Featured
Latest
Triode-Pentode Modu...
 
Share:
Notifications
Clear all

Forum 141

Triode-Pentode Modulator

13 Posts
4 Users
0 Likes
3,508 Views
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Or Triode-Pentode Oscillator - Mixer

I thought I would re-check my work before shipping off a kit of parts. Which also allows me to document it.

The Russian Rod Pentodes can ONLY exist as "pentodes". Real Triode, Tetrode, Heptode, Octode or Hexode valves are not possible. This is because the g1 & filamentary cathode forms a "gammatron" like structure causing the electrons in the "space charge" of the filament to be squirted as two slice shaped beams at right angles to the pair of g1 plates either side of the cathode filament.

The g2 and g3 are not a real screen grid and suppressor grid at all. They are rods either side of the beam to form an electrostatic lens to focus the beams on the two anode plates on either side

Cross-section Diagram of the two beams and all the electrodes
(not to scale at all)

Side view (not to scale)

Almost to scale 1j2b4 top view (similar to cross-section)

There is a spring at the top to tension the filament as it changes length from cold to hot!
Unlike the B7G, Rimlock, Loctal and Noval / G9A the tubes have no complete internal screen. Just a few rods to balance the capacitance of the filament support rod and add extra mechanical support. They provide a slight screening action. These connect to the "sh" Shield or screen pin if present. On some Rod Pentodes they connect either to "f-" or "f+". On the highest performance types they have a separate connection to somewhat reduce inter-electrode capacitance.

Since there are no grids in the conventional sense and only the edges of the pair of "g1" plates "face" the Anode but are close to g2 rods, the "miller effect" is lower than similar capacitance conventional tubes. Partition noise as created mainly by g2 and slightly by g3 and g1 is virtually non-existent as there is NOTHING blocking the smooth flow of electrons from the cathode filament to the anode plates.

Similarly as the g1 can't be a variable pitch spiral or vary in distance from "cathode" filament, a "remote cut off" or variable mu Pentode isn't possible. All the Russian rod Pentodes are sharp cut off. This would appear to make self mixer/osc and AGC nearly impossible.

Actually one solution is the 1j37b. Its two g1 plates come out on separate pins. There were others (1j42b) like this but not available. This makes it into a bizarre and unique hybrid of Gammatron/Wunderlich and Pentode. It's almost the exact opposite of a differential amplifier. It amplifies common mode signals more than differential ones. So it can be used as mixer or for AGC.

However there are other ways of varying gain and thus having AGC or a Mixer.

My first experiment was to use g2 as an Anode and then the actual Anode and g3 as subsidiary controlled amplifier like a Hexode, Heptode or Octode. This only works at very very low Anode voltages, about 6V! In fact the device will work in a Heptode socket as Oscillator / Mixer where g1 & g2 is the Oscillator, g3 is modulation or RF in and Anode is output. On an existing circuit a large resistor in series with anode bypassed by capacitor will reduce the 60V to 85V to 6V. The output impedance is very very low though, so as a frequency changer it's poor.

As a Modulator you can't have more than about 50% AM, otherwise the oscillator may stop! Also you will get a strong FM component. All Hexode / Heptode / Octode modulators have this problem as well as Oscillator leakage out of the RF in connection.

Best reliable modulation depth with a single Rod Pentode as Oscillator and Mixer.

A single 1j24b Rod Pentode wired to plug into a DK91, DK92 or DK96 socket
(the DK92 / DK96 screen isn't connected so this does work in a DK91 socket, the DK91 shorts out HT in a DK92/DK96 socket and DK92/DK96 don't work in a DK91 socket).

Anyway using this scheme you can make a single tube modulator / oscillator, but it's not very good and while it can run off about 12V, the supply voltage dramatically varies modulation depth.

The "Classic" AM modulator is just an Audio amp with its transformer secondary in series with the RF P.A. (or Oscillator, but not as good) HT rail!

But for most non-professional applications the Triode Hexode was the supreme solution to Mixer/Osc. The Heptode really remaining only for battery sets were it saved a filament. US mains radio designs used the Heptode longer than in Europe.

