Tube DIY Asylum

Do It Yourself (DIY) paradise for tube and SET project builders.

Return to Tube DIY Asylum

Message Sort: Post Order or Asylum Reverse Threaded

Here I go again ...

205.188.116.16
Posted on April 12, 2010 at 21:49:43
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002

Tryimg to get everything from an amp.
It has elements that will please everybody.
TWO stage for simpicity and sonics.
Critical H in the first choke with lowish DCR
and a low DCR 1H on the second choke.
VR tube regulation for the first stage.
Direct coupling for cleaner sound.
If the driver tube dies then bias only goes up 5V.
Mega capacitance in the PS for solid bass.
Ultra-path cap to eliminate the negative of the above.
Time delay on the B+ for the 300Bxls longevity.
Enough driver swing to max out the output tube plus 28%.
I am thinking a seperate PS chassis because of size.
Fairly cheap except for the OPT and 300Bxls of course.
It has some elements for everybody but there will be people
that will think any element is not enough to satisfy them.
I think it just might be better than most in all areas.
Let the flames begin ...

DanL

PS I probably will need a grid stopper
on the driver tube though.



 

Hide full thread outline!
    ...
RE: Here I go again .....nice amp I was thinking of building, posted on April 12, 2010 at 22:53:49
Minussss3db
Manufacturer

Posts: 586
Location: so cal
Joined: January 28, 2010
 this same basic amp a few months ago...

you are going for high capacitance on the output...and then Ultrapth

so it will have nice bass...dc..nice bass, ...battery bias...nice bass

etc...could you tolerate a Edcor OPT for the right price...$83 I think???


Also do you want a .1uf Cap (high quality) across the Od3???

In generally this is my next amp, except I may have to use a Direct Heated Pentode to drive it instead....

Nice drawing too DanL...thanks

Have Fun,
seriously...
-3db

 

RE: Here I go again .....nice amp I was thinking of building, posted on April 12, 2010 at 23:06:39
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 -3db

I never think it is a good idea
to skinp on the OPT IMHO.

The .1uFd cap would be a good idea
if it was resistive loaded but since
the choke has such a high PSRR
I doubt you will get much
improvement with a bypass cap.

DanL



 

Needs more work !! ..........., posted on April 12, 2010 at 23:48:24
drlowmu
Manufacturer

Posts: 9730
Location: East of Kansas City
Joined: January 10, 2005



Dan,

I think your design needs much more thought applied to it.

You are running about 490 VDC across the tube, at 100 mA., which is 49 watts dissipation. The expensive (lovely) tube would be happier and last longer, and sound better at less of a dissipation, assuming a good design.

Your power supply needs to be redone.

The 300B XLS is a low plate resistance tube, and you are driving it with a supply that is as slow as slug, and as high in Z, as I could imagine in my worst nightmare. The Z of your supply is 1,353 Ohms, ( drops over 20 VDC with a 15 mA. current change) how well can that drive a sub 1,000 Ohm Finals tube??? Also, your supply takes 3,000 msec. to recover from a step, whereas fellow Forum member John Swenson is measuring and claiming 50 msec. is necessary for a good sounding supply.

I had to do an estimate as to Hammond values, but I show what it looks like in a PSUD 2 simulated step test - with a 15% current change.

I have been working for two years, on a similar design (for my AV-32B DC amp), and have come up with MUCH lower supply Zs, and step recovery times. It can be done!!

The 5R4Gs, even doubled up, are too high in Z, use two 5V3-As. 420 uF as a Final C will sound like MUD, will NEVER do justice to a piano, or drum solo, certainly not from a high Z, 3,000 milisecond settling supply.

On the positive side, I like your choice of driver and finals tube. Cheers.

Jeff Medwin

 

RE: Needs more work !! ..........., posted on April 13, 2010 at 00:33:15
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Jeff

The power supply will sag much less with this design.
Sure it will recover slower but changes much less.
At 20Hz the B+ will drop 5V before the waveform
reverses the slope and the PS recovers.
And that is the total worst case scenario.
I want minimal music signal on the PS.
It is the complete opposite of your methodology where you
tune the PS to the system I want the PS to be seperate.
There is a typo the 1st choke is a 193NP 3H 36 DCR and
the 2nd is 1H 5.75 DCR with gives a nice slow ramp.
With a total of about 40 ohms it will drop very little.
Making the PS less part of the music and more a DC supply.
The ultrapath cap will provide the speed needed for the tube.
The 2.2uFd is passing all the signal from 20Hz and above
so that IS the cap the output tube will see.
That is the beauty of the ultrapath caps.

