|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
205.188.117.12
I have heard many posters saying that
lower power operating points sound better
than the "max" operating points.
I was wondering if that had to do with:
1) lower B+ - keeping away from max voltage
2) lower current - lower max dissipation
3) higher Z OPT - flatter load line
Or some combination of the above ?
I am thinking of the 300B for me.
But I have heard that about other tubes as well.
I have not heard this discussed before
and always wondered about it.
Follow Ups:
Right around 250 Vp and 34 ma Ip the distortion spectra of my little 6cw5 SE starts at low levels with relatively high 2nd and 3rd and 4th just above the noise floor, some 85 dB+ down. As power output rises 3rd comes up with 4th close behind until a point is reached at which 3rd starts dropping and only the even harmonics rise. It maintains that balance to near clipping. The biasing of this tube meets neither of the criteria discussed and was chosen specifically for the behaviour demonstrated by the 3rd harmonic. It sounds wonderful and is my fave build of about a dozen so far.In general though it's almost universal in my experience that high and cold results in significantly more benign 2nd harmonic but much less of the far more objectionable higher harmonics above approximately the 4th.
Thermionically addicted.
Ironically I am going to use the EL86 as the
CCS for my 6H30Pi/12B4 DRD drive for my 300B.
In pentode of course.More 2nd and less upper harmonics sounds good to me.
Try for the 300B:450V/50ma/5k
&
350V/80ma/5k
I know this is work, but it isn't the total end for the world:
Build for 450V, tap the PS to 350V, same end choke & caps.then create the two biases the same way: either different resistors, or tap on bias PS.
You are a major scientist with major audio research projects, and major publishing prospects with an audience waiting with major baited breath. So.....
You could also have a lot of fun finding out.
-3dbI was thinking of doing that same thing.
Just got done modelling it on PSUDII and
I need to use a GZ34 instead of the 274B.
Such is life 8^(
Hi Dan,If you do those two......try 350 VDC at 50 mA too.
What do we think the extra 30 mills does to the SE output trannie, of unbalanced current?? :-)
Dan,
"Running the tubes low" might also be referring to the practice of running them with a lower plate voltage at higher current, putting them into "Deep Class A" - the idea being that keeping the quiescant operating point more (than not) to the upper left of the curve chart (ie. lower volts at a higher current) will keep the tube from going into low current at the high end of the voltage swing where the curves start to lie down - meaning non-linear operation.If you do searches on "Deep Class A" you'll turn up lots of stuff. Paul Joppa's work on tube operating points is also very good to read. The link is to one of Paul's posts that I thought particularly interesting because of his remarks at the bottom about Deep Class A operation and its implications with respect to the output transformer,
'hope this is useful
> > I was wondering if that had to do with:
1) lower B+ - keeping away from max voltage
2) lower current - lower max dissipation
3) higher Z OPT - flatter load line
< <
None of the above. A lower operating poing is less linear. This can be either good or bad depending on where it ends up. If you are delivering it to a 16 ohm 115 dB horn, then you don't need 1000VDC plate curves working to linearize the sound. It's linear sounding at 150VDC.It will "warm up" at the lower operating voltages a tad. If this is what is needed, then this is what you want to pursue. If the sound is still too warm, then go up in voltage and linearity.
I operate 45's at 180 VDC and 15 mA. This is below it's max rating, but I'm running them into 112 dB horns from 250 Hz on up. It doesn't need max power, it needs a nice easy operating point that's still linear to the speaker and won't burn out the 45's quickly. So I like it at that operating point. Plus it's easier to get those voltages and currents. Any more voltage and current and it starts to "gray out" the sound, meaning the harmonics are washing away, something not good to the SET amp crowd.
This is what DIY is all about - tailoring the sound of the amp to the speakers, a good distance ahead of "off the rack" amps and speakers.
I think that you want to stay on the load lines, specifically in the flat sections. Extra low or high voltages can get you into the non-linear portions of the curves. Why put up with the distortion?Current: My experience tells me to never starve my tubes. That doesn't mean I run them all at maxed current but 70-80% of max current is better than low current. Just because you can run it low current does not mean it will still sound good and extend the life of the tubes. You don't want the so-called "perfect sound forever". Low current in the heater and in the B+ should raise distortion and the Radio handbook goes over that. No bargains in good hi-fi. There are those that disagree with me and I have heard a number of amps, preamps etc. that run low current, I think that they are inferior to rich-current equivalents.
