Divide And Conquer

With a High Voltage generator to build, I though it might be a reasonably sensible idea to be able to measure that high voltage. Now, apart from my oscilloscope, which I'm not confident of the calibration of, I don't have any test gear for hundreds of volts at a few tens of microamps.

So, a HT potential divider was called for. This is a straightforward device, simply a suitable chain of high value resistors, in series with the multi-meters own internal impedance. But first, I needed to know exactly what that impedance is!

Now, the manual says 11.5MΩ but how far can you trust a Chinese manual? So I decided to measure it. OK, so how do you measure the internal impedance of a voltmeter? Well, by balance of course!

Here's the set-up - a fixed voltage is applied across the meter and a series resistor. I chose 10V as this makes everything really easy. As I expected the impedance to be in the 10 to 12MΩ region, I started with a 10MΩ fixed resistor in series. If the series resistor matches the internal impedance - the meter would read half the voltage, in this case 5V. It actually read a little more. So I added another series resistor, a 1M5, and it read a little below. So, it wasn't 10MΩ and it wasn't 11.5MΩ! But 11MΩ.

Set-up to measure meter inpedance

So I now knew the exact impedance of the meter, and I want to be able to easily display up to 1kV. So making use of the meters ranges I decided on a 10:1 ratio, so that for instance a 500V HT would read 50V on the meter. This would still give me two decimal places, so 100mV accuracy in the HT reading. The exact resistor chain values were worked out using Ohms law, I wont go into them here but its sufficient to say I needed the chain, including the meters impedance, to give a current of 4.5uA.

As I was using whatever I happened to have in stock, this resulted in a chain consisting of 5x 10MΩ, 5x 8M2, 1x 6M8, and a 4M7 preset potentiometer, set up as a variable resistor (wiper and one end tied together). The use of the preset meant I could swing the final resistor value around the expected couple of Megohms, to calibrate the chain and get exactly 10:1. This was all built 'snake-like' on a bit of Perfboard. I then used my bench PSU, which goes up to 32V, to test and calibrate the chain - adjusting the preset to give a 3.2V reading.

High Voltage Divider Chain (and a random SCR!)

I'll post up a photo of the completed device later.


Ive also had a play with the Euromarine Radiofix unit today. A pair of headphones were found that would work with it. The RDF unit has a mono 3.5mm jack, but although this pair of 'phones has a 4-way 'mobile phone' jack, one earpiece works. Radio 4 LW on 198kHz and RTE-1 on 252kHz are both incredibly strong, but even then the nulls in the antenna pattern are sharp and deep enough to get a bearing! Despite the level of local noise, and the incredibly weak signals, Sherburn (2.3miles), Finningley (21miles) and Leeds-Bradford (21.5miles) can all be received and a bearing taken.

Of course, this unit was designed for maritime navigation, when there were still coastal beacons, and so covers 180 to 400kHz. The aeronautical beacon band extends up to 550kHz, so I will consider retuning it to cover from 250 to 550kHz, if possible.

Bit of G-M H.T.

A long, long time ago, someone kindly donated me a Mullard ZP1481 Mica End-Window Geiger-Müller tube. Whether this was working or not was not known, and it was my intention to build a circuit to test it. Of course, I got sidetracked and its been sat on the shelf waiting for me to get around to it.

So this weekend, it is slated as my no.1 workshop task! My original design, if I recall, called for a rather special pulse transformer, which I didn't have at the time and still don't! The circuit I intend building this weekend uses a much simpler method of generating the required high voltage by utilising inductor back-EMF (flyback) and capacitor-diode multipliers.

With any luck, the circuit will work, the tube will work, and I can then add on a counter/indicator circuit, turning the circuit into a fully fledged Geiger Counter!

No.2 task in the workshop, will be to change the transistor in the EMG channel receiver of the SARBE-5, and see if that solves the poor receive sensitivity issue.

The 2xAA battery holders arrived today, so one is now fitted, with batteries, into the Euromarine RDF unit. All I need to do now is find some compatible headphones!

Spot the mistake?

Yes, I made rather a silly mistake on retuning the SARBE-5 AUX channel - I tuned it to 290.05MHz... which is NOT the frequency for N/E ICF! That should be 270.05MHz!