Also the extra large cathode bypass cap to minimalize
any influence that may have on the bass.
I may double it to make sure.

The tube is rated at about 70-75 watt tube so
at 49 watts it is at about 70% max which is good.
From spec sheet >> Power Output in Class A - 23 Watt
That should have a nice safety buffer for nice sound.

DanL



 

RE: Here I go again .....nice amp I was thinking of building, posted on April 13, 2010 at 01:28:55
reVintage
Audiophile

Posts: 271
Location: Eskilstuna
Joined: June 19, 2008
 Hey DanL,

Did a quick check on the 6GK5 GE datasheet, nice tube.

I am a little puzzled as my calculations says clipping at 65Vrms with ca 7%THD. If 300B wants 90Vptp there is no margin at all. But maybe you have checked the tube IRL.

Tried some other working-points and the tube seems to have its best linearity well below Ug=-1V at a little higher Ia.

How it sounds I have no clue, though ;-).
Brgds
Lars

 

Dan, can you explain the thinking..., posted on April 13, 2010 at 02:52:17
Allen Wright
Manufacturer

Posts: 5652
Location: Schaffhausen
Joined: January 31, 2002
 ...behind the 100:1 capacitance divider on the "Ulrapath " connection.

I cannot see how it can provide much, if any, "ultrapath effect" when you have 220uF to gnd.

Regards, Allen

 

RE: Here I go again ..., posted on April 13, 2010 at 05:53:17
Al Noakes
 The ultrapath cap on the 300B is redundant . The shunt reg and dropper can be removed . The driver stage B+ can then be derived from the filament centre tap of the 300B . Grid stoppers on the 6GK5 may be required . Hideous B+ , this sort of design would be more suitable for 2A3 . Have you considered IT coupling instead of choke loading otherwise ?

Al

 

RE: Here I go again .....nice amp I was thinking of building, posted on April 13, 2010 at 06:05:02
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Lars

With a gain of 80 and a bias of 1.6V,
Choke loading the plate and direct coupling,
it should swing 140Vpk more than enough
for the 100V bias of teh 300Bxls.
I drew the loadlinex and I like
this operating points linearity.

DanL



 

RE: Dan, can you explain the thinking..., posted on April 13, 2010 at 06:10:19
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Allen

The main purpose for the ultrapath cap is
to eliminate the PS cap from the signal loop.
You can use the ratio of cap sizes to have
hum cancellation but at 35mV of ripple
I don't have to do that.
Instead I chose to optimize bass.

DanL



 

Dan, I know why one would use an "Ultrapath" cap..., posted on April 13, 2010 at 06:28:55
Allen Wright
Manufacturer

Posts: 5652
Location: Schaffhausen
Joined: January 31, 2002
 ...but youhaven't answered my question.

Regards, Allen

 

RE: Here I go again .....nice amp I was thinking of building, posted on April 13, 2010 at 06:31:27
reVintage
Audiophile

Posts: 271
Location: Eskilstuna
Joined: June 19, 2008
 Hey Dan,
Checked the GE-curves again and 145V, 10mA and -1,6V is an impossible combination. But if you use 145V, 10mA, -1,4V you get the figures I gave you.
So if you want 10mA, -1,6V you must go for 175V. This point is really good and should give you (noload/choke/CCS) clipping at 85Vrms, 2% THD and a gain of 75x.

On the other hand 6,5mA/145V/-1,6V will also work indicating just below 3% at 81Vrms and 71x gain and is probably the one you went for.
Brgds
Lars

 

RE: Dan, I know why one would use an "Ultrapath" cap..., posted on April 13, 2010 at 07:26:27
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Allen

The 100:1 ratio is irrelevant as I don't need
the ripple cancelation properties you speak of.
The 220uFd is the largest value foil I could find.
The 2.25uFd was chosen for 20Hz and above operation.

DanL



 

RE: Here I go again ..., posted on April 13, 2010 at 07:35:07
Stephen R
Audiophile

Posts: 1428
Joined: January 11, 2002
 Should you take the B+ to the other 6GK5 stage off 0D3 pin 7?

cheers,

Stephen

 

Dont' forget a cap where you raise the 6GK5 heater to 50 volts. , posted on April 13, 2010 at 07:59:36
Chris O
 Without a cap there, it will hum like mad.

Also, 20ma through a 15K resistor will drop 300 volts. Using an 0D3 to drop another 300 volts will not work very good. You need to up that 15K resistor to something closer to 28K.

Use one (or two or three) of your GE motor runs as your ultrapath cap in parallel with the Mundorf and kill the Solen to ground. Don't add new caps, move the ones in the B+.

