Concertina sanity check - 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 26th June 2007, 12:14 AM   #1
PeteN is offline PeteN  United Kingdom
diyAudio Member
 
Join Date: May 2007
Post Concertina sanity check

Hello again all.

Right. Well. After a few weeks perusing and musing over the [in]famous Morgan Jones' Valve Amplifiers I have come to the point where I've printed a bunch of data sheets and put pencil to graph-paper. The world may never be quite the same again.

Fortunately for me the book goes to within a single choice of tube of what I ultimately aim to produce [EL34 output instead of 84], with the Bevois Valley, so I've also been following Kofi and others' threads here to keep up to speed. Unfortunately I found once or twice throughout the book I was left wondering where x or y number/value came from or why it was chosen - I think I've got this down pat now, at least for the concertina stage, but I'd like a sanity check of my concertina just to check I'm not going down the wrong track too far.

OK so I want a concertina phase splitter.
I want to drive the stage preceeding the conertina with the "standard-ish" 2VRMS from an MP3/CD/DVD player, which means linearity over a region about 5.6V wide is desirable (being twice the peak voltage of said input signal), and we want an operating point at -3V or below to avoid grid current on the positive peaks, which will be at 1.44 * 2V = 2.88V (+Vgk), assuming a sine wave.
I investigated a bit and I found a nice linear tube: 6H6Π in Cyrillic which is 6N6P in Latin (dual triode, mu ~ 20).
I'll admit I was looking into the E88CC as per the book and feel a bit guilty for changing this (probably making it more work for myself) - however I've always found I learn best if I'm applying the knowledge instead of reading over a book and thinking "uh-huh. yep. ok". So, despite the book's EL84 and my EL34 requiring virtually the same input signal and me thus able to keep the stage as it is in the book [and modify only feedback], I've decided for the interim to go with 6N6P. It also looks somewhat more linear than the '88, and it's a bit cheaper which is nice.

Anode curves attached. NB the lines aren't totally accurate as they're just a quick transcription of what I have in front of me on paper at a larger scale.

Looking at the anode curves, it's really nice and linear (equally spaced curves) around Vgk = -6V, and best between Ua = 110 to about 150V.
The book states that 2x22kΩ is the 'traditional' values to use on either side of the concertina, as far as I understand to stop problems when loaded with a class B following stage - which this will not be (AB). Also 22k is used to keep power consumption down - however the 6N6P has a pretty high Pamax so I can allow a bit more current. This is me toying with that "tubes have guidelines, not ratings" sig I've seen round here somewhere.

To get the loadline up into the regions where the curves look a bit less bunched, I opt for 10mA from a 250V HT - total loadline is thus 25kΩ from 12.5kΩ on either side of the valve. The 2.5W dissipation is well within the valve's 4.8W limit (incidentally I'm working with 4W as the limit due to using both halves of the tube).
->This 25kR line crosses the -6V curve at Ua ~128V.
->Taken from the HT to give voltage across Ra & Rk = 122V.
->These resistors are equal so they get 122/2 = 61V across them each.
-> Grid will be 6V lower than this = 55V, and this will be the anode voltage of the driver stage.

Under these conditions, drawing a tangent to the -6V curve at the operating point leads us to a conclusion that
-> ra = ΔVa/ΔIa
-> ra = 250-106 V / 32.5-0 mA
-> ra = 4.5kΩ

Looking horizontally left and right to the -4V and -8V curves and reading U(V)a we conclude that
-> μ = ΔVa/ΔVgk
-> μ = 158-95.5 / 4
-> μ = 62.5/4 = 15.625
(gm = μ/ra and is thus 15.625/4.5 = 3.47 mA/V)

Gain (A) is (rather bewilderingly) μRL / { RL(μ+2) + ra } and is ~ 0.87

The next stage (cap. coupled PP output EL34s) have Vgk of -12.5V, thus they need a 12.5/sqrt(2) = 8.84VRMS signal fed into them.
If we want to get that out of this concertina, we need to consider the less than unity gain and thus our driver stage before the concertina needs to take a 2VRMS input and give a (1/0.87)*8.84 = 10.18VRMS output, implying a gain of about 5.09 - which we'll call 5.1.
This driver stage also needs to be linear (fortunately, 6N6P is a dual triode so I can use the other half), and have an anode voltage of Va = 55V. The problem will be in finding an operating point with 55V anode voltage. *Musing* maybe I could increase the OP to -3V from -6V which would require a concertina Vgk of 79.5V which is accessible to the -3V 'curve', at the risk of a few microamps of grid current. Hmmm.

Heellpppp it's all just a sea of numbers! Am I on the right track?
Attached Images
File Type: jpg conc.jpg (58.3 KB, 314 views)
__________________
-- Peter
  Reply With Quote
Old 26th June 2007, 12:50 AM   #2
diyAudio Member
 
Tom Bavis's Avatar
 
Join Date: Jun 2002
Location: Macedon NY
You DON'T have to DC couple the input stage, which would let you use a different operating point. A higher B+ voltage might work out better for direct coupling...

If you want to run Ultralinear (or have the option of triode operation), you'll need more grid swing too. If you have extra gain, it can be used to lower output impedance with feedback, a good idea with pentode operation, which otherwise may give a frequency response as ragged as your speaker's impedance curve...
  Reply With Quote
Old 26th June 2007, 01:07 AM   #3
SY is offline SY  United States
diyAudio Moderator
 
SY's Avatar
 
Join Date: Oct 2002
Location: Chicagoland
Blog Entries: 1
If you're using a degenerated input stage like the Bevois with a feedback connection to the cathode, the input dynamic range requirements are eased. Remember, the input is referenced to ground, whereas the tube operates according to Vgk. So you don't need anything there that biases at -6V (say). You do want something in that first hole that has a linear swing sufficient to drive the output stage to full power (I'd guess 70-80V p-p).

I found that, in this service, I can get lower distortion with an ECC81 than MJ's choice of ECC88. Not as sexy or exotic as a Russian tube with Cyrillic writing, but...

If it were me and I were using EL34, I'd use a different driver stage- the common-cathode->split load is more suited for lower drive voltages. A CCS long tail splitter might be more appropriate, though you'll probably need to cascade two stages.

If you want to use a concertina, remember that with 100% degeneration, the linearity is less of an issue. The choice of load will be determined by swing requirements (lower R means lower swing) traded off against load changes when the output stage transitions out of class A (lower R means less unbalance when that happens).
__________________
And while they may not be as strong as apes, don't lock eyes with 'em, don't do it. Puts 'em on edge. They might go into berzerker mode; come at you like a whirling dervish, all fists and elbows.
  Reply With Quote
Old 26th June 2007, 04:01 AM   #4
diyAudio Moderator Emeritus
 
ray_moth's Avatar
 
Join Date: Jan 2004
Location: Jakarta
Quote:
So, despite the book's EL84 and my EL34 requiring virtually the same input signal
Where did you get that idea? EL34 requires significantly more input signal that EL84, to the point where a concertina splitter ceases to be a good choice as a driver. You're too close to the cutoff point, at which the concertina starts behaving poorly. You could use a concertina splitter if it's followed by a PP or differential driver (Williamson style) but that adds another stage and may prove to be a problem with NFB.

As SY says, an LTP splitter (using a CCS in the tail) could serve well as a combined splitter/driver. A 12AT7/ECC81 would do a good job in that posistion and you could use a high-mu triode as a directly coupled input stage to complete the amp, with enough gain to apply some NFB for ultralinear or triode mode. If you want to use the EL34s in pentode mode, then a pentode input stage (e.g. EF86 or 6AU6) would be a better choice as the input stage, because it would give more gain so you could apply significantly more NFB (you'll need it).
  Reply With Quote
Old 26th June 2007, 01:33 PM   #5
PeteN is offline PeteN  United Kingdom
diyAudio Member
 
Join Date: May 2007
Quote:
Originally posted by ray_moth
Where did you get that idea? EL34 requires significantly more input signal that EL84, to the point where a concertina splitter ceases to be a good choice as a driver.
Can you explain this please?
Here's the logic btw:

EL84 - Book states it needs "8VRMS" for full output (later we're told it's 8.636). Peak voltage of an 8.636VRMS sine wave is 12.213..V.
EL34 - Biased at -12.5V. Allows 12.5V swing either side of the operating point (limited only by Vgk being 0V), thus a peak voltage of 12.5V, which means an RMS (again assuming sine) of 8.838..V
Hence my "practically the same" (0.2V difference).
It would seem somewhere in my logic there is a bit of a gap. Well, maybe more of a crevice?

Quote:
You do want something in that first hole that has a linear swing sufficient to drive the output stage to full power (I'd guess 70-80V p-p).
That's 70-80v pk-pk input? Or output swing?


If you think the concertina isn't such a good idea for this level of "powah" I'll look into the LTP and Cathode coupled types. I'm trying to keep the tube count down (cheaper power tfmr) so I'd rather avoid an additional stage!

