Proposed screen-drive amp design... comments, please? - Page 4 - diyAudio
Go Back   Home > Forums > Amplifiers > Tubes / Valves

Tubes / Valves All about our sweet vacuum tubes :) Threads about Musical Instrument Amps of all kinds should be in the Instruments & Amps forum

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 29th March 2008, 12:14 AM   #31
GordonW is offline GordonW  United States
diyAudio Member
 
GordonW's Avatar
 
Join Date: Mar 2005
Location: Marietta/Moultrie GA
I think this MIGHT actually WORK, this time!!

Changed a lot of stuff.

Went back to the 6SN7 for the LTP... it just could swing the voltage better, while being able to use the right DC grid voltage (relative to ground) to match with the output of the cascode (direct coupled).

With the 6SN7 as the LTP, that led me back to the 6BL7 for the cathode follower... no reason NOT to do that.

Still have the 6N30P as the input in cascode... but with a PARTIALLY bypassed cathode resistor. Got the gain up to 127...

The LTP now has a gain of about 8.77... with a CCS and some corrective anode resistors, balance should be pretty good. Bandwidth of the LTP should be around 210KHz... OK, I would think.

NOW I have enough gain, to get me around 20dB of negative feedback, AND still be at the same input sensitivity for full output as the Berning original circuit (about 1.5v input for full output). Depending on the outputs (which may give more than a gain of 3, according to my estimation), I MIGHT get even another few dB of open-loop gain, which might allow as much as 23-24 dB of total feedback.

My main trick I want to try, is as someone suggested (Douglas?), using a PENTODE as the CCS in the LTP. Looked around, and the 6EJ7 looked good... could handle the current, and had a suitably high plate resistance. HOWEVER- I would REALLY like some advice, in how to set up the pentode. I know I need 8ma of plate current... but EVERYTHING else is still unknown to me... what negative rail voltage on the cathode (I'd prefer to be able to use the same -200V as the negative rail on the cathode follower, if possible?), what positive voltage (assume probably something like 100v?) on the screen, what cathode resistor, what bias voltage (to set the resistors on the signal grid), etc... any assistance GREATLY appreciated!

I do thank all involved with their patience with me in this affair... I'm learning a bunch as I go along here.

Regards,
Gordon.
Attached Images
File Type: jpg screen amp 3-28-08.jpg (80.8 KB, 342 views)
__________________
Speaker Design, Restoration and Repair- since 1985.
  Reply With Quote
Old 29th March 2008, 12:33 AM   #32
SY is offline SY  United States
diyAudio Moderator
 
SY's Avatar
 
Join Date: Oct 2002
Location: Chicagoland
Blog Entries: 1
Hmmm, I'd still want more gain from the driver but...

Let's start with the cascode. Your choice of that topology means that the power supply rail will have to be REALLY tight and clean- there's almost no power supply rejection. The RC circuit in the cathode could be replaced by an LED which will outperform the inevitably crummy electrolytic cap; it looks like your required drop is pretty close to a 1.7V red. The grid of the upper tube needs to be returned quite a bit more directly to AC ground. And RF ground, too, if you don't want a VHF transmitter.

The driver LTP can be simplified. If the CCS is worth a bucket of warm spit, the plate resistors (and grid resistors of the next stage) can be fixed and equal on both sides. Matched.

I'd just set the candidate CCS tube up on the bench to determine its operating conditions. Screen currents vary a lot. Of course, I probably wouldn't use a tube there anyway...
__________________
You might be screaming "No, no, no" and all they hear is "Who wants cake?" Let me tell you something: They all do. They all want cake.- Wilford Brimley
  Reply With Quote
Old 31st March 2008, 07:32 PM   #33
GordonW is offline GordonW  United States
diyAudio Member
 
GordonW's Avatar
 
Join Date: Mar 2005
Location: Marietta/Moultrie GA
For the LED voltage-drop idea- I need more like 2.4 volts... but what, if anything, would be the problem with using a series combination of an orange LED (2v drop) and something nice like a .4V forward-voltage Schottky diode? That'd be pretty much exactly the right voltage...

Also, with diodes/LEDs... since there's NO switching involved here (it'll always be on), the normal issues with diode noise should not be any sort of an issue, right? I'd think that most ANY diode should work pretty well, right?

For AC-grounding the grids of the top cascode tube- John Broskie has a suggestion one of his pages at TubeCad, about putting caps in parallel with the resistors... which can be "ratioed" to cancel out some of the B+ noise, if done right. I'd think that might take care of both problems at one time?

One other thing- it was suggested that I change the grid-stop resistors on the cascode sections to 1K, instead of 47K. That was duly noted, and will be done...