RichardOK 70-80% max rated current.
What about voltage still 70-80%?
On the 300B:
400V/50mA/5K
VS
450V/60mA/5K
I forgot to ask a question. How much power do you want? The thing about less voltage and less current is you get low Wattage out. I was thinking of drivers where I go for max swing and about 80% of the heater current capacity of the tube. For the 300B, I would go for the linear portion of the curve at 75-80mA, where I think you can do 450V ( which gives 34-36W of dissipation on the plate, which is safe and gets the max power. just looked at the curves, whoa! what a linear tube!) You can always drop down, but design to drive the worst speaker load. Also, if you use the Russian 300B, which is the same as their 2A3 evidently, it is a current pig and low current gives lean bass. However, the tubes will not last as long as they would if you dropped the current and ran them at 450V and 60mA(which makes sense for the expensive 300B's).
http://datasheets.electron-tube.net/sheets/163/3/300B.pdf
I am running 93dB speakers and I don't play loud.
I would still like at least 8 watts though.
I like to pick the loadline OP point that
splits the zero bias and zero current point
so that when it goes to crap it really does.
It also gives maximum output before distortion.
400V would give me 10W and 450V would net 8W.
I haven't decided on which 300B to use.
I am leaning toward the VA 300B titanium.
Dan, you'll get all sorts of opinions but ultimately you know the drift, you'll have to make your own mind up. FWIW, I personally think that many differences of opinion arise simply because there really is no absolute rule that can be applied to every case.There’s no doubt that in most triodes generally higher current results in more linear operation. Also, higher voltages result in higher linear voltage swing, but this is not always required. Many factors need to be taken into account and a good starting point is always to plot the load line and set the operating points along the most linear portion of the line having regard to the required voltage swing.
However … when it comes to OP stages or any stage where a lower than optimal impedance load must be used then the previous stage becomes way more important and IMHO the pair of stages should be viewed as one and the operating points should be set according to the best overall result. For SETs, that generally means that lower operating current for the driver though less linear can often be used to offset the higher expected distortion of the OP stage due to opposite phase cancellation. This is both measurable and very audible IME. Since distortion in the OP stage is usually dominant higher current normally works better here as does higher load impedance. I also think that this is the reason that so many report such good results with (less linear) Pentode drivers for SETs.
Cheers,
Naz
NazI am thinking of the more relaxed vs
strained sound some declare at lower OPs.
I was thinking that may be the reason that I
prefer the KT88 over other tubes in my SE amp.
At 22W, the KT88 is just cruising while
the KT77/EL34/KT66 are near maxing out.
Since I seem to prefer that sound
and my next amp will be a 300B,
trying to bring what I know and
applying it to my next project.
So I guess I am asking what strains
the tube sonically more -
Dissipation or Voltage or Low Z OPT ?
How about a 300B at 450V/60mA/5K OPT?
To be specific.
Dan,Since you asked: I've confirmed on Greg's 245 SEer that dissipation and high voltage both are culprits produing a "strained sound" in the SE output stage, and it effects the SE output trannie too !! A low Z trannie is less of a sonic problem IMHO.
See my prior posts by Dennis Fraker from Serious Stereo from 2006 on, re: 5/8ths throttle, etc., do a forum search. `
You may just wanna stay with the KT-88 Dan.
I'd personally optimize the amp's supply and lay out before more tubes were swapped. SEers are critical as to lay out and parts choices.
JeffThis is for a completely different system.
My living room has the Monica 2, 12B4 Preamp
and the SE KT88 with Rythmic Plate subs
with a phono hopefully coming soon.
My kitchen has CAL Alpha/Delta CD sorce,
Marantz SR-66 and DIY Audax TL speakers with
Dynaudio D260 tweeters & GE 97F MR caps.
The Marantz was just given to me and
I have to say it has really good sound for a SS,
hey even some tube systems wouldn't compete.
Which brings me back to the 300B amp.
