CB squelch problem

[G0HZU_JMR]

Originally Posted by Damo666

At the yellow arrow, D103/R125, it was 0.938V with the squelch fully open. With squelch set maximum, the voltage here was 1.208V. The voltage increased linearly as I turned the squelch from open to fully on.

With squelch set mid range, the voltage at D103/R125 is 1.140V, dropping to 1.135 with a weak signal, and dropping to 1.120 with a local strong signal.

Those results look a bit suspicious to me so it does look like a fault is here. If we assume that D103 is a 1N60 Ge diode then we can predict it Vf characteristic within a realistic ballpark assuming the circuit is all healthy. In the absence of any RF the diode is forward biased from two sources.

The first is a (maximum) bias of maybe 100uA via R125 (33K) and R127 (47K) and the 8V receiver rail. But some of this bias that gets to D103 will be reduced because of the tapped bias into the receiver AGC line so I'd expect much less than 100uA into D103 here. The other way it can be biased is from the 0.7V Vbe of the squelch transistor Q206 and R126 (22K). But this would be less than 20uA for a healthy diode with a low Vf.

With such tiny currents available, I'd expect to see just 200mV Vf at the anode pin of a healthy/typical 1N60 (yellow arrow TP in jpg). So 0.6V-0.2V = 0.4V and this will bias at about 20uA.

But you are seeing 1.2V Vf in the diode which looks suspicious. This implies that the diode is either faulty (open circuit) or something is very 'different' about your radio compared to the schematic. I'm also assuming that the resistance of the transformer secondary winding is healthy (and very low) here and the transformer is not open circuit or badly soldered? I therefore suspect that D103 is open circuit?

However, can you check the transformer with an ohmmeter? Do you see low resistance from the transformer secondary (D103 cathode) to ground? I'm expecting this resistance check to be OK or the s meter wouldn't work properly.

[SiriusHardware]

As the S-meter is working, surely T106 secondary is OK? (Ah, you beat me to it).

[Damo666]

Again, thanks a bunch for the continued input as I'd have been completely stumped without your kind assistance.

The 1N60 Ge Diodes will hopefully be here tomorrow, so I'll pull the original out & replace it with a new one and report back.

I have 10 x 1N60 coming, and as I'll only need 1 (I hope) - if either of you are short of them and want some, I'm happy to send the remainder out FOC as a gesture of my appreciation.

Cheers again.

[Damo666]

One more thing, and my apologies about the novice questions; When I was checking the voltage on the Anode of D103, I was checking it with respect to ground - IE, positive lead of DMM on Anode of D103, negative DMM lead to chassis.

I presume this is correct as opposed to checking the voltage across D103 itself?

[SiriusHardware]

You did it right. Jeremy deduced the forward voltage drop across D103 from the fact that the cathode end of the diode is connected to 0V via the few ohms of resistance of the secondary of T106.

1.2V is an unlikely, rather high voltage drop to have across a germanium diode, so that high forward voltage points to either the diode or T106 secondary being faulty. With the receive meter working, it seems likely that T106 secondary is OK.

Which leaves...

If I were D103, I'd be making arrangements for my funeral right now.

[Damo666]

Aha, I see now. I have these dumb moments were I fail to see these things. It must be old age.

The Diodes haven't turned up yet, but they were only sent 2nd class yesterday - so that's fine.

Not sure what's happened with the schematic I ordered, because that was ordered late last week, so I'm going to phone them tomorrow (shut today) and ask if it was despatched.

[SiriusHardware]

I ordered mine Thursday evening and it had arrived by Saturday. I hope they weren't confused by two orders for the same thing at the same time? Probably the first time in years that anyone had asked for that drawing.

[Damo666]

My schematic still hadn't arrived this morning, so I phoned Knights, he checked his previous orders and said he'd missed it - but I think you might have a point with what you say about 2 orders at a similar time for something as unpopular as the GX-19.

Anyway the 1N60 Diodes turned up today, so I fitted one and the CB is working real well & the squelch issue is now gone.

Not sure if it was yourself or Jeremy that mentioned it, but the faulty 1N60 did indeed just crumble in 2 pieces when extracted - and it was a quick removal so wasn't subject to much heat from my iron.

