Marconi T21A Background Noise

[Voxophone]

Hi All,

I'm in the process of restoring a Marconi T21A which seems to be picking up or generating an excessive amount of background noise.

I can tune into stations on all bands and presets, but especially on MW there is a kind of rumbling noise which can be quite audible and distracting on weaker signals. The noise does not appear to change as the tuning control is moved.

I've tried to show what it sounds like here: http://vid13.photobucket.com/albums/...psdlyuxykz.mp4

The set has had all wax capacitors replaced and all wave-change switches cleaned and tested. Most of the r.f. capacitors have been tested with an LCR meter and are close to the correct values. I've tried substituting V1 and V2 with no change.

The IF response seems adequately peaked to me from tests with the signal generator, but the cores are very stubborn so I haven't tried an alignment yet. I did manage to un-stick one somewhat but moving it only confirmed that the original setting was correct.

Does anyone have any ideas? I've heard that the 'silver mica disease' found in US radios produces a sort of rumbling sound, but am reluctant to open the IF cans except as a last resort.

Suggestions much appreciated.

Cheers

Liam

[Boater Sam]

Could I suggest it may be audio related? Slow motorboating? Or AVC related?

[Graham G3ZVT]

I had difficulty with that Photobucket video, I had to resort to an alternative browser to get it to work.

The clue is the heterodyne whistle as you tune through a station. It seems that your IF amplifier is "taking off", ie oscillating instead of just amplifying. Look towards one or more HT decoupling capacitors as the likely cause.

Graham.
G3ZVT

[ms660]

It doesn't open up on my PC, just a load of spam/ad/cookie and general unwanted rubbish.

Lawrence.

[Voxophone]

Thanks for the suggestions. The IF whistle sort of comes and goes. Sometimes it isn't present.

Is there a way to upload videos directly to the forum?

Liam

[AC/HL]

Not videos. Most people download to Youtube and post a link to it.

[Voxophone]

A follow-up on this issue.

I noticed that the MW performance of this set was fine when using an external tuned loop aerial. Comparing the background noise in the video against my other sets when off-station, I found it sounded about the same, but was not generally noticeable because the station volumes were so high in comparison. I started to think that the problem might be due to the aerial circuit not peaking.

Based on the above I started looking at the aerial circuit to see if might not be peaked properly on MW. Sure enough, no matter what frequency I inject into the aerial terminals across the MW band, I can’t get the circuit to peak at any point of the range. In fact the signals were greatly attenuated (20mVpp output for 1Vpp input, measured at the V1 top cap). Conversely, SW and LW both peak very well, with output exceeding the applied signal amplitude when tuned to about the right points on the scale. Moving the MW aerial trimmer has no effect on performance.

I examined the MW aerial components to see if there was a problem. The MW aerial trimmer, C35, gives a range of about 9-60 pF when measured at 100 kHz on an LCR meter with one leg disconnected. The MW aerial coil, L4, has an inductance of about 152 uH and a Q of 3 when measured at 100 kHz. The DC resistance is on spec according to the Trader sheet.

Can anyone tell me whether the above measurements are typical for components in an MW r.f. circuit?

I did a rough calculation based on the LC resonance equation, which suggests that a 152 uH coil would need approximately a 100 pF capacitance to peak it at 1300 kHz (the MW alignment frequency suggested by the Trader sheet). This might not be relevant since the components were disconnected from the associated wiring, but could it suggest that the trimmer is not giving a high enough capacitance? On the other hand, 100 pF seems big for a trimmer to me, especially since some swing would be needed either side.

Any thoughts?

Cheers

Liam

[G8HQP Dave]

Q of 3 seems rather low, even at 100kHz. Has some damp got into the coil former?

A fairly ordinary solenoidal coil might be expected to have an unloaded Q of around 50-100 in its working range, assumed to be 1MHz in this case. Inductive reactance varies like frequency, and resistance will vary like sqrt(frequency) - assuming normal skin effect dominates. So at 100kHz it might have around 1/3rd of the Q it has at 1MHz. This is all back-of-envelope calculations but it shows that you should be seeing something like 15-30 when you measure Q at 100kHz.

[Skywave]

Originally Posted by Voxophone

The MW aerial coil, L4, has an inductance of about 152 uH and a Q of 3 when measured at 100 kHz. The DC resistance is on spec according to the Trader sheet.

That Q-value seems very low. What instrument and what method did you use to measure that Q?

Al.

[Voxophone]

Originally Posted by Skywave

That Q-value seems very low. What instrument and what method did you use to measure that Q?

I used a DE5000 LCR meter set to measure Ls and using a test frequency of 100 kHz. Q is measured automatically by the instrument.

