RA17L: C108 Balancing Cap, 37.5MHz Amplifier

[russell_w_b]

What is the subjective effect of a de-tuned balancing capacitor C108: the one that minimises interaction between the 37.5MHz amp. and the 40MHz pass-band filter?

I've just replaced all the decoupling caps, C107 (changed because of small 56mV positive grid voltage on V9) and all the resistors bar one on the 37.5MHz / 2nd mixer board. After I'd done this I saw a improvement of 2.8V to the previous 1.8V in output at TP3 (V9 out) when swinging through the megs.

I then removed the 1MHz Xtal, V5 as well as V9, and injected 2.5 millivolts-worth of 37.5MHz into TP1 whilst looking at TP3 on my CT471 (suitably padded to 12pF at the probe), and started to peak L50 after setting C108 to 50% mesh. After adjustment of both until I could eke no more peak, I checked the -6db points, and found the bandwidth narrow at only 171kHz, as opposed to the book setting of between 229 - 300 kHz. The mean deviation from 37.5 kHz is 7.5 kHz outside the recommended -6db 20 kHz max. deviation too, and there is a +0.5db peak at 37.521 kHz before it rolls off going up the way.

But the set works! The voltage at TP3 with V9 removed as I swing through the megs is now in excess of 4V, and I was listening to U.S. amateurs on SSB and CW on the 18 MHz band, as well as the odd broadcast station on 6 MHz. Both compared subjectively with a Rohde and Schwarz EK070 RX on the same aerial.

I haven't touched the 37.5 MHz bandpass filter and nor shall I until I can get the Wiltron out, but I was wondering if I might've unbalanced the 40MHz - 37.5MHz interaction avoidance capacitor simply by peaking up the 37.5MHz amp. stage, and whether I might come across 'nasties' up the band because of it?

The manual doesn't say much about C108 other than the role it plays in preventing such interaction.

[russell_w_b]

Originally Posted by russell_w_b

'The manual doesn't say much about C108 other than the role it plays in preventing such interaction.'

Further to my earlier post, I have since happened across a copy of the Electrical and Mechanical Engineering Regulations (EMER Telecommunications E724) which is most helpful, and which states that adjustment of L50 affects the 40MHz filter: arghhhh! It also says what to look for when aligning C108, which requires adjustment of one of the capacitors in the 40MHz filter (C88).

I shall return L50 and C108 to their recorded 'as found' positions and once again measure the -6db bandwidth and centre frequency of the 37.5MHz output at TP3. But I shall adjust L50 and C108 with the set tuned to a signal, and see what happens, out of curiosity.

[Colinaps]

If this upload works.... here's a trace from one of my 40MHz filters with a response on the left at about 37.5MHz. The cap in question shoves that response around.
Find a good birdie and try it - I don't have space on the bench to prove just which birdie right now!

Cheers,

Colin.

[russell_w_b]

Thanks for that, Colin: I might do that this week sometime. I really need to get the Wiltron out but I've too much on right now, and I'm loathe to do any further adjusting until I can. The EMER document has an illustration of an ideal response setting in it.

Today I peaked up C108 (just a tad’s nadger from where it had been originally) with L50 core at its original position and the -6db points now at 37.411MHz and 37.675MHz giving a bandwidth at -6db of 264kHz. Mean of the frequencies is 37.543kHz: 23kHz over-deviated from +20kHz suggested in the manual. The set receives as well as before.

When I reset L50 I wound C108 right round, and when it was fully unmeshed the noise came up but not to the extent of swamping the signal. But I didn't try all the bands. China Radio International on 13M Hz and BBC Radio Cumbria on 756 kHz. I'm sure I'll find a birdie or two!

[Skywave]

Originally Posted by russell_w_b

I'm sure I'll find a birdie or two!

The Racal RA-17 and birdies: a common cause of a 'birdie' is deficient screening. Not only is it essential that all the mating surfaces are clean but that all the screws (and washers) are intact and are tightened securely. Years ago I had a persistent birdie that was 'tunable' (region of 17 MHz) by adjusting the torque of a screw on a screening plate. Tighten the screw a fraction - birdie has gone? Nope: just shifted HF by 0.5 MHz (approx.) Cure was a thorough clean of mating surfaces and new, clean screws and washers.

Al.

[russell_w_b]

Aye, well... I noticed a mild whistle at Megs +750kHz running down the way after doing some S/N tests. I happened to glance down the hole in L52 only to find the slug missing!

[amornummi]

The purpose of C108 is to stop the 37.5 meg signal getting backwards into V7 .
(not sure if it could do much harm there ? )
Adjustment set to balance Cout of V10 so that minimum 37.5 is at the centre tap of L30.
Adjusting C108 can appear to bring up the gain , but you are just messing up the 40Meg BPF.
Whilst you in there I suggest replace R50 2k2 it's usually high , and makes the filter impossible to tune flat.

[russell_w_b]

Thanks for that, Amornummi. I'll open up the 40MHz filter and do things in there when I've time to get the Wiltron out and can sweep it. I'd rather not rush it!

The EMER document E724 suggests monitoring, at TP3, the injected 37.5MHz with a padded-up VTVM and rotating C88 in the 40MHz filter. If the voltage deviates above or below the recommended 1V by greater than 0.3V, then C108 needs resetting.