Thread: Codan HF4000 Series HF Marine SSB Radio – Repair and Refurbishment

Codan HF4000 Series HF Marine SSB Radio – Repair and Refurbishment

Some tech notes and repair history.
Presented to possibly assist others, as information on this system type is somewhat rare.
Also posted on another website.

After an initial assessment to determine whether this work was worthwhile, I recently refurbished this equipment to a reliable working system.
It is quite old, circa 1982, and was not operational at first.
The refurbishment was for a very good friend, and so normal considerations of repair viability of such an old unit were suspended.
And, it was a memory lane trip for me.

A full Service Manual was available.

The system consist of –
Type 4101 HF Exciter/Receiver and Control, 256 channels, 2-25 MHz
Type 4401 HF 150 watt PA
Type 4402 HF Antenna Auto Tuning Unit

All equipment runs on 12 V DC; a 24 volt version was available.

General
The receiver is a direct conversion type, except for 2182 KHz (only) AM, and a separate parallel superhet RX path is included with a 455 KHz IF.
It is USB only, except for 2182, where AM and USB are selectable for TX and RX.
The 4101 has a very sensitive receiver on USB.
MDS <0.5 uV all bands, except 2182 AM MDS 1uV.

The transmitter is a direct conversion type, with a PLLO delivering 4x Fo to the “divide by four” phasing modulator for each channel, and with the modulator operating at final frequency.
It can receive standard HF Marine frequencies for coms, some Aust (and some NZ) BOM weather, and WWV on 5, 10, 15 MHz.
At initial purchase, owner selectable frequencies were available.
Most of the channels are now out of date, but the project was still considered worthwhile for weather reports and long range emergency coms.



4401 Exciter Faults, Repair and Adjustment
Most front panel switches were sticky due to the external rubber seals going yucky with age.
The push button caps suit C&K 7000 series switches only, and if PCB type switches are not available, then you can trim down a C&K solder lug type to give one half of the lug, and it suits the PCB holes ok.
I discarded the seals and replaced two switches with genuine C&K.
If you are missing PB caps, then replacement switches other than C&K are viable, but you may need to replace all switches to avoid physical mounting difficulties.

4401 Receiver
Receiver PCBs can be removed and components easily replaced.
The Display PCB is removed by removal of all applicable switch nuts, as the switches are directly soldered to the PCB.
The main receiver problems were with old tantalum caps, going short, or exploding off their leads.
Replacing C29 near IC6 brought my receiver back to life.
I replaced a lot of caps “on spec”; I didn't bother to test most.
Tantalum caps should not be run at more than half the rated voltage.
Replaced all the blue 16V tantalum caps (two were SC) with 35V types.
Synthesizer Board – Replaced C7-10, C29. Caps C74, C75 (64MFD) replaced by 105 degree normal electros.
Audio and Control Board – Replaced C24, C28.
Also replaced the OM788 10V regulator with an LM2940-10; an LM7809 will work ok if the 12V supply doesn't go too low (the original is a low drop-out type, like the 2940).
You may as well use the LM2940, as you have to cater for a marine emergency with alternator failure and falling battery volts.
The OM788 cermet encapsulated chip separates with age from the metal mount tab, and overheats. Long term repair is difficult and not worthwhile, as the LM2940 is available

All on-board power supplies and regulators then checked for volts and noise.
Adjusted the Synthesiser VCO using a good Philips Counter and good Marconi Sig Gen, rechecking on received 1 KHz audio tone with Clarifier centred.
Double checked across many channels and for all bands.
The Temp Controlled Crystal Osc, the VCO, and as a result the final freq, are all remarkably stable.
Tweaked the RX input BPF filters ok, and adjusted the Demodulator LSB balance ok at 4 and 16MHz.
Did not adjust the LPFs, as overall RX sensitivity was within spec 2-22 MHz. A Sweep Gen is required here.

