Reply
 
Thread Tools Display Modes
  #1   Report Post  
Tom Schlangen
 
Posts: n/a
Default "AGC" for magic eyes

Hi RATs,

both projects I have on the bench at the moment
have a larger dynamic (read power) range than the
first amp I utilized magic eyes (for merely the
nice look of them when flickering).

You can see the circuit I used in the little
amp half way down the page at:

http://www.ndh.net/home/schlangen/ro...aten/em80.html

Useful reading range voltage at the grid of EM80 is
between 0 volts (shows thin line) up to ~ -18 volts
(full reading).

It is nearly impossible get a reading at normal
listening levels whatsoever from the magic eyes
when the max. reading (-18 volts at grid) is
adjusted for max. reading at max. output power
for a stronger amp.

A simple switch and a voltage divider gives a
"sensitive" range for low amp output levels
and a range for higher output levels, but
having to switch the ranges manually is not
practical.

Adding a zener so that at least the EM80 isn't
overdriven isn't optically nice and practical
either, since there is no change in readout up
from a certain amp output level.

Does some have an idea how to implement a
Real Simple (!) sort of dynamic range compression
or AGC, without making things too complicated?

At most I would dare to add one double triode for
_both_ EM80 present (one additional triode system
per channel).

Any ideas appreciated,

Tom

--
Live is too short to be taken seriously.
- Oscar Wilde
  #2   Report Post  
Robert Casey
 
Posts: n/a
Default

Tom Schlangen wrote:


Does some have an idea how to implement a
Real Simple (!) sort of dynamic range compression
or AGC, without making things too complicated?

At most I would dare to add one double triode for
_both_ EM80 present (one additional triode system
per channel).

Any ideas appreciated,

Tom



Thought comes to mind: Use a remote cutoff tube as a "preamp" for the
power level signal. Cathode follower?

  #3   Report Post  
Fred Nachbaur
 
Posts: n/a
Default



Tom Schlangen wrote:
Hi RATs,

both projects I have on the bench at the moment
have a larger dynamic (read power) range than the
first amp I utilized magic eyes (for merely the
nice look of them when flickering).

You can see the circuit I used in the little
amp half way down the page at:

http://www.ndh.net/home/schlangen/ro...aten/em80.html

Useful reading range voltage at the grid of EM80 is
between 0 volts (shows thin line) up to ~ -18 volts
(full reading).

It is nearly impossible get a reading at normal
listening levels whatsoever from the magic eyes
when the max. reading (-18 volts at grid) is
adjusted for max. reading at max. output power
for a stronger amp.

A simple switch and a voltage divider gives a
"sensitive" range for low amp output levels
and a range for higher output levels, but
having to switch the ranges manually is not
practical.

Adding a zener so that at least the EM80 isn't
overdriven isn't optically nice and practical
either, since there is no change in readout up
from a certain amp output level.

Does some have an idea how to implement a
Real Simple (!) sort of dynamic range compression
or AGC, without making things too complicated?

At most I would dare to add one double triode for
_both_ EM80 present (one additional triode system
per channel).

Any ideas appreciated,

Tom


Since we're not directly dealing with audio signals that our ears will
ever hear, it's probably not heresy to suggest the following:

A silicon diode and resistor tree.

You can tailor just about any kind of response curve you like. Toss it
into a simulator (such as Circuitmaker) and have fun playing!

An example is the "limiter" circuit shown he

http://www.dogstar.dantimax.dk/tubestuf/dzart-6.htm

Cheers,
Fred

  #4   Report Post  
Tom Schlangen
 
Posts: n/a
Default

Hi Ronald,

Maybe something simple as connecting
2 pots meganicaly ......


Surely a simple dual pot with log characteristics
would do, but one has to _turn_ it ... but,
I am looking for some more "automatic"
sensitivity control.

Tom

--
The first rule of magick is simple: Don't waste your time
waving your hands and hoping, when a rock or club will do.
  #5   Report Post  
Tom Schlangen
 
Posts: n/a
Default

Hi Robert,

Did a little simulating in Circuitmaker.
Using a remote cutoff pentode like
a 6AR11 (compactron with a pair of identical
pentodes) as an emitter follower


You surely mean _cathode_ follower ;-)

looks to give good compression of the level
signal. [...]


Thank you very much, this suggestion is interesting
to me since it would provide real signal dynamics
compression (not only limiting).

But I must admit that I can't find said compactron
(and suitable socket) for a reasonable price over
here in Germany.

Are there any alternatives with more "common" tube
types and sockets? I understand that one single double
triode for _both_ channels can't emulate two
remote cutoff pentodes ... but a real compressor
function is appaling, if it could be done easily with
few usual/common parts. No need for audio quality
compression, it is only an optical feature.

Tom

--
Falling in love is a lot like dying.
You never get to do it enough to
become good at it.


  #6   Report Post  
Fred Nachbaur
 
Posts: n/a
Default



Tom Schlangen wrote:
Hi Robert,


Did a little simulating in Circuitmaker.
Using a remote cutoff pentode like
a 6AR11 (compactron with a pair of identical
pentodes) as an emitter follower



You surely mean _cathode_ follower ;-)


looks to give good compression of the level
signal. [...]



Thank you very much, this suggestion is interesting
to me since it would provide real signal dynamics
compression (not only limiting).

But I must admit that I can't find said compactron
(and suitable socket) for a reasonable price over
here in Germany.

Are there any alternatives with more "common" tube
types and sockets? I understand that one single double
triode for _both_ channels can't emulate two
remote cutoff pentodes ... but a real compressor
function is appaling, if it could be done easily with
few usual/common parts. No need for audio quality
compression, it is only an optical feature.

Tom


For your "meter compressor" you don't really need two channels. Just
sample the sum of both channels, unless you listen to a lot of music
with everything in one channel you'll be alright. ;-)

As for common tubes - look for the tubes used in IF stages of AM radios.
These will almost always be remote cutoff, allowing gain to be varied
with DC control voltage (AVC). Here the most common is 6BA6 (and 12BA6),
I don't know what the equivalent is in Euro radios.

Cheers,
Fred

  #7   Report Post  
Fred Nachbaur
 
Posts: n/a
Default



Tom Schlangen wrote:
Hi Fred,


[...]



This is very interesting, since it looks simple and
I have lots of 1N4148 at hand ;-)

But I have some questions regarding the principle.
This is obviously a _limiter_ circuit (you even call
it so yourself), although with a non-linear curve,
generated by several steps of voltage dividers, working
into separated diode pairs with their breakthru voltage
(being around ~0,7 volts, I suppose).

If the ladder would have an infinite number of cells
(with epsilon converging to zero spacings), could this
theoretically be make a (linear) compressor function
with a given constant i/o voltage ratio (above
the breakdown voltage of the used diodes, of course)?


Er.. mm.. yes. I mean definitely maybe.

In other words, I don't have a clue, but it sounds like it might. ;-)

Besides theoretics and since I don't have the Circuitmaker
software at hand, I would be glad if you could give
me a dimensioning hint for the voltage devider resistors
for my application, requiering roughly a 2:1 "compression",
err, stepped limiting.


Well, you can work it out the hard way also. For example, using the
circuit referenced above:

Below 0.7 volts, the "gain" will be close to 1, since the output load
resistance (470k) is so much larger than the series resistor R72 (100k).

From 0.7 to 1.4 volts, the gain will be about 200/(100+200) or 0.67.

From 1.4 to 2.1 volts, the 200k and 120k resistors are effectively in
parallel, or 75k. So gain is 75/(100+75) = 0.43

From here on, each new resistor just gets successively paralleled.

From 2.1 to 2.8 volts, gain is 35.7/(100+35.7) = 0.26

From 2.8 to 3.5 volts, gain is 20.3/(100+20.3) = 0.17

And over 3.5 volts, the curve is flat, with a gain of 13.3/(100+13.3) = 0.10

You could spreadsheet it if you like, or just play around empirically.
Since your output device is not something that could be called
"precision" by any stretch of the term, and you're just going for visual
effect, I think the empirical approach is entirely acceptable here. ;-)

I suppose the 5 ladder stages/steps are enough for the
"automated" eye candy effect I am longing it for. The
input range is expected to be 0 to -40 volts, and the
EM80/6BR5 light/shadow readout angle is quite linear to
grid voltage between -2 to -18 volts.