 

RE: Dont' forget a cap where you raise the 6GK5 heater to 50 volts. , posted on April 13, 2010 at 09:33:07
Michael Koster
Manufacturer

Posts: 841
Location: Eureka, CA
Joined: October 24, 2007
 I read total current thru the 15K resistor as 40mA:

10mA channel A driver
10mA channel B driver
20mA idle in VR tube

The problem with this is while the driver heaters are cold, the VR tube will have to pass 40mA. I guess an octal base ST will do it, but why not reduce the VR tube current to ca. 10mA operating, 30mA during warmup?

Cheers,

Michael

Ps I guess your B+ delay will keep the VR tube from seeing 40mA normally. You would need to pull both drivers from the sockets or have an open heater circuit to get 40mA on the VR tube. Still I would reduce the current a little.

 

Don't forget to ground the secondary, posted on April 13, 2010 at 09:42:00
kurt s


 
If this isolated secondary has a speaker wire and a speaker attached, there's much chance for electrostatic build-up and crackling noises, especially in dry climate. This secondary is floating and can build charges of tens of thousands of volts that can arc over to the primary and you then hear crackling discharge. Or at least a bleeder resistor to ground on one side of that secondary.

-Kurt

 

Why loose the Solen?, posted on April 13, 2010 at 10:03:07
JoshK
Audiophile

Posts: 820
Location: NJ/NYC
Joined: August 3, 2001
 I must ask why you suggest loosing the solen?

http://www.tubecad.com/2008/08/blog0147.htm

Although his 100:1 ratio isn't ideal for ripple rejection, loosing it altogether seems to be going in the wrong direction.




----------------
"When Khruschev said "we will bury you" I don't think he meant with surplus parts." zacster

 

RE: Why loose the Solen?, posted on April 13, 2010 at 10:15:17
JoshK
Audiophile

Posts: 820
Location: NJ/NYC
Joined: August 3, 2001
 More references:

http://www.tubecad.com/april99/page2.html
http://www.tubecad.com/april_may2001/page41.html (right column)
----------------
"When Khruschev said "we will bury you" I don't think he meant with surplus parts." zacster

 

RE: Needs more work !! ..........., posted on April 13, 2010 at 11:43:05
drlowmu
Manufacturer

Posts: 9730
Location: East of Kansas City
Joined: January 10, 2005






Dan,

You say "The power supply will sag much less with this design."

YES, much less than your LAST design. A partial truth. But it needs more thinking applied to it.

A 15% step (increase) in current, and your B+ drops 20 VDC, over 3 seconds - despite having hundreds of uF as the final cap. In Jeff Davison's new amp, the same 15% drops only 6.3 VDC in 300-350 msec., and in my supply, I drop about 3 VDC in 40 msec, 75 times quicker in recovery than yours. The entire power supply in certainly INCLUDED, not excluded, in a SE amp.

Henry Pasternak once suggested "the ideal is an instantly settling supply", and I agree with him in principle. It should ideally be "a fraction of one millisecond", which we can't build.

The filters in your new supply have a Z of over 1300 Ohms, HOW can that high Z drive the Finals tube well??

Your amp's supply will have a LARGE drop in B+, when a transient hits (think of a struck piano note) and, when the next piano note is struck, your supply is at a LOWER B+ level, and not recovering, and this cascades into an amplifier whose supply is constantly getting in the WAY of the music, BIG TIME. You hear the music in the time frame of the mis-designed amplifier power supply, NOT in the time frame of the music. The amp's supply gets in the way of the music.

Jeff Davidson built a faster recovery supply for his SE 320B-XLS amp, (top most simulation) I think with Mike Koster's help. Look at it's approximate simulation, on the same scale as your supply. The first choke is sub-critical L, a Lundahl, the second is a Triad C-40X. It recovers in 300-350 msec, not 3,000 msec. It has 477 Ohms of Z to DRIVE the big Finals tube. Jeff tells me privately his amp sounds "incredible".

Also, I have a design I am working on, two 5V3-As, (bottom simulation) just two chokes, that settles in 40 msec, and has a 204 Ohm Z.

Do you think a 204 Ohms Z filter will supply the low-plate-resistance 300B-XLS quite a bit better than a 1300 plus Ohms Z filter ????????

Jeff Medwin

 

If you have enough uF in the ultrapath cap, you do not "need" another one to ground., posted on April 13, 2010 at 11:43:11
Chris O
 You use both an ultrapath and a cap to ground when you "can't" get a big enough, good sounding cap to fill in the ultrapath slot.