Putting the triode and hexode inside one tube and internally connecting the electrodes reduced Oscillator Leakage.

So how can we use the Rod Pentode as a Modulator / Mixer even with a separate oscillator?

Varying g3 only works at very low Anode volts, below the knee = poor performance and low impedance.
What does g2 variation do? It should vary the focus much more even at high anode voltages

Which it does.

 
Posted : 21/09/2012 5:53 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Here is the Triode Pentode Oscillator Mixer

(click to new tab with middle button to bigger it)

If the v1 and V2 were a real triode and real hexode and the output coils an IFT, with a tuned aerial loop at IN, we would have a classic superhet front end instead of a Modulator (Oscillator + Mixer).

Though the Rod Pentodes are not "real" pentodes, you can sort of triodise them by strapping g2 to Anode. The g3 must still connect to 0V/Cathode/filament, or at least much lower than Anode voltage unlike DF97 and some other real pentodes where g2, g3 and Anode are all connected. There is a 3rd "more accurate" simulation of a triode using a Pentode.

Why Triodise it? Because you can run the oscillator to a higher frequency! Also the Pentode mode is like a constant current source. If we use 6 x PP3, that's 54 to 56V fresh and 36V at "end point" of flat batteries. But the g2 of our modulator needs a fairly constant DC voltage. The Triode mode is almost a square law IV on the "anode" rather than constant current, so there is slight voltage regulation!

If the supply volts is increased the R2 needs increased too keep the V1 current the same (450uA to 700uA) and the V2 g2 volts about 40V.

No problem getting 100% modulation depth!

The filaments are in series and at 2V (flat batteries) the filament current is 10mA. With fresh Zinc Carbon batteries the LT is 3.2V and current is 13mA only! So about 90 hours on Zinc Carbon AA and close to 200 Hours on Alkaline AA. A pair of Alkaline D cells maybe about 1300 hrs! About 450 hrs on Alkaline C cells.

The HT is about 1mA (varies as batteries age from 56V down to 36V) so about 490 hrs on 6 x Zinc Carbon PP3 or maybe 650 hours on Alkaline. Less than 180 hours on NiMH PP3 (7 needed). Or use Zinc Carbon AA x 36 for about 1000 hours!

The Modulation depth increases slightly as the batteries age and overall carrier level falls a little.

 
Posted : 21/09/2012 6:27 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Two of these circuits will make a 405 line modulator. The 1j24b works to about 200MHz, but you would want 90V HT for reliable operation with flat batteries.

If you are clever you'll figure there is even a way to have all four filaments in series. Though using 3V is more reliable.

 
Posted : 21/09/2012 6:37 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Mounting the valves, filament series resistor and grid resistors

Double sided tape holds the scraps of veroboard till the earth straps are added. All "earth" / 0V valve wires are soldered to the coffee tin. The lacquer lable solders well with a hot 25W iron. Pre tin where wires will be soldered.

Lidl Espresso cans are great for screen boxes, screens, small chassis, home made pins or sockets. Pierce with stanley knife and cut with scissors.

Another view

Part II Testing the oscillator on its own (triodeized Pentode)
Rest of parts added for oscillator. V1 g1 bias moved from -3V to V1 f- for lower voltage operation. At 90V HT, it can be on -3V.

Adding the mixer Pentode

Fresh batteries. 4 simulates 6 flat batteries. Works fine on 4 or 6

 
Posted : 21/09/2012 8:15 pm
crustytv
(@crustytv)
Posts: 11869
Vrat Founder Admin
 

Absolutely fantastic project and what a wonderful thread, thanks for sharing it on VRAT. I wish I had something to contribute other than that, If my home-brew skills ever get developed I would love to have a go at the 405 modulator you mention, maybe one day.

Chris

CrustyTV Television Shop: Take a virtual tour
Crusty's TV/VCR Collection: View my collection
Crustys Youtube Channel: My stuff
Crusty's 70s Lounge: Take a peek

 
Posted : 21/09/2012 8:35 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

If I had a 405 TV ... A small one :)

A 405 line Modulator is just two of these with simpler coils. Or crystals. The sensitivity is about OK to feed it direct with video but I think by default needs an extra valve to invert it. As it stands I think it does negative modulation like 625 (no problem here were ALL 625 sets tune down to 42MHz. Even the "approved" DTT setboxes and TVs have to manage Band I, Band III and Band IV/V. However, the video can be DC coupled for negative modulation!