Well, so you have a couple options. Since the output voltage should be
up to about -18 volts, this would require some 26 diodes. As hinted
above, you don't need anywhere that kind of resolution. So what might be
better (if you go this route) is to attenuate the signal so it fits into
the 0 to -3.5 volts, and then amplify it again. Except that now it's
inverted, which will make your display work backwards.

Another option: instead of paralleled conventional diodes, use zener
diodes in series (opposing). Then, five 3.6-volt zener-diode pairs would
give close to your 18 volt output range.

Okay, of course I could swap thru the whole E12 resistor
values line with [lots of]^5-1 possible combinations,
but ... ;-)


Yes, as with most simple things, it could be made as complex as you like.

Thank you very much!


You're most welcome!

Cheers,
Fred

  #8   Report Post  
Robert Casey
 
Posts: n/a
Default

Fred Nachbaur wrote:



Tom Schlangen wrote:

Hi Robert,


Did a little simulating in Circuitmaker.
Using a remote cutoff pentode like
a 6AR11 (compactron with a pair of identical
pentodes) as an emitter follower




You surely mean _cathode_ follower ;-)


looks to give good compression of the level
signal. [...]




Thank you very much, this suggestion is interesting
to me since it would provide real signal dynamics
compression (not only limiting).

But I must admit that I can't find said compactron
(and suitable socket) for a reasonable price over
here in Germany.

Are there any alternatives with more "common" tube
types and sockets? I understand that one single double
triode for _both_ channels can't emulate two
remote cutoff pentodes ... but a real compressor
function is appaling, if it could be done easily with
few usual/common parts. No need for audio quality
compression, it is only an optical feature.

Tom



For your "meter compressor" you don't really need two channels. Just
sample the sum of both channels, unless you listen to a lot of music
with everything in one channel you'll be alright. ;-)

As for common tubes - look for the tubes used in IF stages of AM
radios. These will almost always be remote cutoff, allowing gain to be
varied with DC control voltage (AVC). Here the most common is 6BA6
(and 12BA6), I don't know what the equivalent is in Euro radios.


Thought a pair of eye tubes were desired..... At least I'd want a
pair. Anyway, the
remote cutoff tubes need not be pentodes, a 6ES8/ECC189 variable mu
triode would do.
Heard that someone with a large supply of 6ES8s relabeled them as 6DJ8s
and sold
them as such..... If you happen to have a "6DJ8" that sounds
terrible, it might be
one of these.

Equivalents to a 6BA6 look to be: 5749, 6660, EF93, CV454, 6F31, CV2026,
CV5037, PM04, W727. Shouldn't be too hard to find a pair of these. Also
remote cutoff pentodes intended for TV sets should be easy to find. 6BZ6,
EF183, 6GM6, etc.


  #9   Report Post  
Tom Schlangen
 
Posts: n/a
Default

Hi Fred,

Another option: instead of paralleled conventional
diodes, use zener diodes in series (opposing).
Then, five 3.6-volt zener-diode pairs would
give close to your 18 volt output range.


This seems to be the most practical idea; I'll try
that soon and report back.

Thank you again!

Tom

P.S.: Probably the spec sheet of the final amp
will look somewhat irritating: "It utilizes 7
electron tubes and 25 solid state diodes..."

--
fnord (now you see it, now you don't)
- R.A. Wilson / FZ
  #10   Report Post  
JKoning
 
Posts: n/a
Default

You say you would like to implement something Real Simple ...
Well, the description might not be, the final set up is....

