Russian rod pentode circuits

[Jez]

Any mixer can be implemented as plug in replacement using 2x 1j24b, one as a triode oscillator and that can direct drive g2 of the RF in/Mixer. Less LO leakage than DK91, DK96, DK92 and 1L6.

Interesting! I made a superhet using Russian rod pentodes in 2021 (rather a lot, 13 in fact!). The frequency changer uses a 1j24b as oscillator, driving g3 of two 1j24b's in antiphase, while RF is applied to both g1's together. IF is available at the anodes, this driving a centre-tapped-primary first IFT.

I found that bias on g3 was fairly critical for best results - they're at about -10V (without consulting notes) and the LO swings g3's from -20V to 0V.

For what it's worth, the other ten rod pentodes are: Two IF stages; infinite impedance detector (yes! With directly heated, grounded filament); AF amplifier; phase inverter; push-pull output stage (this using 2 x 1p24b's in ultra-linear, giving 2W).

That brings up to ten... the last three then are a further IF amplifier, driving a leaky-grid detector, with anode bypassed to ground for everything from a couple of Hz upwards... the mean DC level here is used as AVC control voltage. The last pentode is a DC amplifier and inverter (with voltage delay), and the amplified control voltage is applied to g2's of the frequency changer and g2's and g3's of the IF amplifiers, appropriately level-shifted.

LT is regulated at 1.15V; HT at 120V and there's a -80V rail which I needed for the infinite impedance detector, and also potted-down biases the output stage and the AVC line. They're all derived from a heavily-screened flyback converter running from a 12V battery.

It works well (it's playing right now)!

Nice! Is that a short wave communications receiver?

[Kalee20]

... push-pull output stage (this using 2 x 1j24p's in ultra-linear, giving 2W).

Oops! Wrong number, these are 1p24b's, not (probably non-existent) 1j24p.

Mods - if you see this and can amend my previous post, do so and delete this one - thanks!

[Kalee20]

 Nice! Is that a short-wave communications receiver?

No - a LW/MW high-quality home radio.

I should also mention another project, actually a joint one with a friend (she did most of the concept work and built the bread-board amplifier; I designed and made an output transformer, and did a bit of optimisation). This basically was a low-power amplifier, just under 0.5W, using two 1j29b's in a self-splitting output stage and dubbed the Amīcamp: one valve's grid is driven; the other valve's grid tied to ground, and a single transistor constant-current source feeding the cathodes.

This was preceded by a 1j24b as voltage amplifier.

Being directed heated, the 1j29b's needed a floating filament battery. But, rather elegantly, they are biased to consume 11mA HT current which is exactly what the first-stage needs to heat its filament - so still only one LT battery is needed.

I did introduce NFB to the first stage cathode with twin tertiary windings on the O/P transformer, passing the heating current to its directly-heated filament while applying common-mode signal feedback. This worked well, and sound quality is good enough for serious listening.

Eventually, separate batteries were dumped and the whole thing powered from a little push-pull switching supply... trying AC, albeit high-frequency AC, on the output valve filaments (from a 0.6 - 0 - 0.6V winding on the power transformer was successful! We included a nulling system in case the HF appeared on the output, but (based on a sample of one!) it wasn't necessary.

Photo shows the amplifier module (the diecast box houses the first-stage 1j24b) and the power supply module, mounted on a baseboard with the battery. But sorry, I don't have the circuit diagram on my phone!

[Mike Watterson]

Sure they knew about inherent NF of resistances vs resistance when valves were still common.

If I want low noise audio I'd use one of the specialist low noise audio IC op-amps. However it's only mag pickups and some sorts of microphones that need that at input stage. Line level audio isn't so critical. But the only valve audio amps I've designed or built since 1970s was the rod pentode one for an FM radio I was building a few years ago. Repaired plenty of them for bands and restoring radios, record-players and tape-recorders. Only the better tape-recorders used an EF86 (often having ECC8x too). All the '50s record players ceramic or crystal replaced by ceramic pickups and those have a high output and don't need RIAA compensation.

Certainly real Pentodes are inherently noisier than Triodes, but the Rod Pentode is just electrostatic rod focusing, no spiral wire grids, so is much quieter.

Edit:
I remember the BC109 was sold as "low noise", but I don't think it was inherently lower noise than a BC108, just higher gain. I'm not sure that BC107, BC108, and BC109 and graded parts a, b, c for gain were not even all the same silicon originally but graded on test. I do remember that at the start it was challenging designing a low noise transistor amp even for guitar pickups and dynamic mics, but better than a 741 could be achieved.