The PLO VCO runs at 140-210 MHz and has an EM shield on each side of the PCB.
After removing the shields, I replaced caps C34,36-38, C45-47, C67, and then replaced the shields and tacked the sides of the shields down to earth with solder; all this to remove a minor microphonic effect on receive audio. The caps tested ok to 100 Mohms, and with freeze, but may have had a very slight leakage in circuit under voltage, and affecting frequency.
Amazing.

The receiver had a weird problem on 8176 whereby the rx audio (eg AUS BOM WX) sounded funny on antenna, but sounded ok on a sig gen (Fo+1KHz), and even on a sig gen at 8176 externally modulated with RX OP voice audio from a radio station . It would be ok on LOW sensitivity but broken and distorted on HIGH sensitivity.
Other channels were ok.
I eventually found that the EPROM was miscoded (I think) on 8104 and 8176, so that the RX input 8 MHz band BPF filter was switched out (I think in error), and the receiver was running on a Wide Band setting instead.
The Wide Band increased noise level, and its different noise nature, was triggering the Noise Limiter, which in turn was breaking up the receive signal via the Noise Limiter gate.
I double checked the EPROM output back through the buffer and driver chips, and can only conclude a programming error was made as the correct 8 MHz filter is switched in, for all other 8 MHz band channels.
Wide Band is selected for 2-4 MHz, and also presumably for owner selected channels outside the BPFs range.
My EPROM was marked 90-20081-108, version 1.0, and ISO is unclear, A34254, OR, A3H25H, OR, 43H25H.
The Noise Limiter is supposed to trigger a 10 uS blanking pulse on each noise spike, but I found that the Limiter was of a “re-triggerable” monostable type, not a non re-triggerable “one shot” type, and implemented in BJTs not ICs.
The result is a multi blanking pulse of excessive length, a few hundred uS, when triggered by the increased noise of the 8176 channel on Wide Band.
There is an on-board option to “disable” the Noise Limiter, and so that is what I have done, by soldering two pins to the disable points and using a push-on jumper.
This effectively fixes the 8176 problem on this particular receiver, but negates the NL use.
I downloaded the 27C64 EPROM contents with a view to fixing the 8176 programming difference, and to add a few modern channels, but for now, I cannot work out the programming algorithm by inspection.
The file is attached for your mathematical pleasure, but you are warned that the PLO final frequency is 4xFo, and the EPROM contents are “divider” integers for two separate on board freq dividers, not for the final frequencies, and so it is not easy.
I cannot find any information on the web for the programming of such an old TRX.
I have not contacted Codan.
The first frequency is 2182 and the next 2008, 2032, 2112.
Remember that for 2182 AM, the TX osc is 2182, but for this channel only, the RX osc is 2182-455 = 1727KHz.
The EPROM contains data for TXf, RXf, BPF, LPF, TX/NO TX, EMERG, plus memory spacers, etc.

Has anyone ever analysed and decoded this EPROM??

Transmitter 150W PA 4401
I added a simple shunt form of system polarity protection by soldering a reverse biased 1N1200 20A diode across the battery input, near the exciter fuse.
This blows the main supply fuse if wrong battery polarity is applied.
Series silicon diode protection is not possible if the full 150 watts is required as the series diode has too much voltage drop.
FET maybe.

Transmitter PA PCBs have some components screwed to the aluminium base.
Components can be replaced in some instances by de-soldering and re-soldering on the top side only.
Replacement regulators and TO220 transistors can be soldered onto PCB pins inserted and soldered to the PCB.
Replaced driver board C3, C41 tantalums. These caps exploded, possibly with V16 failure.
Replaced switch V16 TP32A with TP32C. Original failed OC. TP32C has higher volts rating.
Physical and general electrical checks eg earthing, SO239 skt ok, DC cabling, fuse, transistor mount screws etc
Standard SSB characteristics checked to spec ok into 50 ohm DL.
TX Power correct for High, Med, Low.
Carrier bal adjusted ok, power set ok.