I am looking to replace/upgrade my Jolida 202
because it is noticeably the worst of the pack
and it is in my bedroom/primary listening room.
I have Acoustic Energy Threes in there - 8ohm/93dB
and Sonic Frontiers SFT-1HK and SFD-1MKII source.
It used to be the best system that I had.
BUT I guess things change 8^D
That is what all of those load lines are about. At different voltages current is swept from zero to past the maximum. What you are looking for is a flat section in the load line that is under the plate watt limit. There is usually one combination which will yield both the highest output and the most linear response. That is usually the "recommended" condition. You can run the tube with less voltage and increase the current, you will get a slightly different sound and usually less power. Fender liked to run the tubes at very high voltages and lower current for the tone.
_______________________________
Long Live Dr.Gizmo
Chris,have you ever put any of the "new" tubes ( Russian, Eastern Europe, Chinese, etc) on a curve tracer and compared to published curves of the equivalent n.o.s. tubes?
AndyIf you want to see real tube curves
for russiam tubes go here:
http://www.klausmobile.narod.ru/testerfiles/index.htmDanL
The curves for the tubes I've measured mirror my own findings. Note how much better tubes work at higher current ... higher gm, lower rp, greater linearity. Note also how important a more horizontal load line is for some tubes, especially if high voltage swing is required. But all of this has to be taken in the context of the load presented by the next stage / OP TX etc and the drive of the preceding stage. I think this is where some draw, not wrong conclusions but incomplete ones when they describe the sound of a particular tube in comparison to another or at particular operating points.
NazYes they look so much better at higher currents.
But there is a point they start sounding strained.
I was just wondering whether the current and
thereby the power dissipation causes the strain
OR the higher voltages causes it.
Maybe the higher impedance OPT used on the lower
powered amps for a flatter load line causes
the more relaxed sound.
Jeff says the high voltage and power dissipation
are the source of the strain.
Most things in life function better
when not pushed to the limit.
So I think I will try a 300B
at 400V/50mA/5K load.
I think everything is a variable so you have to suck it and see how it works in your system.IME higher current has always worked better in SE OP stages of any kind including 300B but I always stay well within the limits. The sound is fuller than at higher current and perhaps it's just that the thinner sound of lower current is providing a better balance with the rest of the system which may be perceived as producing greater detail.
In my book higher HT voltages generally provide greater opportunity to design a more linear stage, especially where the plate load is a simple R. However there is a limit in OP stages with an OP TX, which are another thing again. While a higher impedance will always be more linear the trade-off may come in the form of additional leakage C caused by more windings and resulting in lower BW. It costs considerably more to alleviate this problem.
I have never found less linear operating points producing greater distortion to sound better UNLESS the particular stage is cancelling some distortion up or downstream. It's the total that counts. And, I'm not sure what the definition of "strained" really means. Is it describing noticeably higher distortion or merely excessive weight?
So back to your suggested operating points ... 400 - 450V is fine for a 300B but personally, I'd favour 60 - 70 mA, the higher the voltage the lower the current. Distortion jumps appreciably at currents that are too low and to my mind 50mA is bordering that point. If you really want to idle the tube at low power consider lower plate voltage with higher current, it's more linear.
NazI agree that flatter loadlines are preferred.
I like my 5K OPT on a KT88 vs a 3K OPT.
On driver stages I like really high
henry chokes with low rp tubes.
I have a MQ EX0-99 150H on my 12B4
and a MQ EXO-01 100H on my 6J5.So you are saying my original OP
of 450V/60mA/5K is better?
I stress that this is only my opinion but I'd favour something closer to 450V / 70mA than lower but if you feel that 30W dissipation is too high for you then keep the current up and drop the voltage rather than the other way around. Also, 5K impedance is great for linearity (and still OK for good BW) if you 8W is all you need.Now here's something that is not intuitive, you may achieve better results with lower current in the driver stage, even more so than varying the current in the OP stage IMHO. Worth playing with anyway and I'd be interested in whether your findings approximate mine.
NazIf I thought I could get away with lower power
I would definitely try the "deep class A" route.
With 93dB efficient speakers, I need 8-10 watts.
At 450V/60mA/5K, I can get 10 watts.
At 400V/50mA/5K, I can get 8 watts.
| ||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: