C3G pentode for phono use ?

the road not taken.
Post Reply
gurevise
Posts: 9
Joined: Fri Feb 25, 2011 1:13 am

C3G pentode for phono use ?

Post by gurevise »

Guys,
Friend of mine built a phono with C3G pentode in first stage.
Classic pentode circuit with passive EQ.
It sounds very veiled and not clear.
So, my question is this a generic sounds of pentodes (veiled / muffled and unclear) or we have some other issue? I'm trying to figure out whether we have to switch to FET/TRIODE cascode or investigate existing design. I do not have any experience with pentodes so I have to ask...

Thanks
Sergey
Dayton OH
Geek
Posts: 4
Joined: Thu Feb 02, 2012 1:11 pm

Post by Geek »

Hi,

Got a schematic? It may not be burning hot enough.

Typically best performance with any first stage tube is to get enough mA flowing to make use of the full transconductance and all.

Cheers!
izzy wizzy
Posts: 64
Joined: Sun Jun 26, 2005 8:38 pm
Location: Folkestone, UK

Post by izzy wizzy »

I don't believe in a generic sound within reason. Things tend to work when you use them right. I use a pentode, D3a, in my phono. I like it :) But then I don't know anything about the C3G.

cheers,

Stephen

www.izzy-wizzy.com/audio
gurevise
Posts: 9
Joined: Fri Feb 25, 2011 1:13 am

Post by gurevise »

Schematic is attached
Attachments
C3G Phono sch small.jpg
C3G Phono sch small.jpg (73.71 KiB) Viewed 29722 times
Geek
Posts: 4
Joined: Thu Feb 02, 2012 1:11 pm

Post by Geek »

Hi,

C8 will be critical to sound. Is it a Nichicon MUSE or Elna Silmic type? Bypass it with a 100n polyprop would be good too.
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hey gurevise,

Did you measure RIAA response IRL? My sim check indicates errors.

Are you also aware that Zout is in the ballpark of 4kohm due to the unbypassed cathode resistor?
Brgds
Lars
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

I tend to agree. I did a quick Spice of the comp network and got a reasonably "muffled" response. A "quick fix" is to remove C4, but that isn't the best solution either. Dynamically, with varying signals, the relatively high ratio between "AC load" and "DC load" on the 5687 might not sound terribly wonderful.

BTW, for everyone, you DO know that you can "test" the sound quality in Spice by playing a pre-recorded .WAV file into the simulation, and then creating a .WAV output, right? Check the response via comparing the two .WAV files. It won't pick up the nuances of a "live" listening test, of course, but you CAN hear the differences.
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Using the original working points these are two alternatives that possibly will be closer to corrrect.

I would personally go for the second as I have found the absence of series resistor in the network sound better.

There is also a thread by Dave with a similar solution here:

http://www.intactaudio.com/forum/viewtopic.php?t=1026[/url]
Attachments
c3griaa.png
c3griaa.png (171.17 KiB) Viewed 29678 times
c3griaa2.png
c3griaa2.png (192.56 KiB) Viewed 29678 times
c3griaa2.asc
(5 KiB) Downloaded 467 times
Brgds
Lars
gurevise
Posts: 9
Joined: Fri Feb 25, 2011 1:13 am

Post by gurevise »

OK,
to sbench:
1. What exactly simulation shows? Is top end rolled of some? How much?
2. Could you elaborate on ratio between DC/AC loads comment? Most common cathode tube stages have lower DC load (Rp) and higher AC load. At least, that's my understanding. Do they have to be same and why? This is new to me.

to reVintage:
1. Good observation about 5687 Zout.
2. To clarify first /second schematic: You do NOT like the "second" schematic with R7 series EQ resistor. Correct?

Guys, you all have been very helpful.

Thanks
gurevise
gurevise
Posts: 9
Joined: Fri Feb 25, 2011 1:13 am

Post by gurevise »

Here is the emulation I did. It shows response is down about 2dB at 20kHz and fall fast after that.

Image
Maturin
Posts: 22
Joined: Thu Jul 19, 2012 5:36 am

Post by Maturin »

FWIW, one way to improve all the circuits in this thread is to regulate that screen; dropping down voltage to the screen from B+ with a bypass cap to cathode is in my understanding not the best way to run a pentode.

But, in any case, a phono preamp circuit that rolls off at 20 KHz is not going to sound good, period.
gurevise
Posts: 9
Joined: Fri Feb 25, 2011 1:13 am

Post by gurevise »

screen regulation...
i was wondering about for a while. Is there an optimal screen voltage or voltage range?
What about noise on regulated screen. It may be important in phono stage application. This is a first stage after all. Not sure how much noise we'd get from high voltage active regulator vs passive RC.

Thanks
gurevise
gurevise
Posts: 9
Joined: Fri Feb 25, 2011 1:13 am

Post by gurevise »

To reVintage:
I'd like to clarify plate RIAA EQ calculations.
Are the following equations correct:

3183uS time constant = R2 * (C8 + C9) = approx 3.09mS
318uS time constant = R8 * C8 = approx 320.33uS
3.18uS time constant = R11 * C9 = approx 3.15uS

I'm missing following:
75uS time constant = ?

Let me know

Thanks
gurevise
Maturin
Posts: 22
Joined: Thu Jul 19, 2012 5:36 am

Post by Maturin »

Re: screen regulation: well, in simulations, I just use a voltage source to the screen and bypass it to the cathode. I look at the data sheet for the pentode and the data sheet is usually pretty specific about what the screen likes to be at.

Another sim short-cut would be to use a current source to a resistor to ground; a current source across a resistor defines a constant voltage; bypass the resistor to the cathode, and you should have a well regulated screen for sim purposes, i would think.

A very good EE once explained to me that allowing the screen to "flop around" unregulated kinda defeated the purpose of a pentode - which is to eliminate the Miller effect. When you run your sim, zoom in for a close look at the voltage on the screen - with a simple drop-down resistor and bypass cap, I think you'll see more signal on that screen than you'd expect...if the subcircuit actually simulates the screen well, that is LOL.

As far as a real-world screen regulator, a mosfet source follower - or pentode cathode follower, if you wish to avoid all silicone - will nail that sucker down really well. A further refinement might be to CCS the source, but I think a bypassed source resistor probably works just fine.

It is my personal suspicion - unproven, mind you - that clipping on transients in poorly designed amps is probably far more to blame for bad sound than a well designed pentode stage. There are plenty of reel-to-reel tape recorders with EF86 pentode stages on the playback heads that sounded grrrrrrrrrrrrrreat!!!!!!!!!!!
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Gurevise,

Note Rp of C3g is not infinite!

These are the 20-100k plots of the corrections I advised you. Note scale!

About my preference for the other correction, the answer is in the link I added in my previous post.
Attachments
c3g100k.png
c3g100k.png (149.6 KiB) Viewed 29621 times
c3g2100k.png
c3g2100k.png (136.38 KiB) Viewed 29621 times
Brgds
Lars
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »


gurevise: The sim I did was pretty much what was posted in the thread... steep rolloff at high frequencies, leading to amplitude and phase funnies.

The AC vs DC loads. Since the 5687 was running unbypassed, the output resistance is higher than you might have planned for. The load on that circuit (neglecting the 1 meg) is 8.2k for higher frequencies and 8.2k+11k for lower frequencies. Sometimes this leads to dynamic or transient clipping as was pointed out. (you "clip" at essentially the 8.2k load line rate, but the DC load line is 21.2k. As soon as you clip, the 4.7uF capacitor starts to move around re-adjusting the dynamic load line right along with it. That can lead to muddy sound.

maturin: Agree fully with the dynamic clipping issue you pointed out. That's also hard to spot in simulations, but not impossible. Where it DOES show up is if you use a WAV file as input to the simulation, and output a .WAV file from the simulation and then listen to the resulting file.
Maturin
Posts: 22
Joined: Thu Jul 19, 2012 5:36 am

Post by Maturin »

sbench: thanks for your comments. I will use that tip of inputting a .wav and listening...or perhaps viewing?...the output.

I once did the following: I'd play a passage of vinyl, and take a photograph of the oscilloscope screen displaying the input to a stage at a certain point in the (classical) music. Then I'd reposition the probe to the output of that stage and replay the music and take a photograph of the oscilloscope screen again.

It often took several tries to get the same (more or less) shot - but once I could display both oscilloscope photo's one above the other on my computer screen, it was astonishing - and revelatory - to see the real-world differences...especially in the high-frequency transients.

For those who wonder at the obvious question of why I didn't just display both input and output at the same time on the oscilloscope screen, what I remember is my cheap analog scope had such a small screen I couldn't see enough detail unless I used the full screen for each screen shot.

Now that I have one of those PC oscilloscope rigs, maybe I can do it again but this time stack both displays at the same time on a REALLY BIG flat-screen TV LOL.

Sorry if this is straying off-topic; when I get new setup set up (see above) I'll start a new thread on it.
Post Reply