Caution driver to PA PCB cable plugs are not wired pin 1 to pin 1.

Replaced connecting cable latching D plug shells with screw attach metal Jaycar types due corrosion.

Auto Antenna Tuning Unit 4202
You really need to fully understand how this works, before repair and test and tune.

You need to tune the ATU for minimum SWR as seen by the PA, for sixteen ATU “bands”, and store the resultant steps into one memory bank per antenna.
You cannot assume that one antenna situation will suit another.

Read the book.

This ATU is a mix of memory stored “manual” major steps of Z transformation, and the tuning L and C, and then an automatic tuning of the minor steps.
Codan should have bypassed this version and made it fully automatic.

But it works.

Replaced the Ant 1 and Ant 2 antenna connect M5 screws with M6 316 Stainless Steel “bolts” made from M6 SS rod, and a nut, silver brazed onto one end.
Much more solid. Originals were corroded.

Replaced top board IC27 OM788 10V regulator on spec with LM2940-10.
Replaced capacitors on spec -
Top board C45 100Mfd
Top board C57 33Mfd tant with electro
Top board C46 470 nFd
Mid board for SWR F and R detector, C1 tant 2M2 replaced by new 4M7 35V tant
PL11 for Ant current was reverse installed - fixed - for techs only - no impact on ATU operation.
Replaced PA-ATU RG58 interconnect cable.

Caution – ATU bottom mounting screw length is important, else screws may contact internal components. I replaced mine with SS due corrosion.
Caution – the ATU earth lugs are DC isolated from the RF Input SO239 socket, and DC battery negative, on purpose.

The ATU “FINE” switch was intermittent on steps 3 and 4. Expensive to replace.
Work around it.

Tested OK into 9m sloping (45 deg) wire antenna with single artificial earth wire on ground under the antenna.
This simulated the intended installation in a 35 ft steel sailing boat with 9m back stay antenna.

Setup requires a VSWR meter which is ok down to 2 Mhz, which almost all are not.

Eventually I made up a HF bi-directional coupler with FWD and REV power meters, and each switchable in five steps from 2W to 200W FSD, and ok from 100 KHz to 50 MHz within 1dB.

You need to know the “next available” EPROM bank for memory storage of the rough tuning steps.
I did not know this for my unit (due to confusion) and trialled a bank with the MSD switch S3 advanced two positions, and the RHS switch S5 set to 1, eg LHS MSD switch changed from 4 to 6, RHS switch set to 1. Worked ok.
If you cannot work this out, you may need to decode the EPROM looking for a free bank, or erase the whole EPROM and start at bank “00”.

Overall
After initial exterior cleanup, I used liberal amounts of Lanox lanolin spray to coat the metal cases exterior and then wiped off.
Repeated. Soaks into paintwork cracks etc and nooks and crannies.

I used 13 mm2 copper cable from battery connectors to the PA, via an Anderson plug/socket, an inline Narva 25 Amp MaxiFuse, and crimp lugs for cable and battery connectors.

You need a 1 Amp 12V supply to test the exciter unplugged from the PA and ATU. Analogue PSU is best to avoid stray PWM whistles etc on RX.
You need a 20 Amp 12V supply to test the whole system and TX power output, otherwise the PS current limit will clip the TX RF on voice peaks. SMP Ok for TX testing.
You can safely run the 4101 Exciter on its own with a PS to the DB25 skt.
You can safely run the Exciter and PA into a 50 ohm load without the ATU. Test the TX into DL not the ATU.

I did not find one dry solder joint in this equipment; worthy of note.
This equipment is well made for its age. Performance is good.

Soak Test
Tested for two months on BOM 8176 RX, each time that I went into the workshop.

Unusual Acronyms
On spec = “on speculation” that it might fail due age or other factors

ATU Manual, Circuits, EPROM file, in following posts.

And now to install.