As for the CCS tube- I actually went back and did the voltage drop calculations correctly, and determined that the 6EJ7/EF184 doesn't have enough B+ voltage handling (250V DC max across plate and cathode). So, I'm thinking something like a 6AU6... 300v B+ max... and a negative supply between -150 and -200 volts. Should get it in its "happy" range...

Thanks again, as always!
__________________
Speaker Design, Restoration and Repair- since 1985.
  Reply With Quote
Old 1st April 2008, 03:16 PM   #34
GordonW is offline GordonW  United States
diyAudio Member
 
GordonW's Avatar
 
Join Date: Mar 2005
Location: Marietta/Moultrie GA
Started designing the power supply last night.

I had decided at the start, to give the output tubes/transformer stage it's own power transformer... to avoid having to drop a lot of voltages with resistors. So, all the "front end" stages are on a secondary power transformer.

As for power supply noise on the cascode and LTP stages- I used a 260-0-260 transformer with a full-wave voltage doubler, and a LCLC stage for the +700 for the LTP (something like 16 microvolts ripple/noise, as modeled in PSUDII), and an additional RC stage to drop the voltage to 550V for the cascode input stage (something like 60 NANO-volts of ripple/noise there). Since the input stage is Class A operation (all voltages/currents average to the SAME value no matter what the voltage swing), RC filtering for this stage should be just fine, AFAIK.

On the same 260-0-260 winding, I used a bridge rectifier across the 260-260 leads... which on the positive side, used a CLCLC filter, to give +350V high-current rail for the cathode follower driver stage (less than 100 microvolts ripple), and used a swinging choke LCLC filter to get the -200V for the cathode follower (about 100 microvolts ripple), with an additional RC stage hung from that to get -150v for the LTP CCS (with about 20 microvolts ripple). Again, I think an RC stage is fine for this, since it's driving a CCS (INHERENTLY constant current- no chance of droop).

For the outputs, I'm using a 400-0-400 465ma transformer with a full-wave rectifier and a CLCLC filter, to give +500V with less than 300 microvolts ripple, and regulation within 10% at full power (507V @ 200ma draw). Since I only need to swing 400v across each output tube for full power (60w), I'd think this should be fine... it can't get into the high-current-draw region of screen current.

SY, in your experience, would you think that 60 nanovolts ripple should be low enough for the cascode to work cleanly? According to my calculations, that's like well over 120dB down from rail voltage... I honestly can't think of why it shouldn't be fine. Of course, I'll have to build and test the power supply to make sure it works as calculated... but I have pretty high hopes that it should be fine.

Oh, and for those who are counting... that's two power transformers, EIGHT chokes and an output transformer PER AMP. Good thing I found a source for CHEAP surplus transformers and chokes!! The outputs will be new (Edcor), but the others are going to be like half of new prices, if that much... which makes it a LOT more practical!!

Regards,
Gordon.
__________________
Speaker Design, Restoration and Repair- since 1985.
  Reply With Quote
Old 1st April 2008, 03:40 PM   #35
SY is offline SY  United States
diyAudio Moderator
 
SY's Avatar
 
Join Date: Oct 2002
Location: Chicagoland
Blog Entries: 1
I'd still actively regulate the input/driver if possible- it's easy, cheap, and will be nothing but good for LF stability. Noise is much less of an issue than impedance (as you surmise, the PSR of the LTP and cathode followers is very high and the noise is all common-mode), and regulators do a better job at DC and low frequencies.
__________________
You might be screaming "No, no, no" and all they hear is "Who wants cake?" Let me tell you something: They all do. They all want cake.- Wilford Brimley
  Reply With Quote
Old 3rd April 2008, 11:17 PM   #36
GordonW is offline GordonW  United States
diyAudio Member
 
GordonW's Avatar
 
Join Date: Mar 2005
Location: Marietta/Moultrie GA
May have found a better front-end...

What do you folks think about this:

5687 in cascode

745V B+ (I wound up with that for the LTP, due to transformer choices... a little higher still worked with that, without problems)

60K anode resistor

Cathode has orange LED (2.0v drop) in series with 30 ohm resistor

There's right at 10ma current through the stage, and gain is around 135. Bias voltage is about 2.38 volts or so. That should give me about 7.5 dB of voltage headroom on the front-end, between max output of the amp and input stage clip.

By raising the B+ of the first two stages to 745V, I also got an LTP output voltage of around 165V... which helps there, too. DC voltage carried between the cascode pentode and the LTP grid is about 139v, above ground, with about a -18.5V bias on the LTP (gain still around 8.7 there)...

The 5687 has a LOT less Cgp than the 6H30Pi... 4pf as opposed to 6pf. Quite a bit more bandwidth, with about the same gain...

Better still, the 5687 can be had CHEAP, compared to the 6H30P...

Regards,
Gordon.
__________________
Speaker Design, Restoration and Repair- since 1985.
  Reply With Quote
Old 7th April 2008, 05:47 PM   #37
GordonW is offline GordonW  United States
diyAudio Member
 
GordonW's Avatar
 
Join Date: Mar 2005
Location: Marietta/Moultrie GA
Default Wholesale changes...

I'm frankly tired of banging my head against a wall, trying to get this thing to work with a LTP phase splitter. Obviously, this is not something that can easily be done with tubes that are commonly available, or with reasonable B+ voltages.

Decided to try something different... a concertina phase splitter with equal output impedance (cathode follower on the "top" side), and a second voltage amp stage... more like the original Berning EA230 configuration, but with steps taken to get around the limitations (voltage swing, bandwidth, gain) of the original design.

Once I decided to do this, I discovered FIRST off, that the 700V B+ rail is NO LONGER needed. I AM still going to use separate power transformers (one for the output, one for the rest of the amp), but I no longer need a voltage doubler on the second transformer.

Also, I was able to use MUCH MORE NORMAL tubes. Some of these changes might actually make the whole thing more linear than before...

I'm getting MUCH more voltage swing out of this, without the extreme B+ voltages (the most across ANY stage is 610 volts, for the second voltage amp and output cathode follower driver). Like, this thing MAY be able to put out as much as 80w/ch, depending on the behavior of the output tubes. I'm getting like close to 180v available at the screens of the output tubes, by the sims!

Also, the predicted bandwidth is MUCH better. Doing the Miller-capacitance calculations for likely alignments of a few of the different options for input pentode tube, I'm getting something like 450KHz... and the REST of the amp is "greater than 1MHz" according to TubeCad. And this time, the math actually agrees on paper! In this, the output transformer will be the only "wild card" as far as bandwidth is concerned... and I am pretty confident that I can take care of that situation with the Zobel on the input stage or the cap in the global feedback loop. If needed, I can always add a cap from a tube plate to the global loop, or try some of the things Bricktop did to "fix" the Heath W4... using the plates voltage on the output tubes as a "partial drive"/feedback for the second-stage voltage amps (see this thread on AudioKarma: http://www.audiokarma.org/forums/sho...=149678&page=2 )... stuff like that should be able to take care of things, I would think, at least in theory.

The astute viewer will notice that there are 3 LED's... in the cathode circuits of various stages... I will set whatever voltage drops I need (and whatever LED's I need to string together in series to do that) depending on whatever voltages I wind up needing, to properly "partially bypass" the cathodes of those tubes, to get the gain I need. That does seem like a much better idea than a resistor and an electrolytic cap!!

I don't have resistor values for the first stage yet; I'm planning to set it up for a gain of right around 100. Should be no problem for any of the tubes I'm thinking about. Gain of the cathodyne/concertina is of course just below unity (.95, IIRC), gain of the second voltage stage is about 13 or 14 (depending on what values I finally wind up using), the output driver cathode follower has a gain of about .9... and I'm assuming AT LEAST a gain of 2.5 from the output tubes (even in screen mode, they should do THAT much, if not more). If I wind up with lots of gain on the outputs, I'm going to drop back the gain on the first two stages a bit, so that total feedback (from the two loops, combined) is around 17-20 dB.


Wow... it's nice to actually have GAIN TO SPARE, at least in theory!!

So, any thoughts on this one????

Thanks!

Regards,
Gordon.
Attached Images
File Type: jpg screen-amp-4-7-08-jpg.jpg (69.2 KB, 247 views)
__________________
Speaker Design, Restoration and Repair- since 1985.
  Reply With Quote
Old 15th April 2012, 12:31 PM   #38
AJT is offline AJT  Philippines
diyAudio Moderator
 
AJT's Avatar
 
Join Date: May 2003
Location: Palatiw, Pasig City
any updates?
__________________
the best advertisement for a good audio design is the number of diy'ers wanting to build it after all the years....never the say so of so called gurus....
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
How do I design "SCREEN GRID DRIVE" ANALOG GUY Tubes / Valves 27 26th December 2013 02:35 AM
Opinions on proposed design? leadbelly Multi-Way 13 18th September 2006 09:16 AM
Need comments on a proposed sub design cdwitmer Multi-Way 16 23rd August 2004 09:56 AM
proposed 833 design James D. Tubes / Valves 0 25th March 2004 12:42 PM
Comments invited: proposed remote volume control tcpip Parts 2 2nd February 2004 07:07 AM


New To Site? Need Help?

All times are GMT. The time now is 11:06 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2