[Kalee20]

Certainly real Pentodes are inherently noisier than Triodes, but the Rod Pentode is just electrostatic rod focusing, no spiral wire grids, so is much quieter.

I'd dispute that, though - g2 in the rod pentode does draw some current, so partition noise is possible.

It doesn't matter if g2 is rods, wire spiral, or perforated mesh - if it draws current, the noise mechanism is there!

(I'd concede that the current is less, proportionally, to anode current than in spiral grid pentodes, so the actual noise may be a few db less. But it's still there)

[Mike Watterson]

Hi.
Michael...... Did you ever get a 1J42A to oscillate by space charge?

Didn't try. I simply did some DC tests.

Things are a bit disrupted in Ukraine and esp. Odessa where one of the better suppliers where due to the daily unprovoked Russian attacks. I'm amazed they have managed to get post running again.

I'd be amazed if it was possible to buy Rod Pentodes from Russia (and there was only ever one guy in Moscow) due to sanctions on payments, shipping and post. The Bulgarian ones were more tarnished/corroded leads.

I've only enough 1j42a for one project.

I might be able to spare a couple of 1j29b. I have 3 kinds, but I think only differ in glass envelope and mica supports for what's translated as "reliability" or "endurance". There were variations on USA tubes such as 6146 with similar descriptions. Not sure what the variation was on E8x** types compared to ECC8x.

I might also spare a couple of 1j24b

Both only if ordering from Ukraine or Bulgaria fails. Allow for delays inc. from Brexit Customs even from Bulgaria. There maybe was also someone in Poland or Baltic states selling. I sold some to Germany years ago when I had the Web Shop.

[Mike Watterson]

Certainly real Pentodes are inherently noisier than Triodes, but the Rod Pentode is just electrostatic rod focusing, no spiral wire grids, so is much quieter.

I'd dispute that, though - g2 in the rod pentode does draw some current, so partition noise is possible.

It doesn't matter if g2 is rods, wire spiral, or perforated mesh - if it draws current, the noise mechanism is there!

(I'd concede that the current is less, proportionally, to anode current than in spiral grid pentodes, so the actual noise may be a few db less. But it's still there)

Yes, still noise. But Joe Sousa claims that in theory and practice it's less on a Rod Pentode than a regular kind. I've also no idea why the EF37 and EF86 were claimed to be low noise, but Jez is first person I've read claiming an EF86 isn't, or pointless. Maybe I could measure but I'd take Joe's statements.

I wrote "quieter", not quiet.

You get more noise from any electronic device than equivalent resistor, unless it's run in liquid nitrogen or something. Current noise figures on Ku Band LNBFs of 0.2dB are likely fantasy; they are probably cherry picking, also the difference between 1db and 0.2dB NF isn't very significant. You need a bigger dish if rain is an issue.
Similarly it's a truly rubbish HF receiver if receiver noise is an issue in last 30 years.

[Murphyv310]

The Rod Pentodes are insanely quiet, with one exception, microphony! Long leads can be the issue as can the way the valve is mounted.
I've wrapped foam around them and clipped the whole assembly to whatever board I'm using, even adding heat shrink to the wires gives a damping effect so microphony is eliminated.
I had microphony on my Pye BB which gave feedback on high volume, carrying out the above was a complete cure.
There is no hum, noise or hiss, checking with my Tek scope it's a clean as a whistle, so is my 3-3 stereo amp with the alleged noisy EF86's I use Telefunken ones.

[Mike Watterson]

Also magnetic fields. The Russians used steel sleeves. Most B7G and B9A valves you see have an outer electrostatic (mostly) shield. The tube or mesh you see isn't the anode. Earlier valves used metallisation. Many sets, and all commercial / military, used clip on spring loaded outer electrostatic screening cans (aluminium), and continued to do so for B7G and B9A, which also ensured rough handling didn't shake out the valves. Some valves used metal outer envelopes, especially later USA octal types with button rather than pinch bases, not just EF50 etc.

The 1S5 (DAF91) was very microphonic too.