 

RE: If you have enough uF in the ultrapath cap, you do not "need" another one to ground., posted on April 13, 2010 at 11:50:29
JoshK
Audiophile

Posts: 820
Location: NJ/NYC
Joined: August 3, 2001
 I am not trying to be argumentative, but that statement without explanation of why doesn't answer the question. If you read the link, doing so will inject in PS noise into the cathode, which gets multiplied by the mu of the tube and superimposed on the output. Essentially you are amplifying the power supply noise to the output. Explain how suddenly making the cap larger circumvents this issue.
----------------
"When Khruschev said "we will bury you" I don't think he meant with surplus parts." zacster

 

RE: Needs more work !! ..........., posted on April 13, 2010 at 11:58:40
Caucasian Blackplate
Industry Professional

Posts: 8313
Location: Seattle
Joined: June 18, 2004
 That supply has a ton of ripple, nearly 500mv?

 

A couple of thoughts of the 6GK5 operating points..., posted on April 13, 2010 at 11:58:49
kellymon
Audiophile

Posts: 596
Location: N. California
Joined: September 30, 2005
 DanL,
On the 6GK5 operating points, I notice you are running 1.6v bias, which is getting kinda high and into an area of the plate curves that looks pretty non-linear, I've used a NiMH @ 1.3v or RC at 1.1v and been pretty happy with that.
Also, your ni-cad voltage is 1.5v and your bias is 1.6v which means only 5ma across your 20 ohm RK. Is that correct? Because you are showing 10ma higher up on the plate.... And again, I have found the 6GK5 better at more than 5ma. Especially at higher bias v.

You have helped me several times as I learn this stuff and I am in no way critisizing anything here. But since EliD suggested the 6GK5 to me a while back, I've run it though a lot of different scenarios (still continue to)and I'm interested in your choices of these operating points.
Maybe the more horizontal loadline presented by the direct coupling sounds better to you at these points?

I'd like to say thanks for posting this, it looks like a cool project:)
robert

 

RE: Dont' forget a cap where you raise the 6GK5 heater to 50 volts. , posted on April 13, 2010 at 12:03:18
Caucasian Blackplate
Industry Professional

Posts: 8313
Location: Seattle
Joined: June 18, 2004
 An 0D3 is rated to pass up to 100ma during warmup, as long as it is operated within its limits for at least several minutes thereafter.

 

Exactly!, posted on April 13, 2010 at 12:09:20
Chris O

The ultrapath cap is sized to inject a specific amount of out-of-phase noise from the B+ into the cathode to cancel out the in phase noise on the plate. This is similar to how a push-pull amp can cancel out the heater hum and be dead quiet.

Jack Elliano popularized the "Ultrapath" connection for the cathode cap. Over the years has has increased the uF value of the cap quite a bit. If you have enough uF in the ultrapath it fufills the cathode cap roll. I use around 100uf and it sounds good (to me).

 

RE: Here I go again ..., posted on April 13, 2010 at 12:21:40
Caucasian Blackplate
Industry Professional

Posts: 8313
Location: Seattle
Joined: June 18, 2004
 I like it Dan! It's not quite the way I would build it (obviously), but still quite nice and a good reflection of your design priorities. I would, however, wonder about AC at the plate of the 0D3. For the extra 10-12 bucks it would cost, using an 0D3/dropping resistor for each channel might offer better channel separation. Certainly a bypass on the 0D3 wouldn't hurt in that regard either.

If you haven't bought the motor runs yet, you could save some space with these, and the price might be similar to that of six high voltage motor runs...

 

RE: Needs more work !! ..........., posted on April 13, 2010 at 13:43:57
drlowmu
Manufacturer

Posts: 9730
Location: East of Kansas City
Joined: January 10, 2005
 Correct, 400 or 440 mVAC as I recall. NOT a problem with a low mu Finals tube, inaudible when using a 5 or 10 watt SE amp for INefficient speakers.

If you run 105 dB horns, you really should NOT be using a tube like an 300B-XLS, or a 320B-XLS, or a AV-32B. Use a 45 or a 2A3. These big 300B-like tubes are only ideal for lower efficiency speaker loads.

JLH says 250 mVAC is inaudible on his horn sysetem, and I expect that can even be relaxed some.

Ripple is OVERDONE, make it dead quiet and you have a DEAD sounding amp !! Dan needs to design and build the supply right, before worring about any aspects of the audio circuit. The supply is the foundation of the amplifier!

Jeff Medwin

 

RE: Exactly!, posted on April 13, 2010 at 14:19:09
JoshK
Audiophile

Posts: 820
Location: NJ/NYC
Joined: August 3, 2001
 But the plate noise doesn't get amplified and the cathode noise does so it isn't an equal and opposite force. Sizing the ultrapath cap has no bearing on the amplitude of noise injected since it isn't being divided, only on how low in frequency the noise can be injected.

I think we may be arguing different things. I am discussing how loosing the solen looses the noise nulling. I think you might be really discussing sizing the Ultrapath cap large enough so that the AC current loop encompasses a wide enough bandwidth so that a cathode bypass cap isn't needed (for AC gain rather than noise nulling). I still contend you loose the voltage division required to null with the noise on the plate when you loose the bypass cap.

Think of it this way. You inject 1 part of the B+ noise into the cathode and one part into the plate. The 1 part in the cathode gets multiplied by mu (say 3 in this example) and the sign is flipped because the stage is inverting. So therefore

1*(-3) + 1 = -2 which is not zero, actually its gone UP two fold. We amplified the noise, not nulled it. In order, for the noise to null the noise injected in the cathode has to be divide by 3, so that 1/3*(-3) + 1 = 0. Now, Dan said the voltage division wasn't about nulling but I contend with the Solen you are getting at least some instead of none. I.e. 1/100*(-3) + 1 = .97 < 2.


----------------
"When Khruschev said "we will bury you" I don't think he meant with surplus parts." zacster

 

Don't forget about B+ noise that is amplified in earlier sections., posted on April 13, 2010 at 18:33:04
Chris O
 You have to look at the whole system, not just one part.

 

Question, posted on April 14, 2010 at 03:52:00
Michael Samra
Dealer

Posts: 36118
Location: saginaw michigan
Joined: January 30, 2005
 Dan
Why is the 2.2uf in series with the 220uf solen? Where the B+ being applied to the 300b? Could you maybe get 20 watts out if you go with fixed bias? I wonder if this amp can power your Bro's 1.6 maggies?
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken

 

RE: Question, posted on April 14, 2010 at 04:34:01
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Mike

I got sick just after I posted this schematic.
Good ole fever, Big V and Big D and
slept for over 24 hours.
I can almost think now so here goes ...

The 220uFd is the cathode bypass the larger
it is the lower the F3 for deeper bass.
The 420uFd PS caps are for a stiff B+ rail
again for a solid bass.
The 2.2uFd is for the output signal loop
so that the afore mentioned HUGE caps
are not detrimental to the audio band.

According to EML you can squeeze 23 watts
out of this tube if you want to.

I would think it could drive his 2.7s,
I personally prefer the 1.6s myself.

DanL



 

RE: If you have enough uF in the ultrapath cap, you do not "need" another one to ground., posted on April 14, 2010 at 04:55:22
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Chris

A big utrapath cap would defeat
the purpose of an ultrapath cap.
With my choice of cap sizes
above 20Hz the 300Bxls will see
only the 2.2uFd cap not the other
huge caps - 220 or the 460uFd.

DanL



 

RE: Don't forget about B+ noise that is amplified in earlier sections., posted on April 14, 2010 at 05:37:54
JoshK
Audiophile

Posts: 820
Location: NJ/NYC
Joined: August 3, 2001
 Indeed. I completely agree. I just don't still see how it comes out.

No problem. If it works, it works. Easy enough to try out. Measure the noise on the primary or the secondary of the OPT, with and without the solen to ground.
----------------
"When Khruschev said "we will bury you" I don't think he meant with surplus parts." zacster

 

Dan, you musthave me completely confused with someone else., posted on April 14, 2010 at 05:51:26
Allen Wright
Manufacturer

Posts: 5652
Location: Schaffhausen
Joined: January 31, 2002
 I have never mentioned ripple cancellation, and it's not something I need to get into.

I still want to know what use is the "ultrapath " cap when you have such a large cathode cap to gnd.

Surely any effective closing of the loop from the bottom of the OPT is nullified/made inefective by the 100uF to gnd?

Regards, Allen

 

RE: Dan, you musthave me completely confused with someone else., posted on April 14, 2010 at 06:47:31
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002

Allen

The only purpose to mention ratio of
the cap sizes is ripple cancellation.

From link below -
Lynn Olson
"the majority of the current will always flow along the low-impedance path. This is one of the primary advantages of the "alternative" topologies shown below; the audio-frequency (AC) currents are no longer forced through the low-quality power-supply B+ capacitor, leaving it perform the much simpler task of filtering ripple from the power supply."

The 2.2ufd cap has a lower impedance than the
220uFd / 420ufd in series so the tube will see the
2.2uFd as the componant for the current loop return.