The tuning range as is of these receivers is maybe 20MHz wide, so it maybe that I will have to modify the inductors to get the frequency I want!

I'll start the scanner running though and see if I can find an active frequency close to the original 282MHz, to tune to for testing off-air.

As for the poor sensitivity of the main EMG channel, ive found little fault-finding information, so other than dry joints, my only target at present is the transistor. These of course are obsolete types, but luckily, the Tx side uses a few of them, that I can pinch.


Update - Although im yet to repair the EMG channel, I have found that the AUX channel, designed to center at 282MHz, will tune down to the NE ICF frequency, and in fact lower -Ive successfully tuned it to 262.95MHz - the Zone frequency for RAF Leeming. Whether the 20uV sensitivity is sufficient to hear much, well, that remains to be heard!

Partial Success!

After working out what I was doing wrong in the test set-up, and also discovering the very small note in the service manual that explained that the AUX channel receives regardless of the state of the 'receive' button (yes it is in the instructions on the back of the radio, but mine are in Swedish or something!), I've managed a partial success in retuning the SARBE-5 - my units AUX channel is now no longer the NATO on-scene SAR frequency 282.8MHz, but instead North-East Initial Contact Frequency (ICF N/E) of 290.05MHz.

C54 AUX Rx tuning

As these are single active device receivers, retuning a channel, so long as its within the range of the LC circuit, is just a case of adjusting the appropriate trimmer capacitor. The corresponding trimmer capacitor for the test oscillator also needs adjusting if the self test is to work! If the desired frequency is much further out, then adjustment of the little tapped toroid inductor will also be needed.

Waiting for a signal!

 The AUX channel is working nicely with the test set. I haven't had it on antenna for long enough to see how well it actually receives aircraft yet - that test will need to be done in comparison with a known good receiver, such as my PRC-344, and will need several hours of listening!

C45 EMG Rx Tune

Unfortunately, the main EMG channel is somewhat deaf! Ive adjusted its trimmer capacitor C45 for best reception and its not a patch on the AUX channel. I don't know a lot about self-quenching super-regenerative receivers, so at the moment I don't know what to be looking at to solve this. I suspect since there is so little in the circuit, that its either a dry joint or a poorly transistor.

The SARga continues...

Following on from my last post of the strange behavior of the SARBE-5 with the antenna input connected to the Marconi 2955 test-set, ive now tried it running on battery power, with a spare 12v 7Ah SLA battery (charged to 11.1v). The issue still exists, but its form is slightly different, sufficient for me to be able to tell the 'problem' tone apart from the 1kHz modulation. The 'problem' tone is maybe 800Hz.

I also found that the set worked properly if the BNC was 'just' disconnected from the test-set, such that the center pin was physically unconnected, effectively becoming a small capacitor.

So it certainly seems its something to do with isolation from the test-set. Now, the direct connection ive made does bypass a 10pF ceramic capacitor at the antenna, but the antenna connection is still isolated from either the Tx PA or the Rx circuits by more capacitors - the antenna connects at one end of what is essentially a 5-pole low-pass filter, with the Tx connecting at the other end, and the Rx connecting in the middle (so the Tx 'sees' a 5-pole LPF, while the Rx 'sees' a 3-pole Pi-filter. I can only imaging that without the antenna capacitor, somehow this is allowing a feedback path or loading the circuits in some bizarre way. Maybe im missing something in the circuit! This is a 2-channel unit, and I have been working from the single-channel unit schematics for clarity. Is there something going on in the 2-channel set? Maybe its something to do with the AUX channel working even without pressing the receive button?

I'll tack a 10pF capacitor into my test-rig antenna patch lead and see how I progress from there.


And later...

Well, adding the 10pF capacitor into the test-bed antenna connection has solved the issue. I still have no idea why a direct connection from the test-set to the antenna filters - with no DC present - caused such a problem. I can only surmise that it was somehow loading the filter, and causing problems with the operation of the receiver circuits.

So with the receivers responding properly to the test signal, I have found that the AUX channel, currently 282.8MHz, has an MDS (minimum discernible  signal) of around 6uV, and is all but fully quieting at the specified level of 20uV. The EMG channel however, currently 243MHz, is not quite as healthy, with an MDS of around 60uV, fully quieting by 220uV, and a distinct 'squeg' before it starts to receive. With any luck this is just a bit off-tune.