You're probably fed up of me saying this, but again, a massive thank you to all for taking time out to resolve this problem for me. I hope I can repay a favour one day.

[SiriusHardware]

Glad it works, and honestly, we had fun working through it with you. Enjoy your radio.

[G0HZU_JMR]

Yes, a good result all round and it was fun to look at stuff like this again

SiriusHardware:
Yes, normally they'd mount diodes like that vertically if the holes are too close together. They must have been treated quite roughly to get them to fit horizontally and if D103 turns out to have failed, that may turn out to have something to do with it. (When removed, it may just fall apart into two pieces).

Yes, some Ge diodes are quite sensitive to heat damage with an iron or from stress from leg bending. Maybe these diodes were fitted to that Pama CB by a different (2nd) operator at the factory because the legs look like they would have to be bent a lot by hand to make them fit the narrow hole spacing on the PCB. So maybe these parts were fitted after the main board had been done by other means? I'm guessing that it's possible to store (springy) stress in the part if this is done badly and maybe this was a stress failure over time? As you say they would normally be fitted vertically if there was a hole spacing issue. A bit like the resistor in the image below?

Back in the day, the recommended tool for fitting these diodes was a heatshunt like the one pictured below. This is my old one and I must have had this for 35 years or more. I haven't used it in ages but it has been in with my stash of tweaker tools all this time. I was never really sure how effective it actually was but these days I have a thermal camera and I suppose I could test how quickly the resistor heats up with or without the shunt. It does grip the component leg quite tightly but I'm not convinced it will make much difference...

[Damo666]

I have something very similar to the heat shunt you have there; they're similar size, but tweezers (self gripping) with wooden handles and somewhat longer.

Around D103 I did notice the soldering was different, so I think you might be right when you say this could've been fitted by a different operator. On the whole, there were many flaky solder joints that needed touching up - quite the opposite of the the Japanese rigs built in the 80s.

On a different note, I noticed last night that one of the Channel readout segments has just become intermittent on this GX-19, but I know where the problem is. I just need to remove a nut on the microphone socket & channel changer to remove the fascia, but how I do this is beyond me as there's very little space around the nut to actually fit a socket in there from the front to remove the nut. I think I'll try using some long nose pliers with a rag on the outside to avoid scuffing the plastic entry points.

[Biggles]

I once had someone drop small spanner on an FM tuner (don't ask!) which promptly went silent. My first thoughts were that a connector had been dislodged. On further internal investigation a glass diode was visibly snapped in half. I can only think that the diode may have been soldered to the board under tension, and the sharp shock, which wasn't much really, was just enough to physically snap it clean in half. A new diode fixed the problem. I wonder if something similar had happened to your set, or maybe it was subject to extreme cold in a vehicle? We will probably never find out, but a good result in the end. We all love faults like this!
Alan.

[SiriusHardware]

Originally Posted by Damo666

On a different note, I noticed last night that one of the Channel readout segments has just become intermittent on this GX-19, but I know where the problem is..

To get at the nut, it sounds like you need what is traditionally called a 'box spanner', which is a metal tube with one end shaped into a hexagon outline.

First example I found:

http://www.toolstation.com/shop/p169...FakK0wodHCgFfQ

A 'long reach' socket set might do it but usually the outer barrel of a socket is thicker than that of a box spanner, so it might not fit down into the restricted space.

About the display problem, what precisely is your symptom?

1) One segment on one 7-segment display intermittent?

2) The same segment on both 7-segment displays intermittent?

3) One entire 7-segment display, all segments intermittent?

The answer to this question will narrow down the fault area quite a bit. In this radio the PLL IC generates the channel display, and it does so by multiplexing the two displays: The sequence is:

Set the data for the tens display segments on data lines A to G and turn on the power to the tens display.

Wait a bit

Set the data for the units display on data lines A to G and turn on the power to the units display

Wait a bit

Repeat.

If you could film the displays in slow motion you would see that each display flashes on alternately, although they do so so quickly that it appears to human eyes as though they are both on at the same time. (If you look at them through the electronic viewfinder of a digital camera or camera phone, you will probably see them flickering).

