I was wondering if ya'll could help me with an issue that I am having
with a regen receiver that I built.

I built the usual regen circuit with two stages of audio coupled with
interstage transformers. The regeneration is controlled by a pot which
varies the plate regen voltage. The grid leak combination is a 6 Mohm
resistor and 45 pf cap.

The regen and first audio stage tubes are 1H5GT triodes. The second
stage audio tube is a 1C5GT pentode. The filament voltage is 1.4
volts, plate voltage is 90 volts and screen voltage for the 1C6GT tube
is 90 volts. The set easily and cleanly drives a speaker.

I have wound my own coils which cover the broadcast band up to about 15
Mhz.

I really like the way the set operates and it is very sensitive.

My problem is that the regeneration often fades out (the radio
essentially goes dead) and I have to touch the grid of the regen tube
to bring it back in. I had originally built this radio only having one
stage of audio and I did not have this problem with that version. It's
kind of like the charge on the grid builds up too high and I have to
"drain" it off with my finger. Sorry for the bad description--I'm
obviously not an engineer.

I did try various other combinations of the resistor/cap for the grid
leak including the classic 2.2 Mohm/150 Pf combination and it still has
the same fade out problem. Therefore,
I decided to stay with the higher resistor/lower cap value because the
receiver works better at those values (when it is working).

I usually power the radio with a battery eliminator, but the same
problem exists even when I use only batteries. For a ground, I am
using the mains ground.

I'm hoping that you can give me some ideas to fix the problem for what
is otherwise a great receiver.

Thanks to all of you in advance.

Dan

Hi Dan,

Have you checked that your 6M ohm grid leak is actually 6M ?

It does sound like a charge build up on the grid. Does the RF anode
current rise or fall when the radio goes dead ?

Are there any components attached to the grid that could be breaking
down slowly - capacitor, maybe even valve base.

Have you tried replacing the valve base ?

Have you tried a different valve ? Maybe broken or cracked valve bottom
(paxolin to valve pin).

Stick the whole unit in the oven (on low) and dry it out for 3 hrs -
moisture getting in somewhere ?

Does sound weird and interesting.

Regards,

Mark.

Hi Mark,

The resistor is definetly 6 Mohm, but you bring up an interesting point
about moisture.

I have noticed that the set seems to fade out more during rainy or high
humidity days. I didn't even want to mention it because it sounded
silly, but I did wonder if humidity might be a factor.

On some clear, dry days, I have not had any trouble at all. So, maybe
you have something there.

I will carefully go back over the items you mentioned and maybe replace
all of the caps in the circuit. Replacing tubes hasn't made a
difference. Thanks.

Thanks Allison,

All of those are great suggestions and I am going to carefully clean
all of my soldered connections to see if it makes a difference in the
grid leak.

Dan

Dan wrote that the regeneration in his receiver is controlled by a pot
which varies the plate regen voltage. I saw this method was popular
mainly in old books (ARRL from 30's and 40's for example) but later the
most popular became a control on the screen grid. Does it mean, that
the second method is better ? Why ?
Sorry if my question is naive but I'm beginner, especially in tube
stuff.

"Dan wrote that the regeneration in his receiver is controlled by a
pot
which varies the plate regen voltage. I saw this method was popular
mainly in old books (ARRL from 30's and 40's for example) but later the
most popular became a control on the screen grid. Does it mean, that
the second method is better ? Why ?
Sorry if my question is naive but I'm beginner, especially in tube
stuff. "

Actually my regen tube is a high mu triode (1H5GT), so there is no
screen voltage to control. A lot of folks like to use throttle caps to
control regen, but pots are easier to mount and don't take up quite as
much space. (I know that a lot of people will disagree with me here on
my choice).

My second audio tube is a pentode (1C5GT) and I use a one meg pot on
the grid to control the volume for the set. I also have a 10 ohm pot
on the first audio tube (1H5GT) filament which also helps control the
volume and nulls out hum.

In case you are wondering why I chose these tubes, they are the tubes
that were used in the old battery sets and have low battery drain
(although I use a battery eliminator anyway for my power source). They
are also very cheap and plentiful.

I also was able to "steal" the engineering from the old battery radio
schematics to figure out the capacitors and resistors to use in the
audio circuits to make the output sound good. I'm basically too lazy
to learn the formulas to do my own calculations. The set sounds as
good as any battery set that I have ever heard.

This particular combination of a 1H5GT first audio stage tube followed
by a 1C5GT second audio tube was used in a lot of battery sets. It
also made selecting the output transformer easy to maximize output to a
4 ohm speaker.

Dan

My mistake, it's a triode, no screen grid. I'm not used to American
symbols, here in Europe we have easier sytem (EC-triode, ECC-double
triode, etc.). For me 1H2GT and 1T4T is the same tube
What are the differences, advantages and disadvantages between methods
with a pot in plate circuit, screen grid circuit, throttle cap ?

The regeneration control either varies the amount of feedback
(throttle cap, movable tickler coil, or pot across the tickler coil)
OR varies the gain of the valve (or solid state device) by varying the
plate voltage (pot or series resistor), screen voltage (consider the
screen of the tetrode or pentode as the plate of the triode and the
plate of the tetrode that of a cascaded second stage), or grid voltage
(yes you can also add negative bias to the grid to lower the gain,
rarely used in a regen using tubes but the ONLY way with FET's).

If the variable gain method is used, then the tickler coil must have
JUST the right number of turns so there is JUST the right amount of
feedback for regeneration to start when the valve (or other active
device) is operating at the optimal gain setting for detection. If
TOO much feedback is provided the gain of the detector must be set
so low that poor sensitivity results. Also adjustment of the regen
will be tricky. If too little feedback is provided regeneration may
not happen or the detector gain will have to be so high that the
detector will be unstable.