Thanks for the help guys
__________________
-- Peter
  Reply With Quote
Old 26th June 2007, 02:04 PM   #6
diyAudio Member
 
Tom Bavis's Avatar
 
Join Date: Jun 2002
Location: Macedon NY
You didn't mention the power you're looking for - or the supply voltage. EL34 biases at about -13.5V at 250V (class A) and about -37V at 400V (class AB1).
  Reply With Quote
Old 26th June 2007, 02:07 PM   #7
SY is offline SY  United States
diyAudio Moderator
 
SY's Avatar
 
Join Date: Oct 2002
Location: Chicagoland
Blog Entries: 1
Quote:
That's 70-80v pk-pk input? Or output swing?
Input swing, but I didn't realize that you were using an unusually low B+ for the EL34s. Mea culpa. Typically, they run with 400-500V on the plates, requiring 30-35V of bias.
__________________
And while they may not be as strong as apes, don't lock eyes with 'em, don't do it. Puts 'em on edge. They might go into berzerker mode; come at you like a whirling dervish, all fists and elbows.
  Reply With Quote
Old 26th June 2007, 02:36 PM   #8
diyAudio Moderator Emeritus
 
ray_moth's Avatar
 
Join Date: Jan 2004
Location: Jakarta
Me too! EL34 in push-pull is invariably operated in class AB1 or B1, with at least 400v at the plate.

BTW, an LTP splitter will not cost an extra tube, because it has gain. A concertina has no gain and therefore needs an extra tube to drive it. It's 6 of one and half a dozen of the other, although you will normally get more gain from a concertina splitter + input stage than you will from an LTP (alias cathode-coupled alias differential) splitter.
  Reply With Quote
Old 26th June 2007, 03:37 PM   #9
PeteN is offline PeteN  United Kingdom
diyAudio Member
 
Join Date: May 2007
I get the idea I'm in way over my head still.

I was going by the "typical" 100mA/250V operation I've seen on both recent JJ and older Phillips datasheets for the EL34 - I guess that's SE though. Looking at the Phillips one again they do indeed seem to hint at 350+V.

Looking at the Phillips curves, at (say) 400V, max power limits current to just over 60mA, sticking a load line through here it cuts right through the curves through or just after the kinks (6.67k loadline).
*snip from here*
That's with 360V on g2. JJ's datasheet has curves with 250V g2 which are noticeably less kinky - if I had to guess I'd say that ultralinearising the stage takes care of these kinks with wiggling Vg2 somewhere near HT, so it's all OK after all?
*end snip*
Bah! Ignore my (snipped) delusional rantings - I was looking at the triode-strapped curves. The Pentode ones are Vg2=250V with Phillips as well, and again the kinks are less pronounced.

If it is operated at this higher voltage, I can understand the OP being as low as -30V.
If the OP was at -30V, surely hitting the Va axis of the graph would limit input to (from just eyeballing it) -40V or so [10V lower], meaning pk-pk of only 20V, and RMS input of 28? How do you arrive at big ones like 70p-p?

Intended power is circa 25W by the way, just one EL34 either side of the primary.

I've just been reading the book again, and the cathode-coupled splitter makes good sense (I'll admit the see-saw one has me a little stumped), and as someone said doesn't require any additional glassware.

What I'd *like* to do is just breadboard a circuit and have a play, but the voltages make that rather a gamble with one's life and wallet.
__________________
-- Peter
  Reply With Quote
Old 26th June 2007, 04:30 PM   #10
kevinkr is offline kevinkr  United States
diyAudio Moderator
 
kevinkr's Avatar
 
Join Date: Sep 2004
Location: Boston, Massachusetts
Blog Entries: 6
Say you're running the EL34 on a plate voltage of 400V, and you are centered on the load line at a quiescent bias voltage of -35V, then to drive the tube to the cut off intercept (0 mA plate current) you would need to swing it to approximately 2X this value or -70V, and to the point of saturation to 0V (assuming no grid current flows) so this is 70Vpp, and for good linearity generally I find it is a good idea to design the driver to swing at least 3 - 6dB more particularly if global feedback is not used.

(Note all values I mentioned above are approximations and for good linearity you normally don't want to operate the output tubes very near saturation or cut off - the exception of course is class B where one tube is fully cut off for a half cycle.)
__________________
"Extraordinary claims require extraordinary evidence." - Carl Sagan
  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
F1 Voltage sanity check rpafenberg Pass Labs 7 28th March 2009 11:42 AM
Sanity Check ..... please Lostcause Digital Source 0 16th October 2008 07:19 PM
XO sanity check for newbie avizandum Multi-Way 2 30th December 2007 09:29 PM
This model can't be right... need sanity check NogginBoink Subwoofers 4 30th July 2005 10:14 PM
Sanity check planet10 Multi-Way 14 7th March 2004 11:15 PM


New To Site? Need Help?

All times are GMT. The time now is 05:19 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