliness of the 600Ω value for LCR's in general ;-)
dave
The frequency response is the first parametere to look just to be compliant with RIAA specs.
Then other test must be done.
Normally it is very interesting for me to discuss around circuit but the consideration over the circuit and test lab must be done, if possible, with the same level of informations.
Not only with concept or simulations.
The use of CF in a circuit like the one I show is fine for different ways.
Under the sonic performance, when well configured with a proper tube it give you a very good results, technical and sonic
The concept of Tango and Pultec ( if we want) is confirmed in this circuit with very good results.
This project is intend to offer soon a kit that is easy to build, no very expensive, with certified results that are not comes from simulations.
In addition, the next step will be a full trafo LCR stage, already done as proto with original Tango stuff.
But will be a personal stuff, always presented on magazine
Some test results, published
This is the freq response
This is the FFT
This is the THD vs frequency
The ratio s/n weighted is around -76/-78 dB, good results.
The max acceptance ad 1 kHz is around 80 mV and the gain is 40 dB (+/- 2dB, due the selection of tubes). This can be increase using a lower Rp tube in CF
The very interesting thing is the use of other tubes (with little chnges) in a gain stage as ECC83, 5965, 7062 ( good results), and the use on CF of 6N6 and 6H30 or 12 BH7 or ECC99; there is the possibility to change the paddle of filemants.
I am also lucky because each circuit I present on Audioreview magazine is tested by numerous friend and ex-customer to get the impressions, at the end is published. One of them is owner of 105 MC cartridge.
With different chain, of course.
This proto was tested deeply from my collegue who wrote the listening test for the magazine, it has two high level phono stage, LCR, it is a well known brand.
The front end was Kuzhma Reference2 , arm 4 Point , head is a Lyra Titan i and Lamm as electronics.
So the level is high. The review was fine.
Walter
I seem to recall that someone made a 10Kohm version of the filter - I don't know if it's still available.
At 600 ohms, there are a few triodes capable of a 600 ohm plate resistance, if you wish to avoid cathode followers and other forms of feedback.
But overall, yeah, this is more an audiophile-adjacent approach, nostalgia for the pro audio gear of 70 years ago. Nothing wrong with that but not the same thing.
My two cents.
Hey Paul,
Both S&B and Silk made/make 10K modules and while the concept of high impedance is a good one. the use of a 10K number is misguided since it requires obscure cap values. Since the chokes will need to be made custom for any impedance, why not pick something that makes the caps easy values to obtain rather than values that need to be pieced together to adhere to some archaic number? The 600Ω impedance units require 5.04µ and 0.126µ. Simply bumping that number up to 720Ω takes the values to 4µ and 0.1µ much easier values to source. The inductors can be made tapped making the filter easy to trim and a single high quality cap used for each position. I ended up choosing a value of around 7500Ω and the caps were 0.33µ and 8200p which are easy to source values.
As to the low Z tubes driving the 600Ω network, I stared at paralleling the halves of a WE 421A (5998) which would give aµ of 6 and an rP of 150Ω but I discarded that idea since it puts you in the same realm as say a 417A into a 4:1. The 6HV5 / WE 416A were another thought to net a gain of 40-50 and a low enough output impedance to drive 600Î through a step-down but they come with their own baggage and ultimately found that the 6C45/437/triode wired 7788 into a 7.5K LCR gave the best results.
dave
In attach the last of four protos I made for tests before this one.
It is full Tango
MCT-999
EQ-2L
NP-206 x 3 ( and 11103 x 2)+ other for test
Good performance and very good sound.
I am preparing the similar with Sowter stuff, for my personal fun.
Walter
Hi
In attach some test lab ( not simulation!)n about a LCR stage driven with different Zsource but it closed to 600 ohm ( as Tango spec)
It is a freq. response.
As you see there is a good answer fir dirrerent source but the level is quite different.
In my circuit the Zsource with CF of 88 is around 100 ohm.
The reason is that with Tango spec the attenuation at 1 kHz will be 26 dB, too much. With 100 ohm we have 4-5 dB more and it will be fine.
In the other hand the Zin is not costant until 200 Hz ( less or more) but is not a problem for our purpose.
If the Zsource is 600 ohm, we will have a costant impedance of 300 ohm for 20 to 20kHz.
This is the response of original Tango EQ-2L with all the components selected, it is fine
Thids is the answer of the LCR with Sowter nd trimmed for best results.
The result is better than the Tango specs, +/- 0,2dB
Thids is the diagram of my LCR phono
The parallel on LCR circuit is to obtain the perfect value as Tango sepcs.
The pcbs and Sowter inductors will be rady in few days , I would like to selle them.
Walter
Do you sell here, Walter?
Edits: 04/25/21
No
I am planning to do something in next two weeks
I am waitng the inductors from Sowter and the new pcb
Walter
It looks great. What tubes ?
idlers and tubes....life is good
Hi
Thx
The tube are ecc81 as gain stage and ecc88 ad C F
But is possible to use other changing the pad for filamenti
Now the gain is 40 db
I am planning to offer the pcb
Walter
Oh yes you stated what tubes, sorry.
idlers and tubes....life is good
Bravo! Nicely done. Can you post a link to your article?
Thx
Not now because it is on magazine that is selling.
I will send some other info soon
Walter
This is the last version pf pcb, just little changes from proto
Walter
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: