diy IF transformers: first attempt.

[Kalee20]

I suppose the small signal handling will be OK, but "slew" with large inputs could be limited?

You got it.

When the signal is going positive, the valve can actively pull up the capacitance of the cable to the scope. But when the signal is going down, the most that can pull the capacitance down is the current in the cathode resistor... And if the combination of frequency and amplitude exceeds a certain figure, the input will overtake the output, the valve will 'let go' and there'll be waveform distortion.

I did see this, at 5MHz. I started with cathode resistor 180kΩ but had to reduce to 82kΩ.

The only way forward is to use a lower value of cathode resistor, and operate the valve at a higher current. But we can't risk grid bias reaching 0, or becoming positive, else we'll have grid current, spoiling the ultra-high input impedance.

Cure for this is to use a beefier valve... But that'll have more capacitance anyway, just what is not wanted.

Suddenly, the reason why high-bandwidth 'scopes are power-hungry, makes sense...

[Amie]

What's needed is a PNP valve!

Active pull down is what's in order though it gets complex if doing it with valves. I've seen operational amplifiers using hollow state, but if you'd have to go that far

Using sand for a load you could do better with a constant current sink than just a resistor.

Transformerless class A P/P stage possibly? May be interesting! But a pig to drive, and would definitely need separate filament supplies....

There may be something from the world of video signals processing that would be a help...

[ppppenguin]

What's needed is a PNP valve!

.....................

There may be something from the world of video signals processing that would be a help...

Many years ago I started to write an April Fool article for Wireless World based on the PNP valve. Never got finished. The physics is too hard to make even a plausible stab at it. Or maybe my imagination wasn't up to do a bit of sci-fi to get round the physics. Positrons or protons. Both are awkward. I was intending to publish under a nom-de-plume: "Anode Ray". To honour the great MG Scroggie who also wrote as "Cathode Ray".

To get good wideband amplifiers with valves was never easy. In 'scopes (especially Tektronix) the distributed amplifier was a complex and expensive (and hot!) way to get both gain and bandwidth. Invented, I think, by Percival at EMI in the 1930s. The White cathode follower helped to drive low impedance loads. Not sure if it was due to ELC White, another colleague of Blumlein at EMI in the 1930s.

[Kalee20]

Constant-current sink: Yes, would be better. The circuit as drawn approximates to that, with the large -ve rail and resistor. But a chip of impure sand mineral, either NPN or JFET, would be a bit more elegant.

What's needed is a PNP valve!

Yes. It's not a level playing field. Armoury of the sand/soot folk is more complete.

Attempts to cause heated cathode to emit holes into a vacuum is a logical absurdity... The only other solution, antimatter cathode emitting positrons is, I'm told, beset by either (a) expense of creating the necessary substances, or (b) the personal jeopardy in visiting anti-universe and returning with an anti-1j24b. And there's the practical difficulties that wiring a 1j24b and an anti-1j24b into the same circuit is likely to result in a flash, a bang, and the disappearance of both

[Amie]

Creation of a black hole with a couple of small valves and a battery supply.

Ultimate weapon!

[regenfreak]

Hi Amie
Very impressive work. Ed Dinning has introduced me to this forum and asked me to check your thread. 

I am a lazy so i used existing IFs from scrap AM radios becuase they are relatively  abandon.

Currently I am building a  valve FM superhet tuner with a triple gang air capacitor found in ebay and have encountered lots of difficulties getting hold of FM IF transformers.

If you want to build an AM superhet without searching in the dark, study existing proven designs; either using either European heptode or American pentagrid convertors. You dont need to re-invent the wheel. It does not matter if it is Hartley or Armstrong oscillator; the circuit is very simple. I am not familar with heptode supehet design but i know the design of All American Five very well and built one last year:


https://www.vintage-radio.net/forum/show...p?t=166316

I did not have as sweep generator at the time when I built this AM supehet last year. I simply measured the inductance and use LC resonance formula to pre-set the If transformers before installation.

Typical All AMerican Five schematics are shown here:

http://www.wa2ise.com/

Many european battery sets use 1R5 battery convertor tube running B+ arround 90V that can be supplied with a series of 9V battery. I owned and restored 6 All American Five AM radios (RCA, Zenith, Emerson etc)  so I know their design inside out; fundamentally they are all the same!!


The hardest part of building the superhet is the tracking problem. You either cheat by using an existing matched oscillator coil, padders and twin gang capacitors  taken from a scrap radio or do things the hard way like me, writing a three-point-tracking program to find the matched solutions for the antenna, oscillator coil and twin gang tuning capacitor. I spent a couple of months testing the different  three tracking methods from the classic radio textbooks (Langford Smith, Tehman and Henry) with my homebuilt supehet rig.

The second hardest part of building a supehet is mastering the art of  good RF wiring. The most common problem is the oscillation of IF amp stage which can be cured by good wiring, ground and RF by-pass caps.

A 10:1 wont load down sweep of 455/470khz with an isolation resistor. Sweep AM if is realtively easy but it is difficult to sweep FM IF transformers.

The static sweep of FM transformers usually does not produce accurate dyanmics behaviours after they are installed into the radio. At 10.7MHz, the  probe loading is significant even an isolator resistor of 100k to 200k is used. In addition, the Miller capacitance of the tubes will affect the tuning of the IF transformers.

Here is the thread that i was mucking around with stagger tuning of FM IFs;


https://www.vintage-radio.net/forum/show...?p=1317462

Currently I am figuring out how to sweep IF transformers using the NanoVNA. This requires matching the input and output impedance of IF transformers to VNA 50 ohms port. The resonance impedance of IF transformes are very high and vary sharply near resoance. it is very diffciult to design narrow band reactive matching pads for the VNA  frequency sweep. Maybe I will try balun transformers.

Good luck.

[Amie]

Hi.

Following written in installments forgive discontinuities

Well done on all your projects, and, thanks for the links, lots of usefull information..

If you want to build an AM superhet without searching in the dark, study existing proven designs; either using either European heptode or American pentagrid convertors. You dont need to re-invent the wheel. It does not matter if it is Hartley or Armstrong oscillator; the circuit is very simple. I am not familar with heptode supehet design but i know the design of All American Five very well and built one last year:

I have some downloads from American/World radio history of "most needed" circuits. Yes they are all very similar in design.

The reasons for the foregoing experiments are not just that I want to build my own valve AM superhet, but also:

*Have a method of replacing failed, unobtainable IF transformers: that's a box now ticked for the most part.

*Build an amplitude stable oscillator for use in superhet front ends. (I've noticed that a lot of commercial designs are poor both in amplitude stability and waveform purity, both degrade performance in the mixer).

*Use off the shelf or at least plentiful spares.

I have been thinking about creating a portable valve RF signal generator that would use Russian rod pentodes using these methods of frequency determination.

I would like to build a superhet radio exclusively with commodity pentodes, (think EF80/6BX6 )

I would like also to build an all triode superhet, something like the front end of the Grundig 80U, (I have an example), a  cascode IF, an Infinite impedance detector, AF amp, self split PP output. That's a possibility from currently available new production components: China production of 12AT7 are easily capable of 120MHz+ use, and in "new old stock" land, 7AT7 are relatively plentiful here and are not it seems attracting much attention of audiophools, or guitarists tube rollers.... yet.

My way ahead with the tracking problem is to work out what's needed, probably starting with the VC. Then choose the signal L and oscillator components from there, if experiments continue to give good results with the moulded chokes, then to make them adjustable, use the next lower preferred value and add a small adjustable loading coil.... By those means copying a prior design is possible too, without the actual components originally used being available.

I don't really see what I'm doing as reinventing anything, it's fun, educational, and  who knows, I may one day find a way to improve something!

Power supplies...
I usually experiment with things running as low voltage as possible, dry hands and <120v is acceptable for no bad nips. I build up all dry battery packs for my portable collection, one set uses 10×9v as you described and 1×F, one, an Emerson hybrid, uses four 12v (A23) and 3 1½v (AA)
The rest (all Ever Ready barring one HMV) use 60×AA and 4×D. All alkaline. Most of my ideas for battery packs came from Michael Watterson. There is also one running on a 6volt sealed lead acid through an excellent inverter, self built and lent to me by a good friend, I also have an inverter that I'd built, but that suffers from switching noise and needs better filtering/screening, though works well enough with a set with a ferrite loopstick.

The battery packs were all fitted with semiconductor current trips for HT and crowbar protected filament supplies within the sets themselves.

Thanks Amie.

[Kalee20]

 Ed Dinning has introduced me to this forum

Welcome! I've seen your posts on UKVRR

If you want to build an AM superhet without searching in the dark, study existing proven designs; either using either European heptode or American pentagrid convertors. You dont need to re-invent the wheel. It does not matter if it is Hartley or Armstrong oscillator; the circuit is very simple.

The 'standard' 4+1 valve superhet has a lot going for it, moderate gain at IF; moderate gain at AF; high gain overall but without being hugely prone to instability. Easy to make it work.

That said, it has room for improvement! The British / European triode-hexode is generally reckoned to be better than the American pentagrid or octode (less 'pulling' of frequency with AVC; less oscillator radiation - which user won't notice but everyone else does

I've noticed that a lot of commercial designs are poor both in amplitude stability and waveform purity, both degrade performance in the mixer).

I'd second that - and also that lots have rather crummy tracking, meaning sensitivity is poor; the AVC whacks up the gain, and circuit noise gets obtrusive...

A good oscillator, providing sufficient but not excessive output, will maximise sensitivity without giving rise to spurious responses.

A good IF response helps reject adjacent-channel signals without incurring treble loss. Again, sweeping the IF response of a 'standard' superhet is rather alarming...

The all-American 5 is a good wheel and it is almost guaranteed to go round and round, as-built. But it can be improved: better static and dynamic balance; precision bearings; pneumatic tyres; correct lubrication. There's lots to reinvent or improve.

It's turned out that I'm carrying on similar activities to Amie - my aim is to make a no-compromise AM superhet, not make a working superhet with the least amount of trouble. I've just finished developing push-pull output stage based on Russian rod pentodes (see another thread); I've got an infinite-impedance detector lashed up (working but not optimised) using a directly-heated valve; and I've just started on the IF amplifier, and Amie's well ahead with oscillator that I've barely thought about. My IFT's are currently based on ferrite RM cores. Why? Because I want to, and I can... if they are better than the traditional IFT I'll put up results; if not, I'll put up results anyway

Finally (regenfreak), I'm interested in your own FM IF work, I hope you keep us informed

[regenfreak]

I would like also to build an all triode superhet, something like the front end of the Grundig 80U, (I have an example), a  cascode IF, an Infinite impedance detector, AF amp, self split PP output. That's a possibility from currently available new production components: China production of 12AT7 are easily capable of 120MHz+ use, and in "new old stock" land, 7AT7 are relatively plentiful here and are not it seems attracting much attention of audiophools, or guitarists tube rollers.... yet.

Hi Amie
I have not logged on since Feb. I was so bad that I forgot the name of this forum. 

  In my early days, I started off building a minimalist MW superhet with only a single IF stage (somewhat deaf)  using a chain of cheap 9V batteries like yours. 

12AT7 is a good choice for oscillator. It is similar to ECC85 used in FM superhet. The 12AT7  was used in the Heathkit RF-1 signal generator kit, schematic here:

https://www.vintage-radio.info/heathkit

I was building my FM superhet tuner but encountered problem using ECC85 as an oscillator:

https://www.vintage-radio.net/forum/show...p?t=178722

I could not understand why the oscillatior output was so weak in spite of repeated experiments with different coils, feedback tappings etc. However I got the tracking perfect though. I built a minimalist FM superhet tuner with a 6SB7(RF , mixer and oscillator all in one). It works very well. To investigate the problem with ECC85, I will build an oscillator test rig so that i can isolate other factors like components interaction etc.  I have a pair of cheap Russian transmitter values 6C11 (EC81) from Aliexpress. They are good up to 750MHz. I am going to test against the ECC85. The 6C11 works really well as an oscillator in my chinese stereo FM kit. I have been doing other homebrew projects at the moment. So it will be my another "to do list" 

Happy experimentation!

Welcome! I've seen your posts on UKVRR

Good to see you here. Thank you

[Amie]

Hello, regenfreak!

I'm not near any relavent information at the moment, but experiments with twin triodes designed for cascode amplifier use do seem a little reluctant to oscillate, on one occasion particularly in my case because I tried to use the section in a type that has its internal screen connected to its grid! I don't suppose you have done anything that silly One aside that worked surprisingly well, was while investigating superregenerator circuits, that they don't take a lot of modification to produce an acceptable self oscillating mixer for a low IF pulse counting FM receiver ...

The home brew IF transformer experiments I will claim as a resounding success, however trying the same shortcut to make a local oscillator for a superhet has proved difficult as the available range of values don't coincide with anything that will allow decent tracking... We can't have it all though.

Thanks for sharing your ideas.