So its back now to the service manual, to brush up on the alignment procedures. I'll also check out whether there is anything froody going on as far as power switching to the AUX receiver, that might account for the constant receive regardless of the receive button!

And, if I can ever work out where my compass set is - i'll mark up and etch some battery terminal plates!

Strange Test Results with the SARBE-5

I seem to have something a little odd going on with the SARBE-5!

Testing it today, to see if all is well before attempting to retune it, ive found a few curious anomalies. First, the AUX channel receiver gives audio output whether the receive button has been pressed or not! The EMG channel doesnt do this and works just as expected - needing the receive button to be pressed to listen.

But more of a worry is that as soon as I connect the unit up to the Marconi 2955 test-set, both receivers give a roughly 1kHz audio tone. Now, this would be fine, if that tone was the 1kHz 30% AM modulation of the test signal from the 2955, but it isnt! They give that tone regardless of what frequency the test set is set to, or what signal level, or even if the test set has its signal generator turned on! In fact, it only stops if I either disconnect the BNC connection to the test-set, needing only to do so far enough that the center pin un-mates, but the barrel can remain connected, or if I turn the test-sets power off!

Strange...!

I have a few theories. Im not too fussed over the AUX channel's indifference to whether ive switched to receive or not - thats a minor trouble. Its the inability to test with a direct signal from the test-set thats the big concern. Using a very limited test regime, due to the frequencies being used, I tried the receivers by close coupling between antennas - the SARBE-5's own rubber ducky antenna and a telescopic whip on the test-set. On both channels I was able to get normal receive performance, i.e. just background noise until the signal level was sufficiently increased (around -65dBm) and then the modulating tone was heard. So it seems whatever is causing the problem is a result of direct electrical contact between the SARBE-5's antenna connection and the circuits of the Marconi 2955. The ground connection from the coax patch-lead doesnt seem to be an issue. My thoughts are that, either there is a loop of some sort being formed, possibly including the bench PSU that is powering the SARBE-5, or there is some strange feedback occurring. Bear in mind that the receiver stages are self-quenching super-regeneratives, it could be that their own oscillations are somehow being fed back around!

So, apart from the AUX channels switching, it looks like the receivers work, but the test set-up is causing trouble! Probably the easiest next step will be to eliminate the PSU as a suspect, by rigging the unit to battery power. Proper SARBE-5 batteries are virtually unobtainable in a working condition, but luckily the radios work on a nominal 12v, so one of my spare 12v SLA batteries will do.

If need be, I will have to mock-up some sort of 1:1 RF transformer to isolate the radio from the test-set.

Rochdale Rally - DF and SAR parts

Visited the Rochdale radio rally today, quite a small rally but a few interesting bits and bobs. Tom insisted on a brew and bacon butty, but it was worth taking him as he was a good bargain spotter!

No one had the BN73-302 ferrites I wanted unfortunately, so the Wellbrook loop clone is still on hold. However very quickly spotted another SARBE antenna! In very good condition, and the seller chucked me in a SARBE-5 battery. Now, the battery terminals had corroded away (although ive just found the cells inside still read 11.55V) but the important thing is that I now can measure the terminals and fabricate replacements, so I can put together a modern replacement battery for the mk5 unit.

Picked up a couple of dogbone insulators, that will complete the receive random-wire antenna for the FRG-100. Tom acquired a few oddball bits free just for interest.

Tom also spotted a Euromarine Radiofix Navigator - one of the DF units i'd been trying to get on ebay recently! Bagged that for less than the ebay ones would have cost with the postage. The battery holder inside has corroded away, and ive got a replacement on order - if only i'd known at the time I could have got one from the rally and had it working already!

LF RDF Receiver

So thats on the bench awaiting the new battery holder, 2xAA, but ive replaced the PP3 clip already.

The state of the workshop bench

Interestingly, the center terminal of the SARBE-5 battery is the negative. Once I have measured the connection plate, which ive now removed from the battery, I will make up a replacement using PCB material. The three studs on the edge of the plate fit slots in the end of the battery housing - slots which also exist on the dummy (red) battery! This means I could turn the dummy battery into a battery eliminator, to allow connecting up an external supply.