C3 was disconnected from L4 and C35 was disconnected at the S4 end. The wave change was set to LW so that S3 and S4 were open. The meter was connected from chassis to the C3 end of L4.

I've tried to show this in the attached drawing.

Thanks

Liam

[ms660]

Mutual inductance might affect the readings, assuming the top end of L4 has continuity with C3 and the bottom end has continuity with the chassis and with everything connected as normal I would set the tuning to mid range on MW then do manual RF sweep in at the antenna socket and 'scope L4 where it's connected to C3, if nothing then I would connect a 200pF (ish) capacitor from that same point to chassis and see what happens when manually sweeping again.

Lawrence.

[Skywave]

Originally Posted by Skywave

That Q-value seems very low. What instrument and what method did you use to measure that Q?

Originally Posted by Voxophone

I used a DE5000 LCR meter set to measure Ls and using a test frequency of 100 kHz. Q is measured automatically by the instrument.

C3 was disconnected from L4 and C35 was disconnected at the S4 end. The wave change was set to LW so that S3 and S4 were open. The meter was connected from chassis to the C3 end of L4.

Classic text-book theory states that to reliably determine the Q of such a coil, it needs to be disconnected from other components and removed from the set - and then its Q measured.

Al.

[ms660]

Originally Posted by ms660

Mutual inductance might affect the readings, assuming the top end of L4 has continuity with C3 and the bottom end has continuity with the chassis and with everything connected as normal I would set the tuning to mid range on MW then do manual RF sweep in at the antenna socket and 'scope L4 where it's connected to C3, if nothing then I would connect a 200pF (ish) capacitor from that same point to chassis and see what happens when manually sweeping again.

Lawrence.

RE: Above...."if nothing" I should have said...if no peak.

Apologies.

Lawrence.

[Voxophone]

I tried the method Lawrence suggested but was unable to get a peak, with or without the additional 200 pF capacitor.

Afterwards I disconnected all leads from coil L4 and the other two coils which share the same former (L2 and L5). I then measured the inductance and Q value of each coil at the bare solder tags, with the following results:

L2 = 9.887mH, Q = 21.6
L4 = 154.5uH, Q = 3.68
L5 = 2.13mH, Q = 46

It seems pretty clear then that something is off with L4, but the DC resistance is only about 0.2 Ohms high against the design value.

Moisture has been mentioned previously as a possible problem. Is there a recommended way of removing this? I’ve had this set for over a year and it’s been stored at room temperature throughout that time.

Cheers

Liam

[Boater Sam]

Could be green spot corrosion between turns. I'd just rewind that coil if its not too difficult.
You could try disconnecting it and putting a small axial wound choke in with its axis inline with the other coils to prove the point. Shunt it with a trimmer and you could adjust it to the correct peak.

[Voxophone]

I think it would be horrendous to rewind as its wave-wound and seems to be in two adjacent parts. It is on the end of the former though. Might be possible to slide it off and replace with a similar scrap coil?

Is there really nothing that can be done to salvage it? I'm aware of moisture problems with TV LOPTs which seem to be fixable by drying the coils out.

Liam

[Boater Sam]

You could try drying it but I doubt it will work. Green spot corrosion is thought to be due to sweat on the hands of the original winder, usually female, add damp and time and the wire oxidizes, usually shorting turns in a green spot, hence the name.
I am unsure if replacing a wave wound coil with a pile wound coil really makes a lot of difference, no doubt some worthy will jump down my throat soon.
If the winding is in 2 parts, can you establish which part is low Q? Then maybe you could replace just one part?
Using a choke is the other way, there are sites on the net where a whole IF transformer has been replaced in this way, using 2 chokes and trimmers.

[Voxophone]

Would a green spot not also affect the resistance readings? I suppose since the coil is only 2.7 Ohms and the spot might be relatively high R in paraellel with part of it, hence could be going unnoticed by the DMM?

Regarding the choke, could this be placed inside the hollow former and the original coil disconnected? It needs to couple inductivey to the aerial coil L2 which is on the same former.

Thanks

Liam

[Boater Sam]

I would not expect a noticeable change in the resistance, until the coil goes O/C of course. I would think that its the slightly shorted turn/s that destroys the Q effectively.
It would seem that just getting the choke alongside the former but in line is sufficient to achieve coupling. In fact one site shows 2 chokes, primary and secondary, just side by side, not even on the same axis!
Have a look at this
http://www.antiqueradios.com/forums/...p?f=6&t=196404

[Voxophone]

Thanks. I might give that a try then. Should be able to wind a test choke by hand and check inductance and Q with the LCR meter.