 
Posted : 21/09/2012 10:28 pm
crustytv
(@crustytv)
Posts: 11869
Vrat Founder Admin
 

If I had a 405 TV ... A small one :)

Get a Perdio Portarama nice little 8" 405 TV, 5th set down here https://www.radios-tv.co.uk/?q=OTT405 . Why not place a wanted ad, you never know someone may have one the are willing to part with. They come up on e-bay from time to time or get over to the NVCF you're bound to find one there.

I think Trevor has one in bits without a CRT, you could build it back together. Guess what I have a CRT (AW21-11) and scan coils for one which funnily enough, I bought from a chap in Ireland.

Chris

CrustyTV Television Shop: Take a virtual tour
Crusty's TV/VCR Collection: View my collection
Crustys Youtube Channel: My stuff
Crusty's 70s Lounge: Take a peek

 
Posted : 21/09/2012 10:33 pm
Refugee
(@refugee)
Posts: 4055
Noble Member Registered
 

It is a good test circuit and if i go for a modulator i would also need to do a receiver to go with it as i do not have a telly license. I do have CCTV and could do a home brew wireless system.
If you look at the picture valve thread in TV the camera PCBs i have are fitted with a 10MHZ video over power transmitter that produced awful results and for what i am thinking of it would have to get well off the short wave bands or be replaced with an emitter follower like it was last time i fired it up into a CRT with PL259s on the back.
I am going to look into fitting a high speed rectifier to an old tape recorder bias oscillator for HT and running the whole lot from low voltage batteries.
I have got a box with quite a lot of bits from a Rank language school that was pulled down in the 1980s with plenty of OC71s and OC81s in it.
I was thinking of a C cell and a number of AA cells with the heater tap from the C cell.
They were made for NiCad but NiMh should be fine for LT.
This will make the power arrangement simple.
Then we can compare them with the ordinary pentodes i have got waiting to play.

 
Posted : 22/09/2012 1:21 am
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Checking the HT feed resistor on the oscillator:

       HT   12K  15K  22K  47K
Low    40   33   32   30   25
High   58   48   47   43   35
change 18   15   15   13   10

This is testing with 4 or 6 fresh PP3 cells.

25V is as low as we want for screen/ g3 volts on the Pentode!

So 47K is optimal. This results in 850uA total HT current at 58V and only 10V screen volts variation for 18V HT variation

Checking the coil.
Ideally wind two identical coils, one for oscillator and one for Pentode Mixer/Amp and use a dual gang MW cap with matching sections.

I wound another coil, this time two very fine layers of 25T for primary and about 100 Turns for secondary (should be about 150T) This gave about 1.5MHz as the centre frequency, so 150 to 200t is needed on a 6mm former (wave/random wound).

What about a MW aerial coil? I had one off a transistor set with a smaller isolated winding (blue and yellow wires) to drive the low impedance transistor base. So I put that as a the primary on the screen/anode to HT 47K ohm decoupled feed resistor and the main tuned winding (White and Black wires) to the variable capacitor. Worked fine on second attempt when I reversed the blue and yellow wires.

I also reduced the grid capacitor to 22pF from 47pF.

There is no doubt that for a Superhet "Rod pentode" Radio all the screen grids could be driven from a decoupled HT feed for the Triodised local oscillator to reduce the variation of gain with battery ageing significantly. I'll examine the circuit at 48V to 77V (8 x PP3 or 48 x AA cells) which I think is ideal HT for a Rod Pentode Radio.

I'm also wondering if 2 x 1j24b in class B with feedback to screen grids from anodes would be a worth while HT current saving for Audio out (10mA to 13mA @ 3V!). It would only be about 300mW compared with 450mW or 1W for a single class A 1j29b. They would start in class A, but you'd use a rectified feed from local oscillator to reduce the bias.

