STC, yes,

My thought is that if the IFT is a dud, then the correct amount of AGC is not being generated to throttle the controlled valves.

Then, the bias is low, valves amplify at full gain and may be unstable!

Thats a sure way of producing whistles, and poor performance.

Fred.

Sounds right, screen bias gets up to 105v when oscillating.

been busy with house repair and family, might get a go this afternoon.

STC, another move I make with a chassis that is out of control is to feed an external voltage from a bench supply into the AGC line.

The external supply of course has to be floating from ground and feed by a similar isolating resistor say 1 megohm.

then you have the ability to manually control the AGC and so the valve gains and also get a grip on a typical AGC range.

The set of course may be pretty lifeless if one IF can is dead, but the point is to get control of the valves.

Once you find out why the IFT does not resonate and fix that problem, then the AGC developed can get back to controlling the valve gains and with normal gain.

Yes, with an RF and three IF stages the set should be a rocket!

Fred.

One thing to remember is that 6U7 is one of the crankiest tubes ever made. Unshielded it is a nightmare & with its other manifestation 6D6 even putting a Goat shield on it, can be trouble. It is one of the most common causes other than lead dress, for spurious oscillation & commonly generates huge amounts of AGC.

I did, on the American forum mention on one post Screen decoupling. Where RF tubes share a common screen supply it is imperative that it is decoupled; should that cap be missing, there will be serious issues. There can be more than one screen decoupling cap to eliminate "standing waves".

If there are issues with AGC. It is a SW set. Often AGC is not applied to Pentagrids like 6A8 on SW as it will destabilise them, so align on SW. AGC may still be on other tubes: 6SA7 is the exception. Also note I do not find 6H6 to be overly reliable.

Again: With new parts & caps with no "outside foil" and lead dress, The area around the 6H6 and first audio can become quite unstable.

If you think the 6U7G shielding has something to do with the problem, you could temporarily replace the 6U7G with a EF39 for testing purposes.

The EF39 has sprayed-on red shielding, and so the goat shield isn't needed. It does rely on Pin 1 being earthed of course. It's plugin compatible - no need to rewire anything.

Further to my post #17

I tried connecting the grid cap from IF2 transformer, via a 0.001μF 630v cap, to the cathode of 6H6 detector valve. Sensitivity did not appear to be decreased, consistent with IF3 not working, but oscillation the same.

Then went the whole hog and bipassed the IF2 & 3 stages entiely by connecting the grid cap lead of the IF1 transformer to 6H6. Sensitivity much decreased, but oscillation the same.

So it would seam the oscillation is centred on IF1, ie related to 6A8G and/or the OSC transformer, to my way of thinking.

Now I have never opened up the aerial, RF or OSC tranformers, basically because if it ain't broke, don't fix it. But I have checked their continuity, and all are OK from that respect. There are no internal caps in the OSC where the fault might be, but there are in MW aerial transformer, and MW and SW1 transformers. But given that there may be a fualt in MW OSC transformer at least, maybe should have a look, but the retaing nuts for the shield has very difficult accessibility under the wavechange switch - I might get em off but am not sure I would get em on again.

Fred post # 18. Like the idea of external AGC voltage - will see if I can modify my simple DC supply which has earthed DC negative. Failing that could try changing the voltage divider to drop the screen voltage and hence gain.

Marcc post # 19: Am pretty happy with IF transformer and valve shields going on resistance to chassis - zero to 1 ohm.

Re screen decoupling, the cap is there (bluecap bought from DS in the late eighties), but could add others.

Re AGC it is not applied to 6A8G at all, MW or SW, but is to RF, IFs and mixer.

Will look at 6H6 and RF area since I changed the earthing point fo the RC filter between IF3 transformer and 6H6; but pulling 6H6 and 6J7 1st audio do not stop the oscillation -can't hear through the speaker, but it is evident in the full deflection of the tuning indicator.

Robbbert post #20: will see if I have any EF39 - might take some digging.

One ohm to me does not constitute a ground: That is a potential radiator.

Buy a Chinese dinner and salvage the chopsticks. These one can use to poke & probe. If you have an oscillation and probing changes the frequency that's usually the area that needs attention.

Placing a hand on a shield and things change: Says ineffective. If you have no spare 6A8, try a 6J8 but don't expect the band to track. Wind the tuning gang rotor out to ensure its not shorting.

Did you align the IF and how, If its off caused by changing parts that will not help.

Marc

OK will perfect these grounds.

Using a plastic knitting needle, and while oscillating, any wire touched in the circuitt changes the tone, but most sensitive around 6A8G.

I have two 6A8Gs and swapping out doesn't help. Will try a 6J8 however.

I didn't align after IF stages bypassed - will do so.

Am presently trying to improve my crude shielded bypass 0.001μF cap and its lead.

Quite a lot to think about and will leave it for a day or to and work out a plan of action.

Marcc, re my post# 23:

Tried a 6J8 - oscillated in motorboat fashion
Tried a 6K8 - no oscillation, but still dead at the low frequency end of the MW dial. Will try an alignment with this valve to see if I can tame the IF sections.

Have decided to order a 6A8G from the HRSA valve bank so that I have another to play with, in case both of my 6A8Gs are suspect.

How do reliability problems with 6H6 show? I have all sorts of them, steel ( NOS), G, GT and no joy with any of them partnered with the 6A8G.

Plug-in possible equivalents for the 6A8G - they may work depending on the way the radio is designed.

For me they all worked in a PYE that I was testing, but in some other radios only certain valve types work. Your mileage may vary.