SARBE-5 Modifications - Working?

Well, its not easy to say!

After making the initial modifications, that is, removing the crystals and the switching board, and disconnecting the Tx power paths via the Tx/Rx reed switches, I found I wasnt getting any response from the unit! I had a bit of initial success but that soon tapered off! I also noted that I didnt seem to be drawing any current.

It does seem to be though that these problems were due to the somewhat too large for the available space crocodile clips I was using to connect power. After swapping the croc leads for soldered on flying leads, ive had repeatable success, although sometimes seemingly different results for the same tests! Good, measurable currents, and plenty of audio.

Before I can go much further though I have to sleeve the newly created soldered connections for the transducer and power wires, so these cannot short to anything. I also want to remove the battery housing (which means breaching the seal where the power wires go through - not a worry as im not going to be immersing the radio in salt-water!) in order to get to and identify the battery terminals.

The next thing then is to connect a temporary antenna feed from the Marconi 2955 and test the receiver stages with real AM signals! If all is well, then its time to look at retuning.

SARBE-5 BITE Results

Ive had chance to put the SARBE-5 onto the bench PSU, and power it up in its BITE mode. It looks like the BEACON mode fails self test, but receive and manual transmit seem to pass. As its only the receiver im interested in, thats quite fine by me.

The next stage then is to get the antenna socket repaired. This time, unlike the mk6, im going to cut the transducer and power wiring, rather than unthread it from the loom, as it is much smaller gauge wire than the mk6.

Empty case showing antenna socket

 The antenna socket has a small piece of PCB material, plus a small ceramic capacitor. The PCB is slotted and serves as a keying point for the main board. With the PCB and capacitor, and the transducer, removed, the antenna socket could be reseated and the nut tightened. The transducer was then reinstalled, and the PCB and capacitor soldered back on.

Top side of electronics

Under side of electronics

 The above views of the electronics package show the complexity and compactness of this unit. The two Tx crystals are easy to see, and being socketed were very easy to remove! Next step before refitting the electronics is to totally disable the Tx by disconnecting the reed switch contacts feeding the Tx and Beacon sections with power.

The above is the beacon switching board, and its specialist IC. Ive removed this as it is only used in the Tx function, and the IC itself is of interest. It shouldnt have any effect on the receivers.

SARBE-5 and 80cm Loop Antenna

I have now received the SARBE-5 (BE375). It is complete, with safety pin and antenna, but it has some internal damage that requires repair, this being a stuck fixing screw (however the stud had broken off so it isnt hampering anything), and a loose, and hence broken connections, antenna socket.

This is a twin UHF unit, with 243MHz primary and 282.8MHz auxiliary channels. In the case of this radio then, unlike the V/UHF SARBE-6, the two channels are entirely independent of each other - there are two Tx crystals, and two Rx sections. This gives me the option of re-tuning the receivers for two different monitoring frequencies.

My first task then, will be to repair the antenna socket. My second task is to ascertain what the battery terminal connections are! That said, I can temporarily connect the bench PSU to the internal battery connection posts. This should allow me to test the unit using the BITE functions, even without a working antenna connection.

One thing I am interested in, is how the safety pin actually works? It was my thought that withdrawing the pin, activated the beacon by 'pulling' the on/off switch around 90° so that its magnets aligned with the reed switches, but the pin is a flat bar, and slots into the control switch with it in the 'ON - MAIN' position. At this stage I can only think that the metal of the safety pin somehow modifies the magnetic fields.

Another slight issue with it is that at some point it has had a horrible sticky blue paste added to the seals. The original seal is cork gasket sheet, which I can still get from my local auto spares supplier. Since my car is due its service and MoT, I will be visiting them anyway for filters etc, so will pick up some gasket sheet at the same time.


I now have the 80cm loop antenna working, after a fashion. It is hung up on a curtain rail near the radios, and feeding my FRG-100. Im finding however a great deal of impulse noise across HF. Now, im not convinced this is an issue with the antenna itself, as im also getting this when working mobile HF, even with the engine off, but it is making testing rather trying. I can however notice the directionality of the antenna. Some investigation is needed into the source of this horrendous impulse noise.