TO-R S15, Do I really need a new Project ?

[Jim - G4MEZ]

Excellent work Well done!

[tri-comp]

Pre-amp renovated.
All components around the ECC83 were removed and the PCB cleaned.
Two more Hunts, this time electrolytic cap's found under the Volume-pot.
Of course both were leaky and they completely lost capacity.
The Mustard-cap's all measures close to rated value so they were just de-gunked and put back in.

Next it'll be time to focus on the FM-tuner (Torotor) and the IF.
Cathode voltages on both EF89's are low indicating they're not drawing as much current as they should. G2-voltage is also too high; same problem.
If the resistors check-out I'm afraid both valves are showing age. NOS are expensive and I don't have any

[Nuvistor]

Is the AGC turning the valves off rather than low emmision? Although I cannot fathom from the circuit how the AGC is derived or how the magic eye works.
I have another look at it, any suggestions gratefully received.

edit. Possibly from the grid of the last IF acting as a limiter? There should be a negative voltage on the control grid depending on signal strength.

Frank

[tri-comp]

The AGC works fine but of course the base may be off even if I don't think so.
With a strong antenna-signal (still using the 50cm piece of wire) AGC rise to -4V. Between stations it's -0,7V.
It's not too hard to understand how it works.
You need to locate the ground-end of the grid-leak resistors for the IF-valves.
It's actually the two resistors R11/R12 under the tuner-section schematic.
The larger the IF-signal the more summed negative voltage is formed across the grid-leaks.
Perhaps unconventional, but it works.
I just need to check all the resistors to know if it works a 100%.
(Voltages measured with a digital VOM, Internal resistance = 10MOHM, so of course it's loading the circuit and causing somewhat false readings.
The Eye clearly drops a little when doing the measurements on the IF grid-circuits)
The AFC didn't work as it should.
The base is set by R10 to 4V but it was only just past 2V so I though C2 may be leaky.
It wasn't but I replaced it anyway. The fault was R9 all out of spec's at 224KOhm.
I don't think something similar is causing the low current in the EF89's as when the AGC-voltage drops between stations the cathode-voltages hardly rise.


EDIT:
I found my trusty old B&K Model 277 FET-VOM (Ri = 15MOhm) and did another test of the AGC-voltage.
This time no reaction from the Eye when touching the circuit so loading is very minimal with this meter.
Strong signal AGC voltage now measures -5,5V

[tri-comp]

I renovated the circuit around EF80.
All original resistors are replaced and this had an effect.
The resistors were off rated value, some more than other but way past the nominal 10% variation.
The effect of replacing them was somewhat unexpected.
Of course you expect maximum AGC voltage to appear spot-on station but it didn't before. I had to slightly detune the station to obtain max. AGC.
After renovating max. AGC is now exactly following 0V DC output from the discriminator. I'm not sure the S-curve is fully symmetrical as I feel there's more room to tune above the station than below before distortion is noticed. Is that at all possible when max. AGC appears at 0V DC-out ?
As i don't have a suitable sweep-generator I'll leave the discriminator as is.

After having run the set for several hours the pair of ECL86' started acting up. One pentode started drifting, drawing about twice the normal class-A current.
The other pentode developed a sort of shot-noise, possible some periodical internal short that would raise the cathode-current for a fraction of a second.
No problem, the Miniwatts could just be replaced with a pair of Philips that came out of a butchered B&O set a long time ago.
After about 30 min's both pentodes were drawing far too much current, about twice expected.
The set is now running on pair of NOS Raytheon ECL's and all is fine.
Cathode voltages are perfectly matched (5,27V & 5,22V)
Just goes to show you shouldn't trust used valves unconditionally. They may work perfectly when applied but after running for a while expect whatever...!

Btw., you won't believe the going price of ECL86 these days !

[Nuvistor]

Grand job.
Wouldn't tbe max AGC depend on tbe shape of tbe RF/IF curve not tbe discriminator?
Since the AGC voltage is derived from the EF80 limiter.
Really need a sweep gen to check its shape.
ECL86's were not the most reliable, but that could be due to them very often being run nearly flat out and yes I do believe the price these days.

I don't know if it's because we did not see this cabinet style in uk designed radios very often, if at all, I always liked the look the Scandinavian designs.

Has a said grand job, should give many hours of pleasure now with the work you have done.

Frank

[tri-comp]

If dug deeper into my stock of used valves.

A couple of EF89's showed themselves.
One is extremely clean; possibly a NOS Miniwatt w/o box that came with used valves I bought over time and was forgotten because of the vari-mu characteristics. Looks and performs like new and pushed the AGC to -6,2V, a 0,7V increase.
The 2nd. original Valvo EF89 is absolutely fine, no change in AGC voltage when replaced with the newly found specimens.

Something equally interesting fell out.
An absolutely mint TFK EM87, possibly NOS but w/o box, that outperforms the existing EM84 by far.
It's more sensitive and obviously has a lot brighter output.
The TFK has a 6,12mm wide gap on -6V AGC, the EM84 a 8,9mm gap.
The TO-R S15 actually exists in two variants. There's another, newer version (S15u) with extended FM-coverage to 108MHz, ECLL800 as output-valves AND the EM87 for the magic-eye.
I see why they went with the EM87 and the TFK stays in mine.

The first picture shows the EM87 'upgrade' and the second picture the original EM84.

Also a lot more ECL86's appeared. All used and in unknown condition. Time will tell...

Now I'll go and replace the rest of the carbon resistors in the IF and see what happens.

[tri-comp]

A small update:

I had to check the FM-tuner ECC85 as I also found a pile of those.
The original Miniwatt actually performs very well. A used TFK did just a LITTLE better, AGC voltage at -6,4V and THEN something really wonderfull happened.
I found a Siemens ECC85, stamped B&O 1961 and that pushed the AGC to -8V.
The eye is almost closed, gap at 2,1mm.
WoW !

Back to replacing carbon resistors...

[Nuvistor]

Gets better and better.

[tri-comp]

IF all done.
NOT a lot of off-value resistors but still enough to push the AGC to -8,2V with new resistors in place.
Or perhaps it was the pair of ceramic decoupling cap's for the heater-circuit that did the trick.
Next in line for an overhaul will be the FM-tuner.
I got lucky and found a way to securely fix the dial string wheel during removing of the tuner.
Ty-Wraps do come in handy at times

[tri-comp]

Tuner away and I need a bit of advice.

As you can see the tuner is a mess.
The soldering side of the PCB looks like it could do with a nice cleaning and so does the components side.
However, I know it will change the alignment of the tuner if I do.
The Scale is now off by about 2MHz to the high side and I should adjust the oscillator down a little.
As you'll notice there's no easy way of doing this; i.e. no trim-cap.
Alignment was done at the tuner-factory Torotor by adding solder to the oscillator PCB-coil.
Obviously I can't hope to copy that.
There's another issue.
The original (Vitrohm ?) resistors are all carbon composite types. I've been using almost solely vintage Philips carbon resistors for the renovation. Do you have any idea about how replacing the resistors will influence tuner alignment ? I have to say that all the original resistors are within their 10% tolerance margin. Should I just leave well alone ?
All tuner-cap's are tubular ceramic types save a pair of ceramic disc decoupling cap's.
I have NEVER come across one of these cap's off spec's. So, if it's not the cap's shifting the oscillator what is ?
Grime ? Conductive Pertinax PCB ? Or... ?
The AFC Varicap measures OK as a diode. Forward voltage drop = 590mV.

[Nuvistor]

If the resistors are within spec I would leave them, I would not expect the value to affect freq with them being in spec but the physical shape style change could.

Do different valves affect the freq.?

Frank

[SteveCG]

I suggest you keep to using carbon composite type resistors in the tuner section. Other types may have a spiral path cut into the conductive material of the resistor and this inductance will affect the alignment at these frequencies.

[tri-comp]

Meet 1ea. de-gunk'ed Torotor-tuner.

