HMV 901 restoration.
 

And here it is, the HMV 901. This job is not going to be as easy as the Marconi 702. For starters there is no raster, the 5KV EHT is apparent at the anode cap of the CRT, so at least there are no concerns about the mains derived EHT supply.
So the first task is to get the CRT screen alluminated. Possible causes of the no raster fault is no focus volts or the biasing of the CRT is incorrect causing the CRT to be cut-off. The brightness control is in the cathode circuit of the CRT. 20 volts of postive going video is supplied to the CRT grid.

The cabinet presentation is considered by most as being not as attractive as the Marconi version.

DFWB.

Re: HMV 901 restoration.
 

The measured volts on the focus control slider is 1000V, that's OK.
The voltage on the slider of the brightness control is 252V irrespective of the contol setting. Refer to the attached circuit diagram. The brightness control is VR1. It's most likely that resistor R34 is open circuit or there's possibilty that a wire has become disconnected somewhere.The circuit diagram shown here is from the later production receivers which do not have the "Baird" valve.

DFWB.

Re: HMV 901 restoration.
 

Hi David,
I felt at first that the sync unit looked overly complicated... Then I realized that 'sync' unit actually means both timebases, not just the sync separator.
Silly me:-X

Andy Beer

Re: HMV 901 restoration.
 

Hi David,

Don't know if this is significant but...

Teaching granny to suck eggs this, but the red and black wires (VIR cable) connecting the sync unit to the power supply are drawn wrongly and swapped over on the diagrams I have.

Peter

Re: HMV 901 restoration.
 

The raster is restored again. R34 was OK and the fault turned out to be a disconnected wire to the component group board which carries R34.
There still no picture and sound so that will be the next stage of the restoration.

Hi Andy,
That's right, the whole timebase and sync separator assembly is described as the sync unit. Noteworthy is that the sawtooth drives to the scanning coils are described as high sync for the line and low sync for the frame.
It was common practise in many pre-war TV receivers to employ separate sync separators for the line and frame timebase generators. The type of sync separator used in early EMI TV receivers is sometimes called an amplitude filter.

Hi Peter,
It is important to observe the connections of the VIR cable from the power supplier and the focus control. The "hot" cable from the bleeder chain should be the red one and the black lead goes to the control slider.

DFWB.

Re: HMV 901 restoration.
 

The no vision fault turned out to be a badly seated MSP4 valve in the second RF amplifier stage. The attached picture shows two side by side test card images.
The line timebase is running too slow and will not lock into the 10,125c/s sync pulses. The timebase is running at 5,062.5 c/s.
it is possible that the resistor which is in series with the line hold control R22 (50,000ohms) has gone high value.
The frame sync is poor and it is well known that that these EMI sets have excellent frame sync lock. I'll check for the presence of sync pulses at the anode of V1, the frame sync separator.

Another fault to sort out is no sound, just a hiss. It is possible that the local oscillator is not working.

DFWB.

Re: HMV 901 restoration.
 

If the line was running at half-speed, would it not show two complete test cards, side by side?

Re: HMV 901 restoration.
 

That's certainly right. However, when the line timebase frequency is reduced the picture width is increased because sawtooth forming capacitor has a much longer time to ramp up the waveform before the flyback is intiated.

I've reduced the width of the picture but the images on the screen are still overscanned.

The line frequency of the Baird 240 line system was 6,000 c/s.

DFWB.

Re: HMV 901 restoration.
 

Waveform of the integrated frame pulses on the anode of the MSP4 frame sync separator valve. The frame sync pulse amplitude is 100volts P-P. The 100 volt negative going underswing is in fact the flyback pulse reflected back from the blocking oscillator trnsformer. I'll compare these with the waveforms of the Marconi 702. That set has much better frame hold.

DFWB.

Re: HMV 901 restoration.
 

In the Marconi 702 the amplitude of the frame sync pulse measures 150 volts, the negative going pulse from the frame oscillator is 100 volts, the same as the HMV 901.
The oscilloscope traces are from the output of the RF unit of the HMV 901.

The distorted video waveform is the drive to the sync separators, it can be seen the amplitude is considerably less than the perfect video which supplied to the CRT grid. The active video is only 10 volts and that seems to be sufficient to produce a well contrasted picture.

The CRT in the Marconi 702 is good but the contrast is bad. The video drive to the CRT grid is only four volts. Remember these first generation EMI TV receivers do not employ a video amplifier. High level video is supplied to the CRT and sync separators direct from the video detector load resistor.

The HMV 904 which is being discussed in another thread does have a video amplifier valve, an MS4B RF tetrode. The valve also works as an anode bend detector.

DFWB.

Re: HMV 901 restoration.
 

Remarkably compact these sets, width no greater than the average CRT colour sets of far more recent times. Greater in height of course to allow the long necked CRTs.

