This set was obtained from a deceased estate in Sydney. The owner bought it in 1957 and was actually employed by the company that made it.
The Ekco brand is well known in the U.K., but its time in Australia was short. It appeared here with the introduction of television in 1956. Products were assembled by A.E.I. (Australian Electrical Industries) at Yennora, under licence to Ecko in the U.K.  While in Australia they were mostly known for television, there was also a popular Ekco mantel radio known as the "Gondola".  I have also seen light bulbs stamped with the Ekco brand. For those wondering where the name comes from, it's an abbreviation of E.K. Cole.
There were three series of Ekco TV sets. The most well known are the 17" and 21" 70 and 90 degree sets from 1957-1958. There is also a 9" portable, the TX-275 which can operate from 12V as well as 240V. The last generation of sets were the 21" and 23" 110 degree sets, which were produced up until about 1961. From then on, Ekco ceased to exist in Australia.

The main design feature that set Ekco apart from the other Australian manufacturers, was the live chassis, series heater power supply. This was standard in the U.K and Europe, but not in Australia. Because of the possible shock hazard from this kind of design, it was never popular in Australia amongst manufacturers or technicians. Radios had been produced in a small quantity for DC mains districts using this technique, but it was somewhat begrudgingly. Australians were used to a chassis that could be safely touched and earthed, regardless of mains polarity.
When Ekco appeared on the scene with their transformerless power supply, they thought it would be a good marketing tool. Cheaper, lighter, and according to the advertisements, "uses one third less power". Obviously they didn't realise that Australian servicemen hated the design. It is quite possible that the notoriety caused Ekco to use a conventional transformer power supply with their last generation of Australian sets.
The unfortunate outcome of this is that despite their popularity at the time, few Ekco sets survive in the present day. One can imagine service technicians advising that the set was "not worth repairing" at the first opportunity.
As can be seen from my article on the Ekco TX-287, my attitude is quite different. Common sense would make one think that all one needs to do is plug the set into an isolating transformer when servicing. It appears the Ekco hating servicemen overlooked the obvious.

Note the warning about the chassis being live. The mains connector is not original, but fits. The original was screwed to the cabinet back to prevent operation with the back removed. Unfortunately, originals in good condition are rare now as the rubber insulation deteriorates.

The design of the sets as a whole really appeals to me. The serviceability is excellent, and there's nothing weird about the design to cause problems.
So, when one of my HRSA aquaintances found this set, I had to have it. Especially as it had a cabinet with doors. A deal was made and it was mine.

I've always liked doored televisions, and to have a series heater live chassis model is perfect!

Despite the lack of a power transformer, this set is surprisingly heavy. I think that is largely because of the thick plywood cabinet. A quick internal look showed a new Thomas CRT had been installed in 1972. Interestingly, it was still a magnetic focus type. Often a conventional electrostatic focus CRT would be used as a replacement, in a set originally using magnetic focus, and the focussing magnets would be removed. One possible reason is that the CRM212 has a 12.6V 300mA heater, so one has to also install a 6.3V transformer if one used, say, a 21CBP4 as a replacement. In otherwords, it's easier in this set to use the original type as a replacement. It's fortunate too, because the focus is far superior to that of electrostatic tubes.

The replacement CRT has no ion trap, but is still magnetic focus. Note the Metrosil VDR across the top. It is for EHT voltage regulation.

The line output transformer had been replaced. This was to be expected, as the original Ekco type was made of a plastic that crystalised in the presence of EHT. It would thus crumble apart leaving the transformer core supported only by the connecting wires. Telecomponents made a replacement type of far superior quality with the transformer supported on paxolin panels. One change with the replacement is the EHT rectifier is now a 1S2 instead of the original 6S2.
A further observation was that the Metrosil EHT regulator and EHT filter condenser had been disconnected. The Metrosil is a long VDR mounted across the top of the CRT. Evidently there had been a fault in either of these causing loss of EHT. With the modern CRT, the EHT condenser is not required, and it is questionable as to how essential the Metrosil is. In view of the focus being very dependent on EHT with a magnetically focussed tube, this would probably explain the Metrosil, but in practice it works very well without it.

