Heathkit HFW-1 Wobbulator Woes!!

[1100 man]

Good evening,
I recently purchased this to assist in the alignment of 405 line TV IF's and more importantly, to increase my knowledge of what goes on in the signal stages of a TV in more detail. It's all part of a learning curve for me.
So my range of operation will be 30-40 MHz.
Having studied the circuit and circuit description, there seems to have been some effort taken with the design and the unit is well constructed.

With the sweep control set to minimum and the frequency to 30 MHz, I connected the RF output to the scope and saw pretty much what I expected- a period of oscillation of constant amplitude followed by a period of blanking. So far so good! Increasing the sweep control should increase the frequency of oscillation during the sweep period. This wouldn't be noticeable on the scope, so the trace should not change much as the sweep control is advanced. That is how I understand it anyway.
As the sweep control is advanced, however, the amplitude of the initial part of the sweep increases and that of the latter part decreases so you get a very trapezoid shape. This doesn't seem right as surely you need a constant amplitude over the sweep range?

With the sweep control again set to minimum, I found that the amplitude of the RF output varies radically as the oscillator frequency is adjusted. Even over a fairly narrow range of 30 - 40 MHz, the variation is quite large which seems to explain what I see as the sweep control is advanced.

1/2 of V3 (12AX7) is an AGC amplifier which controls a shunt regulator (V4). This varies the HT supply to the oscillator valve to compensate for changes in RF output
To compensate for different oscillator efficiency on different bands, different DC potentials are applied to the grid of the AGC amp by the range switch.
So they seem to have gone to a fair effort to ensure constant RF output.
As the frequency is altered, the RF output varies massively. There doesn't seem to be any noticeable effect on any part of the AGC system. The anode voltage of the shunt regulator does vary by about 4V when the range switch is altered from range A to D but that's about it.

I have spent some considerable time on this trying to ascertain what's going on. I've measured voltages, checked component values, checked the wiring against the circuit diagram but to no avail All the voltage readings are pretty consistent with the manual. I may, of course be barking up completely the wrong tree The only other thing of note is that the actual frequency of the oscillator seems very low compared with the scale markings. 33MHz actual, 40 MHz indicated. 38MHz actual, 47MHz indicated.

So can anyone shed any light on what is going on or does anyone have one of these that they could make some comparative measurements on?
All help gratefully received as I am now pretty well stuck and am not sure what to do next!!
I will post some scope pictures and also the circuit and description.
Many thanks
All the best
Nick

[ronbryan]

Nick

My HFW-1 seems to reduce amplitude as the frequency sweeps higher, but all of my observations were around 10.7MHz, not 40MHz. I checked / changed various components and bought a new shunt regulator 6CL6 but it made no difference to the linearity of the RF output when swept. I have a couple of basic questions about the measurements, though.

What is the bandwidth of the oscilloscope you are using?

Have you made the amplitude measurement with a x10 scope probe or with a co-ax cable terminated with 50 ohms at the scope input? The latter method is preferable at these frequencies.

A bias current for the 'Increductor' operating point is provided by a selenium rectifier M1. Is that in good condition along with its filter capacitor? If the bias is wrong, it may be that the inductor is operating at the wrong point on its B-H curve.

Ron

[ronbryan]

In the previous post, I added the following.

A bias current for the 'Increductor' operating point is provided by a selenium rectifier M1. Is that in good condition along with its filter capacitor? If the bias is wrong, it may be that the inductor is operating at the wrong point on its B-H curve.

Although selenium rectifiers do have an aging problem and M1 might be faulty, the bias circuit via M1 that I referred to is intended for centreing the sweep frequency, rather than making the sweep linear, so I think my suggestion may be a red herring.

