BUSH SW45 output stage weirdness ?

[yestertech]

Having started this some time ago, and now got to the stage where I have signals, I was perplexed by the lousy audio quality.
Measurements made into the audio stages only ( TDD4 + PEN B4) showed that despite the brute of an OP valve ( PEN B4 is 8W RMS-ish !) I was getting very little power before distortion. The audio quality into the large 10 inch speaker was also noticeably 'thin' lacking in bass. Further checks determined that the OP transformer had been replaced at some point . PEN B4 Anode volts were slightly low also, despite the g2 volts ( = HT rail ) looking reasonably OK.
The transformer ratios were printed on the replacement, but these ranged from 60:1 upwards, so clearly not a very good match. The primary resistance was also showing around 500R, twice the value of the original, quoted in the manufacturers service data.

After a rummage in what's left of my transformer drawer, I found a multi tapped 'WEYRAD' output TX, with mounting holes which matched the originals drilled in the speaker mounting plate, so I figured it might be up to the job. The service sheet unusually quotes the primary impedance as 3K, and the valve data states 3k5 anode load. This gives a ratio of around 34:1 in my book. ( for an assumed 3 ohm speaker ) The transformer already had the impedance taps marked on it and lo and behold, one for 3K, with the secondary printed as 3 ohm. Even the DC resistance was correct ,at 247 ohms so in it went.

Output power into the load box now noticeably much better at around 3W /1KHz before noticeable distortion on the scope. Switching up a range showed a fairly decent performance at 10KHz also. Job done ?
er... no !

At the other end of the scale, the response falls off quite drastically from around 300Hz and is well down at 100Hz. As this is similar to the original output transformer the set came with, I thought the fault may lie elsewhere. The scope trace looks pretty bad at this point, resembling a sawtooth wave of sorts.I rechecked the already replaced coupling cap in the audio stages by substitution - no change.
Further measurements show that the TDD4 output measured at the PENB4 g1 remains pretty flat across the same frequency band, i.e. 100Hz to 10KHz. so I can't blame the generator or other circuitry
So I'm thinking to myself possibly transformer saturation ? The PEN B4 draws a whopping 60-70ma, so maybe my transformer is just not up to the job.
I try a multi- tapped RS 'standard output transformer' choosing the nearest ratios, but this is actually worse, as the primary resistance is also now dropping PEN B4s anode volts.
In desperation, I locate a much larger 'hi-fi' sized output transformer, which states it is 5000:15ohms. Not ideal, so I swap to a different output meter to load the secondary properly.
and re-test. Although the output is reduced at 1Khz, the LF output still falls away below 300Hz
and now the HF is bit more droopy too, but I'll have to ignore that one for the moment.
The scope trace IS a bit cleaner at 100Hz, but still rather grim.

My head now hurts as I can work out whats occurring here. I did try briefly bypassing the PEN B4 cathode resistor ( designed not to have a bypass cap. ) but predictably , after allowing for an increase in gain, the result is the same.

Any Ideas

[PJL]

I assume the measurements are into a resistive load and not the speaker? I would agree that transformer saturation is a reasonable guess. You could try tacking a resistor in parallel with the primary to prove it. The recommended load is 3,500 ohms.

[Herald1360]

Try biasing back the PEN B4 by increasing the cathode resistor to drop the anode current to around 45mA. Max output will come down, but the quality may well improve if TX saturation is the culprit.

Is the grid coupling cap OK? That could make the current even worse.

[Boater Sam]

Is the power supply up to providing this level of current or are you getting too little current at low frequencies when the demand is highest?

[Ed_Dinning]

Hi, some of these transformers were designed for use with single ended or PP circuits, but were a bit naughty in that they were interleaved for PP and had no effective air gap.
For a single ended stage the lams should all be stacked the same way and not interleaved. There is usually a thin piece of paper in there as well.
Try dismantling your Weyrad transformer and adding or increasing the paper for an airgap.
This should reduce any saturation problems.

Ed

[john.north]

Hello,
I have a working sw45. I can check the transformer tomorrow.
There is plenty of output. I don't really want it because the case has been eaten by the dreaded worm. If I remember correctly there is a output tx and what I presume is a humbucker coil. I will check.
John.

[yestertech]

Thanks for the replies thus far
i can clarify the few points raised;
The cooling cap has been replaced.
All measurements are taken into a resistive load/wattmeter
The HT rail holds up reasonably well at the higher output. The anode volts seem in the right area.
Yes the speaker is an E/M type ( ROLA ) with a h/bucking coil.