 
Posted : 24/09/2012 12:05 am
Kalee20
(@kalee20)
Posts: 256
Reputable Member Registered
 

Similarly as the g1 can't be a variable pitch spiral or vary in distance from "cathode" filament, a "remote cut off" or variable mu Pentode isn't possible.

I've been thinking about this - couldn't the grid side plates be slanted? Then they would have different effect along their length.

And again, even assuming that the grid-voltage / anode current characteristic is linear, there comes a point where the current is cut off. But, the cathode has a voltage gradient along it, because it's got to be heated! So, one end of the cathode will cut off before the other. And when one end is cut off, that portion drops out of the circuit, thus the remainder behaves as a proportionally shorter, lower gm device. Result: a linear Ia / Vg characteristic which changes to a curved characteristic as cut-off is approached.

 
Posted : 24/09/2012 8:50 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

Yes, I thought of that. In fact the DM70/DM71 has a slanted cathode filament.

But it's "nanotechnology" the spacing of the two grid plates is VERY critical and very very close to the Filament. As was proved in 1930s & 1940s this concept doesn't work even with an indirect cathode as the electron source is too broad!

Only glass is missing ...
(middle click to bigger it)

The filament has 1.2V the G2 is 40V. There is a variation of grid to cathode voltage, hence the cut off isn't quite sharp (you can get some AGC action!)

Pentode Mode Ia vs Va with Ig2 dotted (fixed Vg2 and various Vgk)

You can see the "current mirror" action very dominant from 0V to 7V Va, and yes you CAN run this valve at +1V Vg at Va less than 20V!

"b" on graphs is Cyrillic Б so is "V" for Voltage
Here is the 2W (800W peak) 1p24b in 3W triode mode (simply tying g2 to A, there is a more elegant Triode circuit) Russian Data Sheets

1p24b "simple" Triode mode

My own measurements

and 1p24b as Pentode

Here is short article showing how to improve simple Triode mode of 1j24b (i.e. earth g3, short g2 & anode)
http://www.radiomuseum.org/forum/1j24b_ ... tions.html

The 1j37b and 1j42b etc developed probably because there was no mechanical way to get true variable mu/remote cut off. You can use the pair of g1 on 1j37b to control mu. Hence a more effective AGC or mixer

Summary Data

 
Posted : 24/09/2012 10:01 pm
Refugee
(@refugee)
Posts: 4055
Noble Member Registered
 

I think they have tried to stop that by keeping the heater voltages as low as they can.
With western valves they just make the grid oval and the cathode round in order to get the MU control.
They would need wedge shaped g1 or g2 rods perhaps?

 
Posted : 24/09/2012 10:18 pm
Anonymous
(@anonymous)
Posts: 16832
Group Deactivated Account
Topic starter
 

no, the oval grid is to have the supports at the side and essentially two opposite "beams" and then cutouts on the sides of the anode, taken to extreme on Beam tetrode. The variable mu / Remote cut off is done by having the grid wire spacing increasing/ decreasing

One I took apart earlier ...

EF89

Oval grids, and supports at right angles to two main electron beams, essentially two anode plates

Close up of G1

Spacing is varied so as to give better AGC action (variable mu / Remote Cut off). It's wider gaps on the wire "turns" in the middle I think. A bit hard to see. But varying the pitch is definitely how it's normally done.

On a beam Tetrode the g3 is replaced by two pairs of plates on either side of the g2 in between the oval g2 and an anode "plate" either side of the valve.

The Grid to cathode spacing is TINY on a rod pentode and of course construction not remotely similar

I don't have a clear photo of the 1j24b filament and two grid plates (my camera isn't good enough!).
Here is the significantly larger 1W 1j29b which has three g1 plates and two filaments
Grid plate and filament layout
Anode
g3 g3
g2 g2
|.|.|
g2 g2
g3 g3
Anode

(These 1j29b photos from Joe Sousa, USA Radiomuseum.org member and senior designer in Linear Technology)

The spacing of the two grid plates ether side of the filament is smaller than the EF89 cathode tube!

 
Posted : 24/09/2012 11:00 pm
Share: