Can somebody explain how these diodes work as detectors and AGC

[LincolnG]

So in this set we have an eaf42 and an ebc41. The diodes on the ebc41 are connected together. The diode ebc41 is connected to the diodes eaf42 via 50pf. This all connects to the AGC, slowed by 47nf, moving the bias ECH42, EAF42 and EL41.

Where are the diodes getting the signal from, where does detection happen, what is actually providing the source for the volts on the agc?

This is hurting my head, it looks like the diodes aren't connected to anything in the signal path.

Please can somebody explain how this works?

[paulsherwin]

The EAF42 diode is doing the delayed AGC. It's getting its signal from the final IFT via the 50pF cap. It's an odd design, and the circuit diagram doesn't follow normal British draughting conventions.

The strapped diodes in the EBC41 are doing the audio detection.

[LincolnG]

Originally Posted by paulsherwin

The EAF42 diode is doing the delayed AGC. It's getting its signal from the final IFT via the 50pF cap. It's an odd design, and the circuit diagram doesn't follow normal British draughting conventions.

The strapped diodes in the EBC41 are doing the audio detection.

It wouldn't follow British conventions, it's not British.

The listed schematic for the radio was not what was on the chassis, it was very different. I cut and pasted the original schematic to fit what was actually there.

I now understand how the AGC works. However, the detection, not at all How does the detected audio effect the grid of ebc41?

[G6Tanuki]

The detector diode load is the 15K resistor and the 750K volume control in series, with the 100pf capacitor from their junction to earth acting as a filter to stop IF signal getting fed to the ebc41 grid.

Bias for the ebc41 triode comes via the 1m resistor , generated from the resistor network in the path to earth from the power transformer centre tap.

[Silicon]

Apologies if you know this already.
The cathodes of all the valves are connected directly to chassis.
To Bias them correctly, a negative bias needs to be generated and then connected to each control grid.

The AGC network is connected to the negative bias network to get the advantages of 'delayed AGC' (a voltage delay).

[LincolnG]

Thank you, I understand that much better.

The 50pf providing the signal for the AGC (EAF42 diode) is, I would imagine quite critical. There was a dogbone style there which got broken, I replaced it with a 47pf MLCC (my meter tests 50pf. With that, on a small antenna it works ok, but when I put it on the big outside antenna (which works fine on my other radios), there is a lot of buzzing around various points of the dial - even more so if I switch on the antenna amplifier.

Is an MLCC not good in this position? Do I need to find another Dogbone, or try to repair the broken dogbone?

Is the value of this cap. actually too high or too low (I have loads, can manually test an alternative), if you think the MLCC is OK.

Bear in mind that there is a great deal of strange buzzing signals on the short waves, which I can only assume are there for some obscure reason!

[LincolnG]

Originally Posted by lesmw0sec

Bear in mind that there is a great deal of strange buzzing signals on the short waves, which I can only assume are there for some obscure reason!

Obviously I am talking about behavour that wasn't there before the 50pf was changed.

[paulsherwin]

I wouldn't expect small variations in the value of that capacitor to make any difference - all capacitors have a tolerance. You should probably check the voltage on the AGC line and confirm it's varying as it should.

Originally Posted by LincolnG

Obviously I am talking about behavour that wasn't there before the 50pf was changed.

Sorry, but it is far from obvious. If you are increasing the gain either by fiddling with components, or connecting a better aerial, you can expect in increase in the general 'clag' which is present on short waves.

[LincolnG]

Originally Posted by lesmw0sec

Sorry, but it is far from obvious. If you are increasing the gain either by fiddling with components, or connecting a better aerial, you can expect in increase in the general 'clag' which is present on short waves.

I never mentioned anything about SW, as for fiddling, I am a service engineer, but forgive me for not knowing everything about technology that hasn't been used for 70 years, many more years than I have lived. Moreover I am not talking about the normal noise on broadcast bands, I am talking about additional noise that was introduced by replacing a broken component, with a widely accepted modern equivilent.

Obviously, this happened as a direct responce to changing the part, I am an engineer I want to know why this had that effect. And since most agree, that this circuit arangment is unusual, it seems my request for clarification was well founded.

What is not obvious is the reason for your unhelpful tone, perhaps if you have nothing useful to say, then saying nothing would be an advantage to all envolved.

Thank you, to the other users here, who have as usual been most helpful.