I was at David's earlier today and saw his beautifully restored Marconiphone 702 model in operation. With his kind permission I took some photos which he's allowed me to upload here.

Cheers,
Brian

Re: HMV 901 restoration.
 

Yes, sorry I forgot to mention that the detector in the HMV 904 was an anode bend detector so you get detection and some amplification for free.

I've read that trying to extract sync information straight from the vision detector was notoriously problematic due to the low level of the signal at that point. But sync extraction after any AF amplification was difficult due to the loss of the DC component?

The use of an anode bend detector seems to be a rather clever way of solving that problem. I don't know if it was used in later sets, but it seems unlikely?

The resurrection of a technique from the early days of radio seems odd even for 1938. In my years of radio collecting I've ever once (many years ago) come across a radio using anode bend detection, that was a Brandes radio from 1927/1928 (they became Kolster Brandes in 1929). Being stumped as to how the circuit worked, I finally found a description of it in the "Manual of Modern Radio" 1933, and even then it was described as obsolete.

The reason for its obsolescence at least in radio work was because it introduced distortion, but in the days of horn and moving iron loudspeakers that was the least of their worries. The distortion issue evidently wasn't an issue for the vision circuit in the HMV 904.

Re: HMV 901 restoration.
 

Does everyone know how anode bend detection works? It occurs to me I only know how it works because I found a 1927/28 radio with it, and found a description in an early book. I've not found it described in any book later than 1933.

At the risk of telling people how to suck eggs. All valves from the triode upwards have a bend (or two bends upper and lower) in their response curve. The usual idea is to bias the valve so that amplification is done outside of the bend within the linear part. Anode bend detection turns that upside down. The bend is logarithmic, lower parts of the bend have very low amplification, higher parts of the bend have high amplification. We can use that to do detection if we bias the valve such that the negative part of the wave hits the low amplification of the bend, and the positive part of the wave hits the high amplification part of the bend. The end result is the negative part of the wave becomes reduced, and the positive part of the wave becomes amplified. In effect you detect (negative wave removed) and amplify (positive part of the wave amplified) at the same time.

But of course the logarithmic aspect of the resultant amplification introduces distortion, which is why it was abandoned about 1930.

Re: HMV 901 restoration.
 

The anode bend detector operates on the principle of biasing the detector valve close to cut-off. It follows that when a modulated signal is present at the grid of the valve, conduction takes place only on the positive half cycles of the signal waveform.
With regard to television receivers, an anode bend detector works only with positive modulation TV signals. In some respects this type of detector should be ideal as a television demodulator. The sync pulses are close to the non linear part of the valve's Vg/Ia curve and the vision modulation is the linear part. Of course it does mean that on weak signals the sync pulses could be reduced in amplitude. However, because the syncs are step pulses it doesn't matter too much if these are in the non linear part of the valves' transfer characteristic.
Anode bend detectors were used in certain low cost radios such as the Champion 784, see attachments.
Returning to my problems with the 901 and 702 mirror view receivers.
The 25Kohm vision sensitivity control in the 901 has failed. To gain access to this part the receiver unit has to be removed. Not an easy task because the CRT has to be removed. A real scary job.
The Marconi 702. I've managed to improve the output from the vision detector, there is now 6 volts of active video present at the CRT grid, but it is still not enough, it has to be 10 volts for a fully contrasted picture.


DFWB.

Re: HMV 901 restoration.
 

In the 20s they tended to use the bottom bend rather than the top bend obtained by operating the valve at saturation.

Re: HMV 901 restoration.
 

A few Philips sets circa 1950 employed pentode anode bend detectors to very good effect,
Models: 385U and 485U employed a UF42 R.F. pentode. Possibly both the same chassis.
Victor

Re: HMV 901 restoration.
 

Yes, the Philips 385U and 485U have the same chassis, which is also used in the 492U and 683U.

Jac

Re: HMV 901 restoration.
 

At last the HMV is reassembled. I have to confess the picture is not as bright as it used to be. Might be a good idea to check those connection to the CRT heater.

DFWB.

Re: HMV 901 restoration.
 

News flash! Smell of burning. The EHT transformer has failed. That's the reason why the picture has become dim, so it's likely the CRT is OK. The transformer windings are very hot. It's possibly the original or an EMI service replacement which was fitted some time in the fifties.
I will have the transformer rewound but for the sake of long term reliability I'm thinking about another method of generating the 5KV EHT in these receivers.
Something that can be hidden inside the EHT transformer box.

DFWB.

Re: HMV 901 restoration.
 

Hi David,

Just a thought, someone on VRAT mentioned an EHT generator module available on eBay from China that uses 3V DC input and kicks out about 7KV, maybe one of these could be used?

Funny how these old sets wait until you put it back in its case then go wrong!

Regards,
Lloyd.