The aerial connection had me interested. Most Australian sets used screw or spring terminals, but a few used a two pin socket of similar pattern to a 300 ohm ribbon connector. For my TX-287 I'd been using a plug, which while not correct, fitted well enough. For the TCX-298, the plug was still in the socket with an inch of wire still attached; this being transparent figure eight mains cable, popular in the 1950's. The plug was made by Belling Lee, and it was clearly evident looking at the cable entry, that it was designed for 'lamp cord' rather than 300 ohm ribbon. This shouldn't be too surprising perhaps, as lamp cord had been used as transmission line with some early British sets. It's perfectly viable for this, but the impedance is closer to 150 ohms. The line from the socket to the tuner is normal 300 ohm ribbon. I'm not sure if the impedance mismatch was meant to be of concern, especially if the lamp cord transmission line is only a few feet to the wall socket.

The mains voltage selector was on 230V instead of 240V. I hope this wasn't done to mask a fault, because such tacticts over run the valve heaters. Strangely enough, my TX-287 was the same when I first got it. No mains cable was present. It uses a two round pin connector; one pin of larger diameter than the other to keep the polarity. Like the TX-287, it appeared the original connector had deteriorated and the mains wires soldered straight to the pins, as there was solder on the ends. I didn't have any spare leads, as both were in use for the TT-207 and TX-287. However, I remembered I had some leads off some long forgotten test instruments that had a similar connector. Much joy when I found they fitted!
General condition inside was excellent; in fact it was really like new. No grime or corrosion. Looking under the chassis showed that some paper condensers had been replaced, and that the selenium rectifier had been bypassed with silicon diodes. The volume/brightness control had also been replaced.

Broken channel knob.
I'd noticed the channel knob felt a bit spongy when rotated. At first I just put this down to the grub screws being loose, but then it felt really spongy and it was obvious the shaft had broken. Apparently, the tuner mounting on the chassis had moved out of alignment causing the knob to bear heavily against the wooden front panel. So, each time it was rotated, the shaft was under angular stress. The early Ekco knobs are rare, so I had two options; replace it with a white one I had of the same style, and paint it brown, or fix the original. As it turned out, fixing the broken knob was not only very easy, but made it far stronger than it was originally. To start with, I superglued the break back together. This would be fine for a volume control, but given the torque required to turn a turret tuner, it obviously wouldn't last long. It so happened that 1/2" copper tubing was a good fit over the whole shaft. So, I cut a piece, but with notches at the end which engaged in the hollow parts of the paddle shaped part of the knob. Thus when the knob was turned, all the force would be applied to the copper tube only, and thence to the collar containg the grub screws. No force would go through the repaired join.
With the tube suitably shaped,  the plastic shaft was covered in Quik Steel epoxy and the tube slipped over. This epoxy is much stronger than ordinary Araldite kinds.
I suspect one reason few Ekco knobs have survived is that the uninitiated are not familiar with how to get them off. It is not possible to extract the chassis until the knobs are off!
Because of the live chassis, normal push on knobs could be seen as a shock hazard - someone pulls one off, touches the shaft, and gets a shock. So, the knobs are secured with grub screws, and these are not accessable from outside. For the early paddle shaped knobs, there is a ring around the knob, labelled as to the function. This is secured with a wire clip around the knob. This needs to be removed, along with the ring. Then the chassis will slip out with knobs still attached. For the later models with fluted knobs, it's similar, but the outer fluted part is pulled off.

Chassis restoration.
Like with any other 1950's television, getting the set going again is largely a matter of replacing all the resistors over 47K as these drift high, especially if they're made by IRC. Also, all paper condensers must be replaced wherever leakage could cause problems. For example, a paper condenser across a 150R cathode resistor will not cause a problem if it's leaky, but when used as a grid coupler it certainly will. And, in recent years, mica condensers have become a problem, so they usually need replacing too. The only paper condenser left in the TCX-298 is the 2uF across the filter choke, and the .1uF across the mains. A small amount of leakage here will cause no problems. The mains filter condenser is a sealed type so I felt reasonably confident leaving it in. I would have replaced it had it been a waxed dipped type. In any case, it's after the mains fuses, so no catastrophe if it fails.
Replacing all the parts in the Ekco is easier than most other sets, because instead of using tagstrips with the wires wrapped around them, insulated pillars are used and the leads simply slip off as the solder melts. I had the whole thing done in a couple of days.