I took great care not to disturb any frequency determining parts except the AFC varicap that was raised when cleaning underneath.
That shouldn't present a problem imho.
2 resistors, decoupled at both ends, were replaced.
Resistors with 'hot' ends were left original carbon composite types.

That's about all for this evening. I'm readying a new Lenovo Win-7/PRO computer that I need tomorrow. Windows back-up's on the Win-7 is troublesome at best but tonight I seem to have nailed the problem once and for all. I actually found out how to overcome this common problem and the solution does NOT include any MS Fixit.
Sorry that was a bit off-topic. If anyone wants the solution to the Win-7 update problem feel free to PM me.

[Nuvistor]

Using printed circuit coils would be ok in the service life of the unit but long term changes in the PCB material due to heat etc could cause problems.
hope the clean up sorts it out for you.

Frank

[tri-comp]

How simple can it be to drop the oscillator frequency down a couple of MHz ?

After the tuner was cleaned I assembled it and found that practically nothing had changed except that it looks a whole lot cleaner.
The scale was just about spot on where I left it before cleaning; off by about 2MHz high.
Time to experiment a little. I took the tuner apart again and added a ceramic 3~9pF trim-cap. across the oscillator-coil in parallel with C12 which incidentally is 15pF in this tuner, not 13pF as per schematic.
Have a look at the picture.
I hoped that even with the added stray capacitance I would be able to adjust the tuner into proper range.
Here's the point where sanity is challenged. Mine at least.
It seems the addition of the extra cap actually RISES oscillator frequency.
How can that be ?
Setting the variable cap at minimum almost brings back the same scale setting as before adding it. Adding capacity moves the station up-scale.
This is not possible in my world.
What am I missing ?

[tri-comp]

I have a good mind to replace all existing ceramic tubular cap's with modern SMD cap's.
I know it's a bit radical but the tuner can always be restored to original if it turns out to be a bad idea.
I expect the new cap's to be still mounted on the components side of the PCB with leads added.
That should probably cause less stray capacitance.
What brand/type should I go for?
Vishay VJ HIFREQ Series or perhaps Kemet CBR series or something else/better/obtainable ?
I really don't have much experience working at VHF frequencies.
Perhaps the above types are not at all suitable ?

[turretslug]

Originally Posted by tri-comp

The Scale is now off by about 2MHz to the high side and I should adjust the oscillator down a little.

T-C, feel free to slap me about the face with a metaphorical wet fish if I'm being dopey and getting things all the wrong way round, but do you mean that the scale is reading high in that a station known to be on, say, 100MHz is being received at an indicated 102MHz? In that case, there is too much capacitance in the frequency determining circuit and so the station is being received at lower variable capacitor setting, i.e. higher frequency indication, than it should be and capacitance needs to be removed, not added.

[dseymo1]

...Which, interestingly enough, would correspond with the discrepancy in C12 value.
I wonder if it's a substitute, or the wrong value was fitted in error?

[tri-comp]

The problem manifests itself like this:

A known station broadcasts at 93,3MHz and I'm receiving it when the scale shows around 95,2MHz; the better part of 2MHz off.
I expected to drop the oscillator frequency 2MHz to cure the problem by adding parallel capacitance to the osc.-coil but as described it works 'in reverse'.

Considering the darkened pertinax PCB and that the grease/grime I removed from the board and components had to come from somewhere, I would expect that the stray capacitance increased over time. 'Cooked' carbon composite resistors may now have random values at VHF; see the attached article from Wireless World 1959.

So, if I in fact I need less capacitance to align the oscillator; a wrong and higher value of C20 combined with aged Pertinax and aged carbon resistors may be what needs to be addressed.
C20 hasn't been replaced at any time as far as I can judge, the PCB can't be fixed but the carbon resistors can.
First I'll remove the added trim-cap and replace the resistors. Re-assemble and see in what direction it goes.
If still not aligned I'll drop C20 in value to the specified, possibly even lower and add the trim-cap back in.

Does that sound like a plan ?