With reference to the "Power Meter" schematic as linked in your post :
1. Replace the 220K pot with a CDS-cell or LDR (Light Dependent Resistor.
Light : Licht, nicht leicht, dennoch nicht schwer :-)
2. Remove the 0.1µF/400V cap. and relocate it to the new-projects box.
3. Replace the 390-470K resistor with the a.m. 220K pot.(P1)¹.
4. Don't connect P1 to tube anode but to OPT-secondary, hot side², instead.
4. Connect the P1/LDR node to one terminal of an additional 0.1µF (C1) cap.
5. Connect the other C1 terminal to the 1N4148/1N4148 node.
6. Solder two Leds³ in antiparallel and ground one terminal.
7. Connect other Led wire via a series resistor(Rs)¹ to OPT-sec., hot side².
8. Leave rest of circuit as is.
¹. see text below.
². assuming the OPT secondary, neutral terminal, is grounded.
³. chose 3mm red leds for easy mounting and lowest Vfwd drop.
For convenience I nominated parts, left to right, LDR, C1, D1, D2, C2, R1.

Both Leds and LDR must be mounted face-to-face in a light-tight enclosure.
I assume while looking to the "modified-schematic" you already got the grasp
of it's principle. While an amp's power level increases, the Leds light up
and at the P1/LDR/C1 node the voltage decreases in relation to the P1/OPT
node because of decreased resistance of the lit LDR. Absolute voltage
increases however, but not in a linear relation to the input.

Typical LDR values are 10Mohm in complete darkness to 100ohm in bright
daylight. The "transfer-ratio" of the Leds/LDR combination is such that
only very little Led current is required to drive the LDR's resistance down
to a value appropriate for the application. Rs chosen depends on the desired
"compression-ratio", Led efficiency and of course the LDR sensitivity. In my
experiment Rs = 10K ...47K, corresponding to only 2.1...0.45mA Led current
at full blast, is appropriate.

In my case the application was an EM84 (Vg = 0...-22V) as "power indicator"
in a 70W amp with 4xEL34 in ppp used as a guitar amp. I say "was", as the
amp. parted house a long time ago but probably is still on duty somewhere on
the globe. Yet found the schematic and the homebrew Led/LDR "optocoupler"
prototype in the junkbox. Some modifications were made in relation to the
values as mentioned above. F.e. C2 is changed to 1µF, for sake of test &
measurement at 50Hz, by means of a 24V PST and Variac, to get rid of ripple.
For P1 a decade box is used instead.

The C1/D1 clamp and D2/C2 combination, a doubler in essence, will output a
Vdc at the D2/C2/R1 node of around 2.8 x Vrms. In power amps of ~30W into
8ohm the Vout of ~15Vrms, or ~44Vpp, is sufficient to drive an EM80 (Ug
= -1...-14V) from the OPT without the need to connect to a source with
higher ac voltage like the anode of the output tube. If the doubler's Vout
is not sufficient, I think a tripler or a quadrupler could be used instead.
As an alternative, and if present, the 16Ohm OPT tap comes in handy
with ~1.4 times the 8Ohm output.

A refinement can be made by connecting the cathode of D1 to a -1.6V
(typical) source, obtained via the output tube bias circuit or rectification
of the filament supply etc. In this way Vg will be at -1V at zero signal,
thus providing deflection at the EM80's indicator (what's the name of the
green "curtain" anyway?) at the smallest signal, or without threshold.

Results of the EM84 "compressor" derivative:
Pout¹ Vrms¹ Vpp Vg¹ Vg² Vg³
0.00 0.00 0.00 -1.0 -1.0 -1.0
0.25 1.41 4.00 -3.6 -3.1 -2.6
0.50 2.00 5.66 -4.6 -4.2 -3.4
1.00 2.83 8.00 -5.9 -5.4 -4.5
2.00 4.00 11.3 -7.3 -6.9 -5.9
4.00 5.66 16.0 -8.8 -8.5 -7.6
8.00 8.00 22.6 -10.2 -10.1 -9.4
16.0 11.3 32.0 -11.5 -11.6 -11.1
32.0 16.0 45.3 -12.8 -12.8 -12.7
64.0* 22.6 64.0 -14.0 -14.0 -14.0

¹. Pout Output in Watts into 8Ohm.
¹. Vrms as adjusted at PST-out via Variac.
¹. Vg @ Rs = 10K and P1 = 20K3
². Vg @ Rs = 22K and P1 = 42K0
³. Vg @ Rs = 47K and P1 = 89K4

At Pout = 0.25W, Rs is chosen such that 30MR-LDR10M.
For R-LDR = ~10M, Rs = 10K @ P = 0.25W
For R-LDR = ~20M, Rs = 22K @ P = 0.25W
For R-LDR = ~30M, Rs = 47K @ P = 0.25W

* For a given Rs, at Pout = 64W, P1 is adjusted to gain Vg = -14V.