Under the chassis after restoration. Towards the front of the chassis is the 20R 20W resistor occupying the space where the selenium rectifier was.

Prior to applying the mains straight to the set, I decided to reform the electrolytic condensers across the B+ line. As the set had not been switched on for many years, and with the mains fed rectifier, there's not much to limit the current if they should be leaky. So, I simply unplugged the CRT so the valve heaters would not draw current, and connected a 40W light bulb in series with the set. After about 20 mins the bulb had dimmed sufficiently to indicate leakage current was now at a safe level.

I had noticed early on that the PCF80/9A8 converter valve was cracked, so this had to be replaced for the tuner to function. I also fitted a set of channel 3 biscuits to the tuner.
As the sets were sold, they only had biscuits for channels 1, 2,6,7,8,9, and 10 fitted. At that time, rural TV services were off in the distant future so only potential capital city channels were catered for. Some set manufacturers did not include all ten channels, probably as a cost cutting measure.
As is well known, that was all to change by the early 1960's with a revamped 13 channel plan, and the introduction of services outside the capital cities.
It so happened that one of my later model spare parts chassis did have all the biscuits fitted for the old 10 channel plan, and used the same tuner. So hence the source of channel 3 biscuits. Channel 3 is one of the output channels for many VCR's and with no off air signals now, it was important to include it. Channel 4 is also used, but I prefer not to use it as it's right in the middle of the now occupied FM band and is thus susceptible to interference. Besides, the old channel 4 frequency is completely different. The other channel used for RF modulators is channel 0 or 1. The closest channel to this in the old plan is 1. This can be tuned down to the present day channel 0, which is closer in frequency than the modern day channel 1.

Power Up.
On full power up, a snowy raster was present which looked promising. Next thing to do was to feed a signal in. Initially I used one of my ancient JVC HR3660 VHS machines on channel 3 but I couldn't receive anything except some sound. I adjusted what I thought was the local oscillator coil but no change. After wasting some time on this I then recalled the Ekco tuner was different, and checking the manual reminded me of what I'd done with the TX-287. What I'd been trying to adjust was actually the RF coil, not the local oscillator.
The local oscillator coil is actually adjusted from the side of the tuner through one of the holes that lines up with each coil. And, so I was able to tune in the VCR.
Picture was reasonable given the 1994 recording of the SBS test pattern, but the sound was weak and full of frame buzz. The picture centering and height and linearity were touched up.

Focus control is user adjustable.

Next was to get a live signal into the set. For this I used a digital box to which I'd fitted a channel 0 modulator. I was able to tune down the channel 1 oscillator and RF coils with no problem, unlike with the TX-287 where I had to add 2.7pF to the oscillator coil to get it to tune down sufficiently. However, some interference was present, at the rate of 50 cycles, modulated with something a lot higher in frequency. It appeared to be due to the switchmode PSU in the digital box. Typical! Moving the box away improved things somewhat.
Just to be sure, I tried my Arlunya pattern generator. It produced a perfect interferance free picture. OK, so nothing wrong with the Ecko; it was the digital box.
However, the sound was still problematic; weak and full of frame buzz. Adjusting the fine tuning helped, but the optimum position for better sound was not the best for the picture.

Sound problems.
Time to check the sound stage alignment. And here, I found it quite a way off. I noticed the core for one of the sound IF coils was loose, so it's quite likely that vibration over the years had changed its tuning. I used my Rhode and Schwarz SMS signal generator to realign the entire 5.5Mc/s FM stage. This fixed that problem and sound was now good.
One thing I'd noticed during the restoration was that the audio feedback resistor between the PL82/16A5 plate and the PABC80/9AK8 plate was 680K. In the circuit it was 1.5M.
I then discovered that some Ekco literature had advised the change to obtain another 3dB of gain. Given that there wasn't a huge reserve of volume, I did the modification and changed the resistor to 1.5M.

First Faults.
At this point the set was basically restored and working. First trouble was the set would go off frame frequency after time. Since I'd changed all the resistors and condensers, it didn't leave much, so tried the 12AU7 blocking oscillator. That fixed that.
After several days of running, it was looking very good. But then, I came into the room one day to find the line hold was way off. I readjusted it to lock the picture, but it was right at the end of the pot adjustment. Well, I'd changed all the resistors and there were no paper caps (so I thought!) so what else could it be? The 12AU7 line oscillator valve could be faulty, so I tried that but still no joy. Out with the chassis and another look. Of course, it  might not be the line oscillator itself, but could also be the AFC circuit, as this controls the line oscillator frequency. There were two mica condensers I hadn't replaced; those feeding the 6AL5. The output from the 6AL5 is a DC control voltage for the line oscillator, so anything affecting the input, such as intermittent mica condensers, could be the problem. Initially, I hadn't replaced these because they looked like replacements, and the voltage across them was quite low anyway. But, replacing them with new polystyrene types would definitely eliminate the problem.
As I was removing the old micas, I discovered I'd been caught out. For underneath them, hidden from view, was an original 220K and a .002uF. These are the filter components for the AFC detector, and if faulty will change the line frequency. Well, guess what, the .002uF was leaky. The fault has not returned since replacing it.

Selenium rectifier.
One thing that had always been on my mind was the rectifier replacement. It's well known that selenium rectifiers do fail, resulting in low B+, so to find two EM404's in series replacing it were no surprise. The strange thing was, that while the diodes had been soldered to the selenium rectifier terminals, the terminals were then cut off the rectifier, leaving the diodes suspended in mid air. An important difference between silicon and selenium diodes is the much greater voltage drop across selenium rectifiers. A resistor should be included when substituting with silicon diodes, but this had not been done. No surprises to find the B+ about 20V higher than it should be. This also explained the excessive picture width.
I found a 20R 20W resistor to produce the correct B+. I also replaced the two EM404's with a single 1N4007. EM404's have a PIV of 400V, and while two in series would appear to increase this to 800V, there is no guarantee the reverse voltage would be equal across both. Better to use a 1000V diode instead. To help protect against spikes on the mains, a .01uF 1KV ceramic condenser was connected across the diode.
Out of curiousity, I did try the original selenium rectifier just to see how bad it was. Imediately I had the answer as to why the terminals had been cut off supporting the EM404's!
A jet of flame shot out between one of the cells around the middle of the rectifier. It left an odour throughout the house for the rest of the day.

Above the chassis. The frame hold pot has been replaced, along with the line output transformer, the frame blocking oscillator transformer, and the volume/brightness switch pot.

Finishing touches.
The tuner mounting bracket was readjusted so the channel knob lined up in the front panel hole. It's easy to see why it had moved, as it's a natural reaction to lift the chassis by that corner. The insulating bushing was missing from one of the back of the chassis mounting screws. I replaced it with bushings intended for power transistors. It would not have been a shock hazard without it, as the bracket that the screw goes into is not accessable from outside the cabinet, but it looks better to have things correct.

Chassis reinstalled. The now unused EHT condenser is visible right at the top right mounted on the CRT bracket.

Now let's look at the circuit.

Video stages.
The IF from the tuner is fed into the video IF strip based on four 6BX6's (EF80's), tuned to approximately 36Mc/s. "Approximately" because stagger tuning is used to obtain the necessary bandwidth. Unlike most other sets, the Ekco uses single tuned IF coils, rather than double tuned transformers, to couple the stages. This makes alignment very easy, and it can be easily done without a sweep generator.
A germanium detector diode feeds the 15A6/PL83 video amplifier pentode in the normal way. DC coupling is used, so bias for the PL83 is dependent somewhat on signal. The plate of the PL83 is also DC coupled to the picture tube cathode. The picture tube is a CRM212. CRM presumably stands for "Cathode Ray Mazda", and 21 being the size. It's a 21" 90 degree magnetic focus tetrode tube. The original used a bent gun with ion trap, but the new replacement has a straight gun with no ion trap; the screen aluminising protecting against ion burn. To fit in with the series heater circuit, the heater is 12.6V at 300mA.

AGC.
AGC is of the gated type, using V7, 6BX6, as the AGC rectifier and control. V7 operates as a variable shunt rectifier diode which develops negative voltage from pulses obtained from the line output transformer. Video signal is fed into the cathode of V7, as a grounded grid amplifier.  However, because the plate voltage appears at line frequency, it means the valve conducts during the portion of video signal when the line sync pulse is present. Thus, the strength of the line sync pulse determines valve conduction, and thus the negative DC developed at the plate. This DC is filtered and fed to the grids of the IF and RF amplifier valves. A voltage divider is used to give greater AGC voltage for the RF amplifier, than for the IF amplifiers. 6BX6's are sharp cut off valves, so applying too much negative voltage will cause distortion. Note the presence of a local/distant switch.  In the local position, the IF valves are given greater AGC voltage.
Contrast control is via the AGC, rather than the video output stage. The contrast pot alters the overall conduction of V7 by controlling its grid voltage.

Sync separator.
V12, another 6BX6, is operated without any initial bias. Its grid is fed with the high amplitude video signal from V6's plate. This causes V12 to produce a distorted video signal at its plate. So distorted that only the sync pulses are produced here. From the plate of V12, the sync pulses pass to the frame and line oscillators.

Frame oscillator and output.
The frame oscillator is of the blocking transformer type. It's based around pins 1,2, and 3 of V15, a 12AU7. Some degree of frequency control is available by controlling the grid voltage. Unlike most sets, the B+ for the oscillator stage comes from the normal 230V B+ and not the B+ Boost which is usually around 500-600V. Pins 6,7, and 8 of V15 are wired as an interlace diode.  Interlace diodes are probably unfamiliar unless you are familiar with 405 line TV receiver design. Outside of British designs you're unlikely to see them. With the 405 line system, loss of interlace causes a very obvious line structure to appear. This is because the lines pair, giving only 202.5 lines. Hence, some care had to be taken to ensure clarity of the frame sync pulses to ensure good interlace. Good interlace can be upset by the entry of line pulses into the frame oscillator. One way this can happen is via stray coupling due to wiring and component layout, and even because of the magnetic coupling between the line and field windings of the yoke. Another cause of the problem is line sync pluses getting to the frame oscillator. And it's here the interlace diode is used. Sync pulses are coupled into the oscillator via the diode, but once the oscillator is triggered, the diode is reverse biassed, thus preventing any of the line sync pulses getting through during the frame scan. Line sync pulses are of course transmitted during the frame sync period.
As the TCX-298 is for 625 lines, poor interlacing is not so obvious, but it's obviously good practice to include the interlace diode.
The output stage is a conventional pentode, V16, type PL82. It has the usual waveshaping components to provide the required sawtooth current in the scanning coils. The height control is just a voltage divider controlling the incoming sawtooth amplitude, while the linearity control adjusts the feedback.

Line oscillator and output.
V14B, 12AU7, is another blocking oscillator, except of course optimised for 15,625c/s. Like its frame oscillator counterpart, grid voltage sets the exact frequency. User adjustment is provided by a pot in the cathode circuit. V14A is simply a DC amplifier which feeds the grid of the line oscillator with DC proportional to the AFC detector voltage.
As is usual, an AFC circuit is provided so that the line oscillator is not directly triggered from the sync pulses. If this is done, it is possible with a noisy signal to have a jagged effect on the outline of the picture, as the line oscillator triggers from noise. Instead, an AFC circuit with a long time constant effectively filters out the noise.
This is based around V13, a 6AL5 dual diode. If it looks reminiscent of an FM detector, thats because it has a lot in common. The 6AL5 is fed with line sync pulses via a coupling transformer. It is also fed with sawtooth pulses from the line output transformer. Essentially, if the line oscillator is out of phase with the sync pulses, a positive or negative correction voltage is developed which controls the line oscillator. C75 provides a long time constant so that interruption of sync pulses for a few lines will not affect the overall oscillator frequency.
The line output valve is a PL36/25E5. The output stage is conventional using an autotransformer. The damper diode is a PY81. Originally, a 6S2/EY86 was used for the EHT rectifier, but with the locally made Telecomponents transformer, a more common 1S2 is used. One unusual feature, as far as Australian set design goes, is the EHT regulation. Here a Metrosil VDR is shunted across the rectified EHT supply. As the EHT rises, the Metrosil conducts more, drawing more current, and thus loading down the EHT. Conversely, if the EHT drops, the Metrosil conducts less. There is also a 1000pF EHT filter condenser. This is not required with modern picture tubes due to the extra conductive coating on the back of the glass.