What puzzles me most is that the bass loss 'effect' is audible at much lower levels, when the output into the speaker is probably only 1/2 watt or less.
Even when I test on the wattmeter at 1/2 watt into 3 ohms, the response is still dropping badly below 300Hz.
I like the idea of biassing back to recheck, and also the parallel restive anode load.
I've never seen such a poor audio response from a set.
The original transformer must have been something very special.


A.

[PJL]

Despite being my idea, the parallel resistor won't work as the value would need to be too low. You would have to use a choke in series with it. I suggest you go for biasing it back to prove it is the transformer. How to find an alternative transformer is another matter. There were a few HMV sets that used KT66 output valves so it might be worth a try in the wanted section.

[Wellington]

Originally Posted by yestertech

At the other end of the scale, the response falls off quite drastically from around 300Hz and is well down at 100Hz.

Originally Posted by yestertech

…the response is still dropping badly below 300Hz.

Originally Posted by yestertech

Any Ideas

At those sort of frequencies, could this be something to do with the humbucking coil? If the original output transformer has been replaced, could the connections have been messed up?

I'm pretty much out of my depth on this thread , but I thought I'd put it forward as an idea. Also, do you think it could be the loudspeaker itself?

[Geoff 555]

Hello, to be honest I do not know an awful lot but an emediate thought was the speaker, I have a set (1934) and the cone has gone rock hard with not a lot of movement and frequency response is not good.
Perhaps another speaker tacked on just to see what the result might be would be an idea.
All the best.
Geoff.

[Hunts smoothing bomb]

The O/P has stated that the response rolls of below 300Hz using a watt meter therefor doesn't look like it is the loudspeaker itself.
Has the PEN B4 been subbed yet? Although reading your original post the anode is drawing 60/70mA so doesn't sound like the valve is at fault unless it is soft of course?

Cheers

[yestertech]

PEN B4 is a bit of a hard 'sub' as I don't have one and there's only one equivalent.
I'll run it up on the valve tester again, but I'm fairly sure the emission is good.
Given that it's generating nothing like the quoted 8W, I may change to for something a bit less beefy. I can't see a valve can cause this fault ,but stranger things have happened.

As to the comments about the hum-bucking coil - good point.
Only problem with that is that since tacking in the replacement transformers, I have run the secondary direct to the wattmeter, so it's not in circuit.

I keep thinking I'm missing something fundamental here, but blowed if I can see it...


A.

[john.north]

Hello,
Just looked at the circuit and then checked my SW45.
Mine has a different set up, Straight transformer 409c and a Magnavox speaker. I think these were an alternative and I think they were fitted to the SW41. Anyway mine is as follows, primary 250r, secondary 10r. The field coil+ hum bucking coil gone. The replaced field coil primary which is needed is now a 100r choke + 500r resistor making up the original 600r. The radio works with plenty of power. If you want to try this set up with the Magnavox speaker let me know. I would suspect a faulty speaker.
John.

[PJL]

The saturation of the core is caused by the DC current and will impact all volumes. Loss of low frequencies would be to do with the impact of saturation on the apparent inductance. As Ed mentioned earlier, this is normally circumvented using an air gap so the coil is not fully enclosed by the core, typically by a cardboard wedge. Lower the anode current to say 35mA by increasing the grid bias resistor and see if the low frequency response is restored.

[Silicon]

When you look at the schematic almost all of the capacitors from C20 to C28 will have an effect on the 'tone'.

Have you checked them all to see if the capacitances match the schematic?

[yestertech]

The signal being delivered to g1 of the PEN B4 is essentially 'flat' at the frequencies lacking in the output, so that should absolve any of the other caps.
I tried reduction of the cathode resistor as far down as 470 ohms ( 34 mA) which made a very slight improvement. The LF end is still rolling off steeply below 100Hz
Unfortunately, the output power now seems drastically reduced ( to be expected ) and although this may be adequate into a speaker,as the quality will still be lacking I see no real gain in leaving it like this.
The only thing left as far as I can see is the valve itself, but on retesting, the emission is in excess of the quoted TAYLOR figure and the gas test is good.

My patience with this heap is now exhausted.
Even when working it has other issues and a lot of intermittent crackling. The audio quality ( aside of the bass deficiency) sounds distorted, even into a different speaker, so thats something else which needs sorting.
My patience threshold is Not helped by having other projects to do, zero time and this set taking up my workbench !!

A.