Ron

[1100 man]

Hi Ron,
Thanks for your reply. Yes, I had checked the selenium rectifier & 40Mfd cap and all seemed OK. The bias voltage across the primary of the Increductor was correct according to the manual. For the moment, I'm keeping the sweep control at minimum and just trying to get my head round what is going on at the RF output at different frequencies.
The RF is currently connected to the scope with a direct lead (RG79 75 Ohm co-ax). I've just tried 3 different leads and they all give different amplitudes on the screen!! There is no termination at the scope end- I don't have a Tee piece or terminator with me. Maybe this is causing problems? I've just tried a scope probe and even on x1, the ampltude is only a 1/4 of what it is direct.
The RF output impedance from the generator should be low, it comes from a cathode follower and is tapped off a 300R pot.
The scope is specified as 40MHz.
I'll make some more detailed measurements this evening as I really should be doing something more industrious than playing with toys
Many thanks
Cheers
Nick

[ms660]

A simple RF detector probe connected to the scope should tell you what's going on.

Lawrence.

[ronbryan]

Nick

An unterminated length of co-ax will undoubtedly give rise to measurement errors at 30MHz. The output impedance of the HFW-1 attenuator is specified as 50 ohms. I use a BNC 'T' and a 'thin Ethernet' terminator for low power measuremets as I don't have a feed through 50 ohm terminator.

If the bandwidth of your scope is 40MHz, I would expect the response of the 'Y' amplifier to be at least 3dB down at that frequency, so Lawrence's suggestion of using an RF diode probe seems a good way to work around that issue.

I've attached a diode probe circuit suggested by Roberts in their R700 service sheeet. No doubt there are many others on the web.

Ron

[1100 man]

Thanks for that, Ron.
I have found a 50R BNC terminator and will pick up a 50R tee piece from Maplin (of all places) when I'm out this afternoon. I've also found a 50R BNC lead, so will do some experiments later.
Many thanks
Nick

[Argus25]

Nick,

There is a very interesting lesson in this, since you plan to sweep TV IF/RF stages.

It is really important that the lead from the generator is terminated in the correct impedance and that when you attach the generator to the antenna terminal of the TV set that the TV's RF input is not incorrectly loaded or the first resonant circuit will be thrown off.

(you can sweep the input regardless of whether it is a superhet with a tuner or a typical TRF circuit)

So often it is recommended that a resistive padder matrix or "dummy antenna" is used to match the generator's impedance to the usual 300 or 75R of the set's RF input, or even just have a 3 or 6dB resistive ladder attenuator to avoid spurious tuning effects even if the two impedances were already matched.

If these issues are not attended to and the frequency response of the actual sweep signal at the set's RF input has a skew on it over the scanned frequency range, you can end up setting up a video IF or video RF stages to look perfect on a sweep, but in reality for real received signals via an antenna, it will be miles off.

Another tip: One place to look at the results of the sweep (which fortunately only requires a low bandwidth scope), where the signal levels are high, is the actual video output on the plate circuit of the video output stage, because sometimes loading of the video detector, when a scope is placed there, can alter the response of that stage even with relatively low Z of the video detector output.

[1100 man]

Hi Hugo,
Thanks for your input!
My main focus at the moment is to make sure that the sweep generator is actually producing a constant amplitude sweep before I start connecting it to a TV! Always test the test equipment someone once said!
I tried a 50R BNC lead and 50R terminator at the scope end but apart from reducing the amplitude (as you would expect) the amplitude variation at different frequencies is very significant. As Ron suggested, it could be down to the Y amp on the 'scope. However, if I use a lower frequency range on the generator, there is still significant variations which would be less likely to be the scope.
I am still sure, though, that the generator is not behaving as it should. I am sure the AGC system/ shunt regulator is not controlling the amplitude of the oscillator.
Unless the sweep is flat, it's not going to be very useful for setting things up!
My next plan is to knock up a diode probe if I can find a germanium diode. I assume this will rectify the signal so I can look at a DC level on the scope for different frequencies?
As you can tell, RF is not my strong point: having been a repairer of TV's during the '80's & 90's, there were seldom RF signal faults to deal with.
The main purpose of this entire exercise is to improve my knowledge of this area of vintage TV's. I now also have a practical need as I have two TV's which need alignment.
So far, its having the desired effect: its making me have to think about what's going on!
I'll post the circuit and description below and try and find a diode to make a probe tonight.
Many thanks
All the best
Nick

[karesz*]

Hi Nick,

My next plan is to knock up a diode probe if I can find a germanium diode.