DanL



 

Dan, I don't seem to get through to you., posted on April 14, 2010 at 07:42:43
Allen Wright
Manufacturer

Posts: 5652
Location: Schaffhausen
Joined: January 31, 2002
 Whatever, I give up.

But your schematic shows a different connection for the 100uF to gnd than does Jack E.

His goes from the B+end of the OPT to gnd, which makes some sense to me. You use it as a cathode bypass cap.

Regards, Allen

 

RE: Dan, I don't seem to get through to you., posted on April 14, 2010 at 07:58:17
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002

Allen

First, I don't have a 100uFd cap so
I am not sure what you are referring to.

Jacks ultrapath caps go from B+ to cathode.
Check the link below if you don't believe me.
Even the preamp has one to the cathode.

I just add a cathode bypass to the table
so the ultrapath cap can be much smaller as
it doesn't have to do the cathode bypass duty.


DanL




 

RE: Dan, I don't seem to get through to you., posted on April 14, 2010 at 10:50:22
Stephen R
Audiophile

Posts: 1428
Joined: January 11, 2002
 "The 2.2ufd cap has a lower impedance than the
220uFd / 420ufd in series so the tube will see the
2.2uFd as the component for the current loop return."

That's clearly untrue. 220u//420u = 144u which has a lower Z than a 2.2u cap by nearly 2 orders of magnitude.

So as far as current returns go, you will end up with a parallel path of the 220u//420u combo and to a much lesser extent the 2.2u.

As Allen says, what is your thinking on this? It certainly isn't ultrapath. It might be some form of ultrapath or somesuch thing IF your final stage was choke fed therefore isolating the 420u from the AC loop around the valve. Much like what Thomas does here http://db.audioasylum.com/cgi/m.mpl?forum=tubediy&n=178729

cheers,

Stephen

 

RE: Dan, I don't seem to get through to you., posted on April 14, 2010 at 11:52:04
Michael Koster
Manufacturer

Posts: 841
Location: Eureka, CA
Joined: October 24, 2007
 I think I see what's going on:

"I just add a cathode bypass to the table
so the ultrapath cap can be much smaller as
it doesn't have to do the cathode bypass duty."

So what I think will actually happen is the ultrapath cap will be marginalized by the much larger cathode bypass cap. In other words, if the ultrapath cap doesn't do cathode bypass duty, it does nothing particularly useful.

Take the current loop diagram from Lynn O's site you posted earlier today and put the cathode bypass cap in like your circuit has, and now look where the current loop goes...

Cheers,

Michael

 

ripple rejection, posted on April 14, 2010 at 11:56:40
JoshK
Audiophile

Posts: 820
Location: NJ/NYC
Joined: August 3, 2001
 Hi Allen,

I sounds like you don't want to discuss the topic, but I'd sure be interested to know your thoughts on the matter.

Josh
----------------
"When Khruschev said "we will bury you" I don't think he meant with surplus parts." zacster

 

I see now., posted on April 14, 2010 at 11:58:21
JoshK
Audiophile

Posts: 820
Location: NJ/NYC
Joined: August 3, 2001
 Good explanation!
----------------
"When Khruschev said "we will bury you" I don't think he meant with surplus parts." zacster

 

RE: Dan, I don't seem to get through to you., posted on April 14, 2010 at 14:39:59
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Michael

You have a 460uFd cap and a 220uFd cap in series
and a premium 2.2uFd cap in parallel with them.
Which of route will the sound favor above 20Hz?
Which will have nore influence over the sound?
You do not need to be a rocket scientist to know.
It will be the 2.2uFd dominating the sonics.

DanL



 

RE: Dan, I don't seem to get through to you., posted on April 14, 2010 at 14:52:59
Tre'
Industry Professional

Posts: 17297
Location: So. Cal.
Joined: February 9, 2002
 "You have a 460uFd cap and a 220uFd cap in series
and a premium 2.2uFd cap in parallel with them.
Which of route will the sound favor above 20Hz?"

The 220uf in parallel with the 370uf makes a 150uf cap.

A 150uf cap will have a much lower reactance (impedance) than a 2.2uf cap.

So I would think the sound would favor the 150uf cap over the 2.2uf cap as a current path. The path of less resistance-reactance-impedance....well, you know what I mean? :-)

Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"

 

RE: Dan, I don't seem to get through to you., posted on April 14, 2010 at 14:58:03
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Tre'

The larger caps have a higher ESR, loss and
higher inductance which is added in series.
As the frequency goes up over 20Hz then the
2.2uFd cap will dominate the sound more and more.
Below 20Hz I will agree the higher capacitance wins.

