Taylor 65B Sig Generator: Dodgy Waveform
 

The output waveform from my Taylor 65B is not what I would expect.
Instead of a symmetrical amplitude modulated wave, only the bottom half of the wave is modulated. The top half is flat.
The 400Hz from V2 is clean.

I have replaced the usual caps and a few out of tolerance resistors but the output is still dodgy. Can anybody shine a light on why the output would be asymmetric?

Thanks.

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

I am not sure what you refer to, but taking a quick glance at the 45B (not 65B) manual reveals "MUTUAL CONDUCTANCE: In this test a choice of anode, screen and grid voltages are provided. These are AC Except the grid voltage is prevented from going positive by a small rectifier"

I only own a printed version of the 45B operating manual

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

That will teach me for not looking properly, you refer to a sig gen and I stoopidly refer to a valve tester- Ignore me! ;D

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

That's ok - I wondered what you were on about ???
Thanks for replying anyway!

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

This normally happens because the modulator or RF PA (OSC in this case) has gone high source impedance, and the plate voltage of the OSC cannot be pushed high enough on the positive part of the modulating waveform.

Ensure heater volts are correct to all valves, and all are glowing merrily on high.

Try replacing the 6X5, or with two 1N4007s etc and a 330R 5W resistor.
This will remove doubt about the emission, and whether it was half or full wave rectification due to possible fault.

Replace 6J5s one by one.
The emission of V3 is high suspect.

You don't say which resistors you have tested or replaced.
There's only a few components.
So, ... test ALL resistors R1, R2, R3 etc.

Have a look for crook earth connection on OSC valve cathode.

If all tests OK, consider if the Phantom Repairer has been here, and wired a component incorrectly.

Wiggle all valve in their sockets whilst probing RF OP with CRO.

Do you have a service manual with expected volts on each valve pin?

Good luck.

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

This is a re drawn version of the osc.

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

You immediately won my heart with the use of graph paper.
.

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

Thanks for the information and advice.

I checked all the resistors and replaced three:
R2 11.0K --> 8K2
R3 59.4K --> 47K
R4 40.1K --> 33K

The RF out pot VR2 seems to be much increased (6K8 instead of 2K), but I can't imagine this would be a problem.

As V2 and V3 are both 6J5's, I swapped them, but it made no difference.

The original instruction manual doesn't include voltages.

Measured Voltages:
V1 Cathode = 194v
V2 Anode = 108v
V2 Cathode = 2.9v
V2 Grid = -1.6v
V3 Anode = 85v
V3 Grid = -11.5v
V3 Cathode = 0v

From these voltages, I'm guessing that the rectifier V1 is working correctly.
Maybe V3 is too low?

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

Just to be clear, here's a picture of the waveform from the Taylor 65B ...
The second picture is from a modern generator - this is what I'm expecting to see.

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

It is just 2 triodes so don't expect perfection. How does it look on other ranges? What does the modulated HT look like on R4? R5 was OK?

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

I don't mean to hijack (so consider this a parenthesis). I wanted to check: the second picture is what you expect to see when the signal is modulated? I ask because I'm attempting my first alignment check, and I'm using a Taylor 68A just acquired through FCS. I get this waveform with modulation turned on. I am thinking I also should be checking the circuitry of my generator!

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

These generators were designed to be just enough to align a set and to find out where the signal stopped in the case of a dead set. They weren't really up to looking into the quality of reception or of audio. That needed something most repair workshops couldn't afford.

Looking at the redrawn circuit on the nice graph paper, the cathode is directly grounded. The valve has to get some negative bias from somewhere, and it does it by rectifying its own signal. g-k acts as a rectifier diode, driving the grid itself negative. This is a very neat automatic level control, setting the level of oscillation, but it has its prices.

That negative voltage gets expressed across R3 and VR2. DC voltage across a resistance? that means current.... average grid current! The grid current will only be on the peaks of the oscillator's sinewave, so the peak currents will be a fair bit bigger than their average.

This peak grid current must come via C5 from the tuned circuit. The current varies over the cycle, and voila! lots of distortion, These generators didn't usually have an uppermost band, it was usually just a scale showing the harmonics of the next band down.

Wait a minute... I said 'a sort of ALC' a bit back, yup. This just happens to try to fight/counteract the modulation, if you turn that on.

The oscillation level is controlled by the grid cut-off bias voltage, as much as by anything. One of those facts that become interesting only after you know about them (therefore why does anyone go looking in the first place?) is that relying on device non-linearity or cutoff to set the level, means that there is a floor to the range of levels you can make an oscillator run. If you try modulating an oscillator, you won't get much variation and the linearity will be crap. Try to make an oscillator run low in level and it just stops, sits there, looks at you, and blows a raspberry.

So oscillators are only happy somewhere in the volts region. If you want low levels, you're into attenuator territory, several of them, and lots of screening. If you want plenty of AM, then apply it to a subsequent amplifier stage... then attenuate as needed.

If you're ever hunting the causes of EMC trouble, spurious emissions, low levels are a clue. Remembering oscillators run biggish levels, the low level you see means either there is screening/filtering, or else the oscillation on your spectrum analyser is a damped ring only there for a short while on a waveform (Quad 303's famous 10MHz oscillation is a short burst at one point on a sine output.)

DAvid

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

Agree with David, the grid will be passing current and that will tie the top of the waveform to a little above 0V where the grid will conduct so the waveform is as expected. The AVO All Wave version is a better design that uses a separate winding on the oscillator coil for the output.

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

The Taylor 65b and the 68A should be perfectly capable of producing a decent waveform. These pictures although irrelevant are taken from the same exact models from a cold garage where they have been stored for the last ten years.

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

Looks like the scope is not triggered? Try TV mode to help you trigger on the modulation signal.

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

I agree with Stephen, that there must be something wrong with Howard's sig gen. There are reasons why that design isn't wonderful at the best of times, but his example is a lot worse than I'd expect.

Changing the usual suspects hasn't helped, therefore whatever it is, it's something already dismissed as unlikely. It's time for slow and reasoned diagnosis, now the short-cuts have been eliminated. Is the modulation waveform OK? Is the RF oscillator OK when not modulated? Try modulating from a different source (audio amp and modulation transformer) You need to look in overlooked places, hoping to find clues.

David

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

David and PJL - thanks very much for your contributions. If I read you correctly, you are both saying that the circuit design of this equipment is responsible for the asymmetric output and that no amount of tweaking is going to improve it. That 'dodgy' output is basically as good as it gets.

If that's the verdict from the theoretical standpoint, it would be good to also get an empirical view. So, calling all owners of the Taylor 65B - could you please show us your modulated waveforms?

Freya - thanks for your screenshots, but I'm not sure they are proving anything. My 65B can output a decent sine wave, no problem, it's the AM modulated wave that it's struggling with.

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

Ok, i have slowed the scope down for you. This is the 65b with what you should be expecting to see

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

Aha! So it can produce a decent output.
There must be something wrong with mine.
That's very useful, thanks a lot.

Re: Taylor 65B Sig Generator: Dodgy Waveform
 

Here's a list of all the R's with reference to the older circuit diagram;