You can use most Schottkys too, ie HPs 2810 or BAT-family...
We have discussed here too and you can find lot of ccts between simplest and highend solutions on radioamateur sites too. In RF jobs is very important a clean work with short as possible component leads and, proper components and bypassings, terminations and shildings...
Karl

[1100 man]

Hi Karl,
That's very helpful: many thanks for posting the links. I'll have a read through later.
All the best
Nick

[Argus25]

Originally Posted by 1100 man

Always test the test equipment someone once said!
Nick

A very smart move.

One thing about these vintage TV sweep generators, they are very easy to use , sometimes easier than more modern ones. Often they have dual tuning controls, one for the sweep and one for the markers. While the sweep generators are great, because they just cover a good range of frequencies, the marker generator is not much chop because its calibration is often just a dial and its off accuracy anyway.

So over the years what I do for setting up the band pass response of TV video RF/IF stages, is to use the sweep generator from the vintage valve unit to acquire the the curve for the bandpass response on the scope, but to give it accurate meaning I use a more modern signal generator with a digital display of frequency like a PM5326 generator to create the markers.

All that I do is get the output from the PM5326 generator and clip the alligator clip onto the insulation of the wire that is the output of the valve sweep generator. This creates a fraction of a pF gimmick capacitor, so that the coupling of the marker doesn't affect the sweep response.

Then I turn the output level of the PM5326 up just enough to see the "birdie" this gives me the accurate frequency marker superimposed on the sweep waveform out of the video detector or video output amplifier.

Of course to get the best out of a TV set, if you want a sharp crisp image it is necessary that the video bandpass response of the IF (for a superhet) or in the RF circuits (for a TRF design) is correct and has the right or adequate bandpass response. (Obviously, also the crt beam has to have a sharp focus throughout the scan and also the video output stage has to have the appropriate bandwidth too)

I became obsessed with setting up video IF amplifiers when I was about 17 years old and I realized I could drastically improve the images on many monochrome TV sets by adjusting them correctly.

I got a bit disappointed when surface acoustic wave filters for TV IF's came out with the correct bandpass response built into them and I could no longer adjust the bandpass response, but that is progress.

Hugo.

[1100 man]

Hi Hugo,
They seem to have thought of the marker accuracy on the Heathkit. There are two marker generators, one variable and the other crystal controlled with a removable front panel mounted crystal. The output of both oscillators is mixed together and the output can be viewed on a scope. It's then obvious when the variable oscillator is tuned to one of the harmonics of the crystal. These harmonic 'beats' are then used to calibrate the variable oscillator over it's whole range. Seems quite a neat idea to me!
Cheers
Nick

[1100 man]

Hi,
I managed to steal an OA81 from the TV I'm working on, and built the components of the RF probe (linked to by Karl above) directly onto the RF output socket of the generator.
So the centre pin of the socket goes via a 1N capacitor to a 220K resistor and then via a short length of co-ax to the input of the scope. Between the cap and 220K is the diode to ground.
With the frequency set to 33MHz, I get a DC level of 350mv on the scope. Increase the frequency to 40MHz, and the reading goes up to 380mv. At 70MHz it goes up to 450mv.
So assuming my measurement technique is sound (which it may well not be!), that's more variation than I would expect or want.
Using the two lower frequency ranges on the generator gives much less variation as the frequency is adjusted.
They seem to have gone to some effort to incorporate AGC and a shunt regulator which according to the text should keep the amplitude over a given sweep range constant. I am still sure that this is not working as it should although I have spent some time investigating/ testing the AGC circuitry.
So my quest for constant amplitude over my desired sweep range, 33-40 MHz, continues
All the best
Nick

[Argus25]

Originally Posted by 1100 man

Hi Hugo,
These harmonic 'beats' are then used to calibrate the variable oscillator over it's whole range. Seems quite a neat idea to me!
Cheers
Nick

Nick, Yes as long as you are sure which beat or birdie represents some exact frequency that is fine too. I find it very handy to have the variable generator connected though, so I can run that across the curve to any point and see what the exact frequency is on any part of the response curve. Its sounds as though your generator achieves exactly the same effect, but its nice to have a digital frequency readout too.