DanL



 

MK wrote: "" if the ultrapath cap doesn't do cathode bypass duty, it does nothing particularly useful.", posted on April 14, 2010 at 15:16:45
Allen Wright
Manufacturer

Posts: 5652
Location: Schaffhausen
Joined: January 31, 2002
 Exactly - and that's the EXACT point I've also been trying to get through to Dan.

Regards, Allen

 

Dan wrote: "It will be the 2.2uFd dominating the sonics.", posted on April 14, 2010 at 15:21:48
Allen Wright
Manufacturer

Posts: 5652
Location: Schaffhausen
Joined: January 31, 2002
 I say BS!

Above 20kHz maybe, if the elctro's are not HF types. But from 20Hz, the 2u2 will not even be noticed by anything or anybody if paralleled by so much other capacitance.

Regards, Allen

 

Everybody has a right to their opinion (nt), posted on April 14, 2010 at 19:15:58
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 DanL




 

RE: Here I go again ..., posted on April 14, 2010 at 22:05:53
Naz
Audiophile

Posts: 2184
Location: Sydney
Joined: September 2, 2005
 Dan, not bad but it's going to be one of those personal taste and experience things so FWIW, here are my personal views:

1) I also agree that the 2uF will be swamped by the much larger value cathode bypass caps and are therefore redundant.

2) I think that the OD3 is redundant in this circuit.

3) The major benefit of DC coupling is negated by the need for a large value cathode bypass cap. Reason, it's easier and cheaper to get good sounding small value caps than large ones and they are effectively in the same circuit here. Personally, I'd have gone for fixed bias on the OP tube and a negative rail for the driver to maintain the DC coupling.

4) A pair of Sic Schottkies in the cathode of the driver will give you 1.6V with better sonics IMO and definitely better reliability. They have lower impedance and don't require bypassing. You could also then lose the series R.

5) I'd probably do the PSU a little differently but that's a whole other subject.

Rgds,
Naz

 

RE: Here I go again ..., posted on April 15, 2010 at 00:35:37
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Naz

1) The larger caps have lower efficiency,
higher resistance and inductance which
are added together when in series.
At very low frequencies yes but above that...
Definitely by 1Khz the 2.2uFd is all you hear.

2) Pull out the 6GK5 without the 0D3 in there.
You fry the output tube big time.
now you know why it's there.

3) Big caps sound slow and dull yet when bypassed
you still don't think the bypass cap will affect the sound.
There is none so blind ...

4) I would play with posibilities for the bias
on the 6GK5's cathode.

5) Most people have and will always do that.
I figure since we have Jeff and TW we should
have a counter-arguement from the other side.
Instead of having the PS as a musical conduit
why not have it as a DC level supply only and
leave it to the tubes to conduct the music.

DanL



 

RE: Here I go again ..., posted on April 15, 2010 at 06:30:27
Naz
Audiophile

Posts: 2184
Location: Sydney
Joined: September 2, 2005
 Dan,

1) At the risk of an argument, then why Ultrapath? Better still if you went fixed bias on the OP tube you can lose the caps altogether and then you'll not hear any cap sonic signature.

2) True ... a seperate supply would be another way of fixing that or dear I say, cap coupling.

3) I didn't say bypassing a big cap with a small one isn't audible, although it's arguable whether this is better than a single small cap (only) to begin with. There are other good reasons for using a small coupling cap vs a large cathode bypass ... size, cost and complexity - eg you could then get rid of the OD3 because the OP tube won't fry if the driver is pulled.

4) Seriously, try Schottkies.

5) Well, PSUs are always gonna be debatable and I have my own ideas on that for another time. One thing I will say is that they become less critical if high PSRR stages are used but of course, there is limited scope in a SET OP stage for high PSRR so it's more audible.

Naz

 

RE: Dan, I don't seem to get through to you., posted on April 15, 2010 at 15:21:58
Tre'
Industry Professional

Posts: 17297
Location: So. Cal.
Joined: February 9, 2002
 Does the ESR get that big?

The reactance of a 2.2uf cap at 1kHz is 72 ohms.

The reactance of a 150uf cap at 1kHz is 1 ohm.

Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"

 

RE: Here I go again ..., posted on April 15, 2010 at 18:57:35
keto
Audiophile

Posts: 1006
Joined: September 29, 2001
 I like the fail safe direct-coupling voltage. Maybe accomodate for a pair of VR tube sockets. 2 0B3's in series would give you about 180V, which would allow you to experiment with more driver headroom, 0B3+0C3, 195V. Then if you stick to the single 0D3, you could have one for each channel. Cool circuit!