The 1U5 has the same electrical characteristics but it uses an improved structure which greatly reduces any tendency toward microphonic effects. In addition, the diode unit is effectively shielded from the pentode unit to prevent "play-through". They are not interchangeable since pin allocation is completely different

—Radiomuseum

I've only seen the 1U5 (DAF92) in USA sets, despite it was released in 1945. The 1S5 was 1939, but B7G 1.4V types not used in UK till the 1946 Romac. The Ever Ready Model C used Octal in 1946 and wasn't B7G till 1947 Model C/A. AFAIK all UK sets & mainland European using Philips/Mullard/Telefunken designations used the DAF91 (designed by RCA as 50mA nickel filament 1S5) till 1953 approx. and the DAF96 (25mA tunsten). RCA designed the original nickel filament B7G 1.4V (release 1939 to 1940) based on Sylvania octal 1.4V. The UK edge connect 1.4V were based on the Sylvania Octal. Philips designed the 25mA tungsten 1.4V B7G series, released 1953. Mullard had been a Philips subsidiary from 1928.

I think all my post war USA battery valve sets use the 1U5. I have some early pre-1946 USA "lunch box" AKA "Personal Radios" and they all use the original 1939/1940 RCA types, so 1S5 (DAF91).

However even the DAF96 is prone to stray pickup (maybe not microphonic). Some Pye sets you have to test if more stable with screen on the DAF96 or the DL96. The DK92 and DL94 were still used up to at least 1957 when better HF (DK92, USA used 1L6) or more audio needed (DL94). One Philips set I have uses 2x DL94 in push pull and higher HT on mains (above Philips own spec!) for really loud audio. It uses a DF91 RF preamp and a second DAF91 as phase splitter. However a later version (with added DM70 Magic eye) uses DAF96 detector and DAF91 as phase splitter, a trap for the unwary. So as to get filament power for the added DM70.
Perhaps no-one in Europe licensed the less microphonic 1U5.
https://www.radiomuseum.org/r/philips_lx527ab_2.html

The Colette is more modest on audio using 2x DL96 in push-pull. However it filters grid voltage of the LO (set to about 6MHz on Gram/Tape input selection) and "wastes" a 2nd DAF96 as amp for the active tone controls. The DM70 is used as a triode power phase inverter and power on indicator, not magic eye. So as an experiment I wired the 2nd DAF96 as a phase inverter and the DM70 as magic eye. It didn't make any noticeable listening difference to the tone controls.
Like the sole Vidor VHF-FM & AM set, it unusually doesn't use a DK96 for AM. Instead the 2nd FM IF (DF97) is a Pentode mixer (LO via g2) and 1st FM IF (strangely a DF96) becomes an RF preamp.
https://www.radiomuseum.org/r/philips_co...62_ab.html

I think more strange things done with battery valve portables in the 1940s & 1950s than mains sets. Some German sets use a 1st IF of 10.7 MHz on FM and then the DK96 converts to 6.5MHz IF. Others simply turn off the the DK96.
Most ordinary non-Personal USA MW only (Broadcast) had a RF preamp, but UK and rest of Europe only added that on better multiple SW band sets.

[Mark Hennessy]

... push-pull output stage (this using 2 x 1j24p's in ultra-linear, giving 2W).

Oops! Wrong number, these are 1p24b's, not (probably non-existent) 1j24p.

Mods - if you see this and can amend my previous post, do so and delete this one - thanks!

I think I've got it right in your post and Jez's quoted version of the same, but do check

I've also sorted out the quoting in Jez's posts, hopefully.


Jez,

You're probably using the post editor in its default "WYSIWYG" mode. It's OK, but far from perfect.

If you're struggling to make it do sensible things, try switching to the "raw" mode. This is where you can see the MyBB code that you're probably familiar with from other forums (where you can use B or I in square brackets to make things go bold or italic, for example). Then you'll be able to see the Quote tags, and move them accordingly.

It could be just a case of not positioning the cursor at the very bottom of the post before typing your reply. But whatever's causing the problem, you can fix it in raw mode.

To switch to (what I've been loosely calling) "raw" mode, you just have to click one button. It's located just above where you type your replies - where there's 2 rows of icons. Look for the one on the bottom right, that looks like a text document. Hover your mouse over it and a pop-up should appear saying "View source".

This only applies when using the "full" interface for replying. If you just use Quick Reply, then you are using MyBB by default. It is possible to quote using Quick Reply, and that's probably easier once you've figured it out. Of course, when using Quick Reply, you can move to the full interface by clicking on the "Preview Post" button.

It's usually quite easy for me to sort it out afterwards, but as it's something that comes up fairly often, especially for new folk, it's worth mentioning from time to time.

There is a "sticky" that goes into this in more detail, with screenshots showing you where the buttons are: https://golbornevintageradio.co.uk/forum...p?tid=4638

Also just to remind everyone that if you prefer to work in "View source" mode by default, that is something you can change in your User Control Panel

Mark