You can try any of these as a 6A8 replacement: ECH33, ECH35, 6J8G, 6K8G, EK2G/GT, VR57, X61, X61M .

Make absolutely sure that pin 1 of the socket is earthed.

A shielded 6A8 should not oscillate and if they all oscillate. Tube is not the issue. Beware with Philips Metallised tubes. Towing them out by the envelope can break the bond between the earthing pin (not always pin 1 check) at the top of the socket. That & loss of metallisation is fixable.

Check your work and the values & placement of new caps. If the capacitor & coils separate you often get supersonic oscillation. The chopstick is telling you that there is a feed back loop. That is more likely a wire away from the chassis, a modern cap or a miswire. Wires should cross at 90degrees. Wires running parallel can induct.

Did you check the oscillator gang for shorting? Rotate the rotor vanes to fully open & feedback if that does not stop it.

The weakest Oscillator signal is at the highest frequency and that is where dud Pentagrids fall over.

Marc

Robbbert: re your post # 25 - thanks for those suggestions, will go on the hunt for these in my spares

Marcc re your post # 26 - Looks like the 6A8G is more susceptible in this circuit than the 6A8. I will try to stick to it once this problem is fixed simply because the tracking is good with it, and tuning stations these days depends on finding its frequency given that the call signs are mostly displaced.

Have replaced only two caps in the current restoration, one a mica (NOS) in the RC network at the 6H6, so that is definitely worth a look. The other the grid feed for the output valve because it is under stress - the previous cap was OK but a bit lower voltage rating.

The mica above is the only one that I have replaced ever, and that because it replaced two in parallel that were inconvenient during the restoration. Recent comments elsewhere have indicated that micas of 1930s vintage are starting to fail. I have thought about this job but put it off because some of these caps are practically inaccessible behind the wavechage switch. I have quite a few NOS micas - can these be used or do they also deteriorate in storage?

First thing I thought of was the tuning gang but can't find any problem. While oscillation is around the 630-680 Kc mark, it does move around a bit, 10kc or so either side of this range. In particular though, running the radio without the RF stage shifts it to around 800Kc. I don't think that this is consistent with imperfect gang shaft earthing; poking around these earths with a probe has no effect either.

This problem is at low frequency so I get your point if weakest oscillator signal is at high frequency.

I do not consider fault finding in a set full of oil filled & waxed paper caps and worn out electrolytics: Productive. A 6A8 should be metal and its earthed body the shield any octal tube with no "G" is metal and "G" glass (6A8G tubes tend to have no shield. Normally pin one will be wired to ground for a metal (not always pin1) https://www.radiomuseum.org/tubes/tube_6a8.html. Franks Electron tube page is down?

With non polarised caps. Valve radio Mica types from the late 30's are rare failures. It the Mica is across a coil unbranded: No touch it is specific to that coil. If one is out of circuit or suspect, the leakage test will be 500VDC using an insulation tester.

With the rest, paper & oil (not electrolytics) if I bother, the leakage test is at 250V 500 for sets with rectifier diodes & #80 / 5Y3, or as close to rated value. 99.99% will fail. Screen decoupling caps should not be below 50Meg and coupling caps not below 200Meg. Valve tester uses 225 VDC and a Neon tube: Pass / Fail. Standard here is if it leaks it goes in the bin.

Leaking caps can damage tubes and cause overloads.

I will often run card between the vanes to clean them of corrosive castles & such as they can arc over on oscillator gangs, which are often alive. After removal nasty chemical things often happen to them here if seriously dirty. Most gangs have a screw and locknut at the rear end of the shaft to tighten them. loose shafts & bent vanes can short. Once isolated the Insulation tester is used at 500V then 250VDC to assess if it can flash over on the mounds, or a bent, or too close vane. Watch the fingers if you try this. Some older guys used the transformer "B" : Scary.

Hello Marcc
There was a total recap in 1989 except for the micas.
It is a 6A8G with shield.
Noted re unbranded caps across coils. Have a resistance tester which I am going to apply today to the fixed padder cap parallel to the padder trimmer (one thing I have never checked, filthy inside) . Will check all micas around 6A8 and the oscillator transformer that I can get to, to start with.
Will also check the screen decoupling cap, and coupling caps since most are 33 years old bluecaps.

Re the tuning gang, did all this when restored in 2022. Took out the bearings and put the gang through the dishwasher, treated corrosion and painted, refitted bearings, lubricated, and cleaned the brass shim shaft contacts. Lubricated the shaft contacts with electronic cleaner and lubricant. If this lubricant got into the gaps of the gangs then tracking would have been off so I don't think so. All the earthing points were remade with new braid soldered onto the cut end of the old braid so I didn't have stick a b great soldering iron into small gaps.

All seemed sweetness and light then, but didn't test the gangs with a resistance tester. Difficult to do now as it would require cutting very thick wires from the gangs to the band switch since the terminals both gang and switch ends are inaccessible.

Hi STC, my advice is: to stop going off in tangents.
Fix what you have found is actually wrong.
You have an IFT that is not resonating and peaking.
That alone will stop the AGC working to control the front-end gain and so will cause valves to run at high gain and be unstable.
The set worked before it became faulty.
You have found a fault.
Fix that fault.
Sorry to sound so abrupt but going off on wild goose chases is not productive and likely to create more faults.

The IFT is not resonating so has low gain.
Pull it out and substitute another IFT of the same frequency type and general vintage.
If that does not resonate, then there is something else wrong associated with the stage.
You need methodical steps. not mad professor, could be this, could be that, could be the other, carry on.

Cheers, Fred.