 

RE: Here I go again ..., posted on April 16, 2010 at 09:34:27
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Naz

1) The problem with large caps is speed.
A small ultrapath cap will give you speed.
I don't believe the speed has anything
to do with PS settling time it has to do
with limitations of big caps.
Ie their inductance and resistance increasing
their response time.

3) Small coupling cap vs Big cathode bypass cap
Not that simple as if you want to remove the
cathode bypass cap and have a small coupling cap
instead then you are talking about fixed bias and
all the complexity that entails.
Also I have no problem with a Solen as a cathode
bypass cap but would never use one as a coupling cap.
The Solen's limitations would be highly noticeable
in the coupling cap duty but barely noticeable
in the cathode bypass duty.
So coupling caps are alot more "in the circuit"
than the cathode bypass cap is.

5) Active loading, CCS and parafeed come to mind.
First 2 need to double the B+ voltage and is expensive.

DanL



 

RE: Here I go again ..., posted on April 16, 2010 at 11:52:47
Mike C
Audiophile

Posts: 1074
Location: Essex
Joined: November 23, 2000
 A thought about the driver cct with anode load ...

As I see it, the anode load choke acts somewhat like a CCS in that it will pass the DC that the valve needs, but will resist AC. I think that should work well; myself, I tend to use active loads to give a similar result.

I believe choke or CCS loads have several advantages but two are:
. high power supply rejection ratio
. very little AC appears across a cathode bias resistor so there is little benefit in having a bypass cap.

I therefore suggest there may be little benefit in regulating the PSU.
Also, a simple unbypassed cathode bias resistor might work just as well as the nicad/R mix?

No criticism, just some ideas to think about.

By the way, I'm building a 300B SET at present, using the somewhat similar 6AN4 as driver (active loaded). But I'm using fixed bias (negative supply to the 300B grid) as I've had good results from this in the past.
It will be interesting to see what results we get ... :-)

 

RE: Here I go again ..., posted on April 17, 2010 at 05:33:04
Naz
Audiophile

Posts: 2184
Location: Sydney
Joined: September 2, 2005
 Dan,

>>A small ultrapath cap will give you speed<<

Let's agree to disagree.

>>Not that simple as if you want to remove the
cathode bypass cap and have a small coupling cap
instead then you are talking about fixed bias and
all the complexity that entails.<<

Agree, but worth it IMO.

>>So coupling caps are alot more "in the circuit"
than the cathode bypass cap is.<<

There's been debate about this. I think it's almost as noticable. Look at your loop diagram.

Not sure what you mean about 5)

Naz




 

RE: Here I go again ..., posted on April 17, 2010 at 10:40:33
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Naz

You were talking about high PSRR methods for
output stages so active loading, CCS loading
and parafeed are the ones that come to mind.

I looked up schottky diodes and at 10mA
they only drop between .2-2.5V a piece
so that means using 6-8 in series and
doesn't their noises add together?

DanL



 

RE: Here I go again ..., posted on April 17, 2010 at 21:23:44
Naz
Audiophile

Posts: 2184
Location: Sydney
Joined: September 2, 2005
 Dan,

OP topology, I get it now. I honestly don't think that high PSRR in the OP stage is all that important because we are playing with high level signals that get divided down again. My comment was a general observation that the higher the PSRR the less audible will be the PSU, which is more and more important the further upstream we go.

On Schottkies, you are right about standard ones but Cree Schottkies will give you about 0.8V drop. I and others here have had emtremely good audible results with Schottkies in general for gain stages where you only need 1 or 2 volts bias.

Naz

 

RE: Here I go again ..., posted on April 17, 2010 at 22:58:35
danlaudionut
Audiophile

Posts: 5480
Location: Schenectady
Joined: June 6, 2002
 Naz

Just read something for you to think about
as pertaining to your post -

1) I also agree that the 2uF will be swamped by the much larger value cathode bypass caps and are therefore redundant.

From post below -

Once upon a day I bypassed a final C (a quality 50uf oil) with a small PIO
cap. I heard a noticeable difference in the symbols and other metallic sounds. I thought, "oh shit". There's no end to this. The circuit was SE with a normal OPT and a cathode bypass cap.

How much better will the higher frequencies be if
BOTH the PS cap and cathode bypass cap were bypassed?
I would think you would hear the difference all
the way down to well into the midrange too.

DanL




 

RE: Here I go again .....nice amp I was thinking of building, posted on April 20, 2014 at 18:35:50
jetrexpro
Audiophile

Posts: 57
Joined: April 19, 2010
 Hi Dan,

Did you build this amp? How did it work out? Thanks.

 

Page processed in 0.048 seconds.

[ Top Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2024, Audio Asylum, all rights reserved ]
[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]