Sound stages.
Returning now to the video detector, a portion of signal is taken via a 10pF condenser to the sound IF amplifier. As well as the video signal, present at the output of the video detector is also a 5.5Mc/s sound signal. This comes about because the vision and sound carriers are always 5.5Mc/s apart. The output of the tuner consists of the video signal centered on 36Mc/s, and the sound on 30.5Mc/s. When both of these are fed into a non linear stage (e.g. detector diode), the sum and difference frequencies are produced. Hence, 36-30.5 = 5.5. The sound IF amplifier is V8, another 6BX6. Its input and output are tuned to 5.5Mc/s. Following this is V9, the limiting stage. Important in an intercarrier sound system is AM supression. This is because the vision carrier (AM) effectively functions as the local oscillator in the 30.5 to 5.5Mc/s conversion. V9 operates under non linear conditions, rather like the sync separator. Again, it is provided with no initial bias. This means that any AM is clipped from the plate signal. Note that the grid of V9 actually produces a negative voltage to control the gain of the sound IF amplifier. Therefore there is a degree of AGC for the sound IF. This is unusual.
Sound detection is by a conventional ratio detector, using the diodes of a 9AK8/PABC80. The triode of this valve functions as a conventional amplifier which feeds the output valve, a PL82. Negative feedback is provided by a 1.5M resistor from the PL82 plate to its grid. A form of frequency correction is provided by the RC network across the output transformer primary. Because of the live chassis, the output transformer must also provide mains isolation. The speaker voice coils are connected to the chassis via a 1M resistor; high enough not to pass a dangerous current, but sufficient to discharge any static build up on the speakers.

Power supply.
The valve heaters are connected in a series 300mA circuit. Depending on mains voltage, various resistors are connected in series to accomodate for the total heater voltage.
Warm up surge protection is provided by an NTC thermistor. The mains is also rectified, originally by a selenium rectifier, to provide B+ of about 230V DC. This is filtered by two chokes; one for the deflection stages, and the other for the rest of the set. Across the choke which filters the deflection stage supply is a 2uF condenser. This tunes the choke to 50c/s for improved ripple rejection. Filtering is of course more difficult when half wave rectification is used.
Because of the live chassis, a polarised mains connector is used. This to hopefully ensure the chassis is always connected to the neutral. Of course, one cannot always guarantee this because of incorrectly wired power points and extension leads (live/neutral polarity was not even officially standardised in 1957). So, the normal procedure for servicing is to use a neon screwdriver to see the chassis is not live, or better still is to use an isolating transformer. Because of the DC component created by the half wave rectifier, it is necessary to use a transformer larger than first thought. This is because the DC flow causes core saturation and heavy losses. I use a 500VA rated unit, but 250VA would probably suffice.
 

Tuner.
To the final part of the description, the tuner is a conventional 12 position turret type. Fitted biscuits as the set left the factory were for channels 1,2,6,7,8,9, and 10.
The aerial is fed into the RF stage via the isolating components mounted on the aerial socket. The 470pF condensers have a high reactance at 50c/s, so provide mains isolation, but at VHF the reactance is low, so the signal passes unimpeded. 1M resistors discharge any static build up on the aerial.
The RF amplifier is a conventional cascode type, using a PCC84/7AN7 twin triode. Interestingly, there is another Metrosil in series with the AGC to the RF amp. It provides AGC delay so the tuner is not desensitised on weak signals. From here, the VHF signal passes to the mixer, using the pentode section of a PCF80/9A8. The local oscillator is a conventional Colpitts type using the PCF80 triode.

Home