For some reason most triodes seem to operate as detectors with
about 30-50 volts on the plate, and pentodes with 20-50 volts on
the screen. Using a regulated supply for the detector is a good
idea, especially if you want to try and receive SSB signals.

The fade out problem with your regen sounds familiar, I had this problem
on an old AA5 radio. It would stop playing after a while and came back
if I touched the signal grid lead (antenna tuning cap hot side) of the
12BE6 tube (I actually held one end of a .01 cap and touched the OTHER
side of the cap to the grid as it was a hot chassis set). Weird!

Maybe I had a bad cap in the agc circuit, not sure.


wrote:
On 29 Apr 2006 06:26:08 -0700, wrote:


My mistake, it's a triode, no screen grid. I'm not used to American
symbols, here in Europe we have easier sytem (EC-triode, ECC-double
triode, etc.). For me 1H2GT and 1T4T is the same tube


What are the differences, advantages and disadvantages between methods
with a pot in plate circuit, screen grid circuit, throttle cap ?


Apples and potatoes. Seriously screen grid valves(tubes) of the
tetrode/pentatode flavor are different animals than triodes so
applying comparison of regen throttle techniques across them
is difficult to compare. However.. with that said.

Throttle cap is usally the best as you can optimize the tube operating
point and then only adjust the amount of feedback. Disadvantage is
there is RF on the capacitor and that means careful layout and good
mechanical mounting for stability.

The other two vary the plate(anode) voltage and vary the screen
grid voltage are similar as they both affect the gain of the tube by
altering it's operating point. Disadvantage is that you may have
to play more with the feedback winding (or tap) so that regeration
occurs at the best point on the variable resistor or performance will
suffer. Advantage, no RF on any of the wires to the resistor so
placement and mounting is more flexible.

Both can be made to work equally well with care.


Allison

wrote:
On Sat, 29 Apr 2006 18:01:58 -0400, ken scharf
wrote:


The regeneration control either varies the amount of feedback
(throttle cap, movable tickler coil, or pot across the tickler coil)
OR varies the gain of the valve (or solid state device) by varying the
plate voltage (pot or series resistor), screen voltage (consider the
screen of the tetrode or pentode as the plate of the triode and the
plate of the tetrode that of a cascaded second stage), or grid voltage


(yes you can also add negative bias to the grid to lower the gain,
rarely used in a regen using tubes but the ONLY way with FET's).


I found that was not the best with fets. However, fets work best when
run at best gain and the amount of feedback (regeneration) is
controlled via pot or throttle cap.


If the variable gain method is used, then the tickler coil must have
JUST the right number of turns so there is JUST the right amount of
feedback for regeneration to start when the valve (or other active
device) is operating at the optimal gain setting for detection. If
TOO much feedback is provided the gain of the detector must be set
so low that poor sensitivity results. Also adjustment of the regen
will be tricky. If too little feedback is provided regeneration may
not happen or the detector gain will have to be so high that the
detector will be unstable.


Mostly what I said.


For some reason most triodes seem to operate as detectors with
about 30-50 volts on the plate, and pentodes with 20-50 volts on
the screen. Using a regulated supply for the detector is a good
idea, especially if you want to try and receive SSB signals.


It does help. I have a 5 tube 80m RX I built that uses a regenative
detector. I didn't need the gain but for selectivity using very loose
couped IF. Stability was hard to maintain without a regulated
screen (6BA6 detector) source from interactions of agc and audio
amp pulling the supply. Once I nailed it down it performs extemely
well and is a bit narrow for AM phone with only 2 IF cans.

Also for radios using tubes with directly heated filliments dropping
the filiment voltage can tame the regneration a bit.


The fade out problem with your regen sounds familiar, I had this problem
on an old AA5 radio. It would stop playing after a while and came back
if I touched the signal grid lead (antenna tuning cap hot side) of the
12BE6 tube (I actually held one end of a .01 cap and touched the OTHER
side of the cap to the grid as it was a hot chassis set). Weird!


The problem was likely an open antenna or oscilator coil, I've seen
that before. The coil opens (broken connection) and the grid floats
till its self bias drives it nuts. Inject noise and you bleed off the
charge for a while.

I suspect the 6M grid resistor is too high for the tube used and it's
building space charge bias till it cuts off. I tried a regen using
some 1AD5 (submini 1.5V filliment tube) and it didn't like grid
resistors over about 4meg and it behaved the same way.
the solution was lower (2.2m) grid resistor and 100PF cap.

In general using a higher grid resister and smaller capacitor results
in increased selectivity and sensitivity with a loss in large signal
handling ability. Visa Versa for smaller grid resister and larger
coupling capacitor. Also the grid resistor can go across the capacitor
or from the grid to ground. If the grid coil's cold end is NOT grounded
the latter arrangement is required. In early tube manuals the
recommended grid leak resistance was usually 1-2 meg ohms, with a
capacitor value of 100-250pf. You have to watch the time constant
there or you might end up with 'super-regeneration' where the tube
breaks into a high frequency (ultra-sonic that is, around 10khz)
oscillation.

"Other tubes I've used (12BA6) that resistance could be 10meg it
behaves
well where with a 12EK6(12v space charge tube) was happiest
with 5meg."

I'm thinking of putting a 10 ohm pot in series with the filament of the
regen tube to see if making the filament less negative in relation to
the regen grid might work, maybe also using a bypass cap across the
pot. I don't have one right now but I'm going to play around with
different resistor values (like around 3 ohms) until I can order one.

Did you think about doing something like this when you couldn't get
your 12EK6's to work with a 10 meg resistor?