If there was a fault on (let's say) the 'A' output drive line from the PLL IC to the display, the top segment on both displays would be off. But where only one segment on one display is affected, that is more likely to be a problem with the actual display. It also depends on whether the display is two individual displays wired as a multiplexed display, or (more likely) a twin 7-segment display.

I do question the wisdom of using such an inherently 'noisy' display driver circuit in an HF receiver - the method used in earlier sets where the channel switches directly generated steady 7-segment data for both displays was better in this respect, although of course it made the switches complicated and expensive.

[dseymo1]

Wouldn't a diode matrix achieve steady-state display switching using only one switch contact per selection?

[Biggles]

Logic chips would save you a few (!) diodes.

[G0HZU_JMR]

On a different note, I noticed last night that one of the Channel readout segments has just become intermittent on this GX-19, but I know where the problem is.

My experience of repairing channel display faults quickly taught me that this part of the CB radio can be a very unforgiving place. Mistakes in probing or wiring can be punished harshly where a poor connection problem can easily turn into "whoops... I just killed an LED segment there..."

I used to run the radio at reduced voltage with a current limit when messing with the display. If anyone ever brought me a broken CB radio I would always promise to fix it for them quickly as long as the channel change and LED display was healthy. Otherwise it could take quite a while to find a replacement part from a scrapper if the display or channel changer was dead or damaged beyond repair.

So my advice would be to do as much of your work with the radio switched off and disconnected from the PSU. If you do decide to try and do any voltage probing then insulate your DMM probes right up to the very tip of the metal probe so you can't short any exposed metal parts nearby if you slip or wobble a bit when probing.

[SiriusHardware]

Originally Posted by dseymo1

Wouldn't a diode matrix achieve steady-state display switching using only one switch contact per selection?

Yes, but, 40 channels, 40 channel rotary switch with a single common and up to 14 segments to provide 'on' states for for each switch position... quite a lot of diodes! The least segments lit (therefore diodes needed) on any channel would be 2, for channel 1, and the most 12, for channel 38.

If the channel switch generated an ascending BCD count (2 X 4 bits) then that output could be taken through 2 BCD-to-7-segment ICs (7447) for steady-state drive to the displays.

Whatever disadvantages the use of a multiplexed display driver might have for HF reception in theory, there are certain signature faults (such as the same segment being faulty on both displays), which are more likely to be caused by a wiring / solder fault than a fault on the display. But if, on the other hand, the problem lies with just a single segment on one display only, then the problem is almost certainly in the actual display.

[Damo666]

Sorry about the late reply guys. I was meant to respond yesterday but one of my kids distracted me.

Re' the segments on the channel display; I can see that there is at least one poor joint on the PCB that's connected to the display, and when it's wiggled slightly, the segment reappears. I think there's 2 segments that intermittently flash - but I didn't really look too much into it as I've been preoccupied with other stuff. I'll report back soon with a more in depth post.

I recall having one of these Pama GX-19 sometime in the late 90s & that also had an issue with a couple of intermittent segments, so I'm guessing it's a design flaw - because looking at the PCB that connects to the display looks a little stressed/forced - as one of the tracks has very slightly lifted.

Sirius, that box spanner looks just the ticket - so I'll buy one soon.

Many thanks again guys, and enjoy your weekend.

[Damo666]

In relation to my previous post I was mistaken; The mic' socket is in fact being retained with this circular type nut with indents as shown below - so does anyone know the name of the tool required to remove this, please?

The squelch, volume, & channel changer are being retained with nuts, so I've purchased a set of box spanners for them.

[TonyDuell]

I would probably use my adjustable 'spanner wrench' that is sold for removing/fitting the locking rings in camera lenses, etc. But that's an expensive tool to buy for one job (I work on mechanical cameras as well as electronics, so I need one). Microtools sell them if you are interested.

The other 'right' tool is tubular with the end cut to have 2 of 4 projections to fit in the notches. You might be able to make one by filing a piece of metal tubing.

Normally you can get a pair of long-nost pliers to engage in the slots. It's not likely to be very tight.