You may notice that at low listening levels there already is quite some
deflection. For 0.25W(!) this is 20%, 16% and 12% deflection respectively,
assuming Vg = -1V = 0% and Vg = -14V = 100% deflection. At a medium level of
4W it is 60%, 58% and 51% respectively. As stated before Rs and P1 values
depend on the specific Leds and LDR as used and have to be re-established
for any other combination, and for other output levels of course, but I
don't expect immense deviation from the values given.

Happy experimenting,

Rgds, Jan.



"Tom Schlangen" schreef in bericht
...
Hi RATs,

both projects I have on the bench at the moment
have a larger dynamic (read power) range than the
first amp I utilized magic eyes (for merely the
nice look of them when flickering).

You can see the circuit I used in the little
amp half way down the page at:

http://www.ndh.net/home/schlangen/ro...aten/em80.html

Useful reading range voltage at the grid of EM80 is
between 0 volts (shows thin line) up to ~ -18 volts
(full reading).

It is nearly impossible get a reading at normal
listening levels whatsoever from the magic eyes
when the max. reading (-18 volts at grid) is
adjusted for max. reading at max. output power
for a stronger amp.

A simple switch and a voltage divider gives a
"sensitive" range for low amp output levels
and a range for higher output levels, but
having to switch the ranges manually is not
practical.

Adding a zener so that at least the EM80 isn't
overdriven isn't optically nice and practical
either, since there is no change in readout up
from a certain amp output level.

Does some have an idea how to implement a
Real Simple (!) sort of dynamic range compression
or AGC, without making things too complicated?

At most I would dare to add one double triode for
_both_ EM80 present (one additional triode system
per channel).

Any ideas appreciated,

Tom

--
Live is too short to be taken seriously.
- Oscar Wilde













































  #11   Report Post  
Tom Schlangen
 
Posts: n/a
Default

Hi Jan,

all I can say is WOW ... because it even saw
real world implementation =:-)
But the principle is clear.

Tnx2u!

Tom

--
Knowledge is power - knowledge shared is power lost.
-- A. Crowley
  #12   Report Post  
Patrick Turner
 
Posts: n/a
Default



Tom Schlangen wrote:

Hi Jan,

all I can say is WOW ... because it even saw
real world implementation =:-)
But the principle is clear.

Tnx2u!


I have a logarithmic amplifier using a 741 opamp,
and a shunt fb path using a diode, and a resistor,
and this powers a meter which has a center zero point.
The log amp allows me to read +/- 40 dB easily in Db, rather than volts.

I use the meter for sound level readings from speakers, which vary
a lot, and to cope with variations, I need a log meter.
But such a simple arrangement could be used to
alter the grid volatge of a magic eye tube.

Patrick Turner.


Tom

--
Knowledge is power - knowledge shared is power lost.
-- A. Crowley


Reply
Thread Tools
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Listen to a 'borg and destroy your eyes Clyde Slick Audio Opinions 0 February 26th 05 09:29 PM
Not happy with the bass in my trunk. Help? Doug Car Audio 86 August 3rd 04 04:23 PM
The Eyes of Texas Are Upon OU! Schizoid Man Audio Opinions 2 October 13th 03 11:08 PM
Just hear Sherryl crowing 'Behind blue eyes' Rob Adelman Pro Audio 47 August 11th 03 01:09 PM


All times are GMT +1. The time now is 06:30 AM.

Powered by: vBulletin
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 AudioBanter.com.
The comments are property of their posters.
 

About Us

"It's about Audio and hi-fi"