[LincolnG]

Originally Posted by paulsherwin

I wouldn't expect small variations in the value of that capacitor to make any difference - all capacitors have a tolerance. You should probably check the voltage on the AGC line and confirm it's varying as it should.

I 100% agree it shouldn't. For some reason this radio really doesn't like having that capacitor changed. I managed to repare the old dogbone and fashion a new end. I put it back, now everything is as before.

The AGC line does vary, both with the original and the replaced capacitor. However within a different range. With the MLCC it's on the higher range and I believe perhaps pushing the bias too high, making everything distort. Maybe the MLCC is just too good, compared to the old dogbone. Both do test the same value though, so I'm still confused.

[LincolnG]

Originally Posted by Silicon

Apologies if you know this already.
The cathodes of all the valves are connected directly to chassis.
To Bias them correctly, a negative bias needs to be generated and then connected to each control grid.

The AGC network is connected to the negative bias network to get the advantages of 'delayed AGC' (a voltage delay).

Yes, I had understood that - but thank you for the helpful information

[Chrispy57]

Hi Lincoln - we generally think of a capacitor as being capacitive and measure for that property, but with a 70 year old component it's best not to overlook the possibility that resistance effects may have crept in. Time constants vary for this reason, and will change the behaviour of components in a circuit that is sensitive to such effects.

Cheers
Chris

[LincolnG]

Originally Posted by Chrispy57

Hi Lincoln - we generally think of a capacitor as being capacitive and measure for that property, but with a 70 year old component it's best not to overlook the possibility that resistance effects may have crept in. Time constants vary for this reason, and will change the behaviour of components in a circuit that is sensitive to such effects.

Cheers
Chris

Yes I can see that, additionally, it only occurred to me when I saw your message, that unlike my other radios, this one, perhaps IS very much more sensitive. I noticed with the new MLCC cap. that it was even receiving some stations without any antenna at all. Perhaps it was never designed to have a big external antenna. Sadly, so much information has been lost to time.

Also. apart from the EL34 all the tubes are new, all the caps had to be changed, most of the resistors were carbon and were changed too, this is basically a new radio in an old chassis, it even had to have a new loudspeaker.

People have commented, that the design is unusual - but nobody said whether it was a good design or not. Perhaps, I'll put the MLCC cap back in and test it under a different set of conditions

[paulsherwin]

The only aspect that's really unusual is the use of the diodes. This sort of superhet would normally use the EBC41 diodes for detection and AGC - that's why there are 2. The EAF42 isn't needed, and an EF41 or similar pentode would be expected in the IF amplifier position.

That said, Philips seem to have overestimated the demand for the EAF42 in the early 50s, and they do sometimes turn up as IF amps with the diode left unconnected. I've never encountered a radio wired like this though. It's neither a good or bad design, just unusual. Maybe it was a Spanish thing - we don't see many Spanish designs from the 50s in Britain.

[LincolnG]

Originally Posted by paulsherwin

The only aspect that's really unusual is the use of the diodes. This sort of superhet would normally use the EBC41 diodes for detection and AGC - that's why there are 2. The EAF42 isn't needed, and an EF41 or similar pentode would be expected in the IF amplifier position.

That said, Philips seem to have overestimated the demand for the EAF42 in the early 50s, and they do sometimes turn up as IF amps with the diode left unconnected. I've never encountered a radio wired like this though. It's neither a good or bad design, just unusual. Maybe it was a Spanish thing - we don't see many Spanish designs from the 50s in Britain.

I work a lot on Catalan & Spanish radios, but never seen this; and as I said, the original schematic didn't show it either - mind you, I'm not certain that any radio wired per this schematic, would even have worked, looks like the ebc41 grid goes nowhere. I put a copy for you to look at.

[paulsherwin]

The EBC41 grid is connected to the wiper of the volume control, which is exactly what would be expected in a 4+R AM radio of that era.

[Sideband]

Well it's connected signal-wise to the volume control via a capacitor, the grid bias for the EBC41 (and all the other valves) is via the resistor network from the centre tap of the mains transformer. A bit unusual doing it this way but Philips tended to favour it. All the cathodes are grounded so grid bias is used rather than cathode bias.

[paulsherwin]

Sorry, yes, I should have made that clearer. Overall it shows a distinct Philips influence, though the circuit diagram isn't the usual Philips style.