Output Transformer replacement query.

[Station X]

I need to replace a faulty output transformer in a military set. It is specified as having a ratio of 15:1 and drives a pair of 800 ohm impedance headphones from the secondary. 15 x 15 x 800 gives an anode load impedance of 180,000 ohms. This may seem high, but the transformer is connected in the anode feed of a 7R7 double diode pentode rather than to an AF Amp Valve.

I am wondering if an RS Components Output Transformer would be a suitable replacement. It will just fit in the space and can be strapped for 15:1 ratio. The anode current (3mA) is no problem, neither is the power handling (50 milliwatts). Unless the original had a very high DC resistance in the primary there should be very little volt drop across the primary, so there will be little effect on anode voltage.

The only thing concerning me is that the RS transformer was designed with primary impedances of 4-15 Kohms and secondary impedances of 3-15 ohms in mind. How will it perform with 180 Kohms and 800 ohms in the primary and secondary respectively? Will there be problems with the inductance of the windings? Is there anything else I need to watch out for?

I would try it and see, but I don't have an RS transformer to hand.

Graham.

[grindrod]

Hi Graham, the RS transformer is likely to be lower inductance. When terminated in 800ohms you may get a rising frequency response as the pentode anode load is a combination of reflected load and primary inductance. To find the LF corner frequency you obviously need to know the primary inductance.

You could flatten the frequency response at the expense of gain by shunting the secondary load. However, I think it is likely to be OK as it is since the headphones will probably have a lower corner frequency of several hundred Hertz and the RS output transformer was probably designed for an LF corner frequency of around 50Hz when correctly terminated.

Give it a go!!

73s Martin G8ILN

[mikej56]

Hi graham,

I have a few r.s. replacement o/p transformers here, but don't have any spec for them! They are PM1, PM3, PM4, PM5, PM6. If anyone has info on these and one is suitable for your purpose, You can have it for the postage cost.

michael

[Station X]

Mike.

Thanks for the offer of a transformer. I think I'll be better off getting a new one with a known spec. The available space is very restricted and I know that the RS Universal Output Transformer will just fit in if I modify the mounting lugs.

Martin.

I'll take your advice and suck it and see. Not being an audio buff I'm a bit puzzled by some of the theory though. As I see it taking the primary inductance (IN ISOLATION) the Inductive Reactance will increase with frequency following the rule XL = 2 x PI x F x L. However this will need to be combined vectorially with the relected impedance from the load. Is this where the rising frequency response comes from?

Unfortunately the original transformer is missing, but I know it was very small and mounted in a can. I should have mentioned that to counteract the inductance the original transformer had a tone correction capacitor? of 0.002MFD wired across the primary. I don't know whether this was designed to give a flatter response, or to optimise the response at around 1KHz. The set was designed for CW (morse) reception. Presumably I would need to change the value of this capacitor to fit in with the reduced inductance of the new transformer? It will need less capacitive reactance and therefore the capacitor will need to be increased in value?

Finally what is the LF corner response frequency? I assume that it's a point on a graph, but none of my text books mention it.

Graham.

[Station X]

To find the LF corner frequency you obviously need to know the primary inductance.

Can someone tell me in simple terms what "corner frequency" is? I've tried googling it, but the explanations I find are immensely complicated and in many cases have nothing whatever to do with transformers or audio .

Graham.

[GMB]

My understanding of the corner frequency is just the point on the frequency response where you reach the end of the bandwidth. When the performance of almost anything is plotted in the usual log/log way (i.e. log freq vs dB) at each end you usually see a point where the graph turns down and keeps on going. The log/log plot makes this look quite a sharp change, i.e. like a corner.

In my experience full-blown circuit analysis is always too hard because you often don't get the full and complete data even from a manufacturers spec. sheet.
So you have to get in the right ball park and then measure the actual result and tweak as needed using the insight that a simplistic analysis gives.

[grindrod]

Hi again everyone, the corner frequency is typically the -3dB point.

Graham, your analysis is correct, usually a transformer would have sufficient inductance for the load seen by the valve to be dominated by the reflected secondary impedance over the frequency range required.

In your case the lower primary inductance will give a higher low frequency
-3dB point than a transformer with more primary inductance, however, communication headphones are probably tailored with a response from 500Hz to 3kHz so it might not matter.

The shunt capacitor will provide gain reduction at high frequencies, the reflected load resistance makes sure that any tuned circuit formed with the primary inductance is extremely overdamped.

Cheers, Martin

[Station X]

Thanks Guys.

I am familiar with -3dB points and the classic inverted bathtub curve showing frequency response. I'd just never heard of "corner frequency", but its meaning is obvious once it's been explained.

As a matter of interest the receiver IF is specified as having a 3dB bandwidth of 1KHz and a 20db bandwidth of 9KHz. IF alignment to be carried out with a carrier modulated at 400Hz. I've tried to compare this with modern filters, but these are specified by 6dB bandwidth. I don't think that AM reception will be very good though.

Graham.

[Ghostuser]

Hi,
From the time when I was designing servo mechanisms, I discovered that 'Corner Frequency' was an Americanism for the turnover frequency (3dB point) of a simple filter. I would certainly just whack in the replacement output transformer and 'tune' the primary shunt capacitor to give a pleasing sound!
Nigel

[Station X]

I fitted the RS Transformer and as expected it worked. However the output to the headphones is far too high, even at minimum volume*. Worse than someone playing their Walkman flat out on public transport!

*I should explain that this receiver has no "volume" control as such. The pentode output stage operates with fixed bias. There is no AGC. The "volume" control is actually a combined RF and IF gain control which varies the bias on the preceding stages which use variable mu valves. The negative bias is derived by passing the HT current through a 500 ohm resistor in the HT line. The voltage across this resistor is spot on at 14 volts representing 28mA HT current. The manual does not quote any grid voltages, but they are all what I calculate them to be at min and max settings of the volume control. Anode and screen currents vary as the control is moved.

As I can find nothing wrong in the rest of the set I am beginning to suspect the new transformer. The original, which was missing, was in a can about the size and shape of a 35mm film canister and therefore much smaller than the RS transformer. I am beginning to think that maybe it had significant losses when compared to the RS transformer. I haven't yet tried changing the value of the tone correction capacitor. There is no hope whatever of finding an original transformer.

What does the team think?

Graham.

[YC-156]

Originally Posted by G4ILN

However the output to the headphones is far too high, even at minimum volume*. Worse than someone playing their Walkman flat out on public transport!

Could be due to differences in either the expected and actual impedances of the 'phones, or different power sensitivities for a given sound pressure level. Some modern headphones are amazingly sensitive.

Originally Posted by G4ILN

What does the team think?

1. Connect resistor between transformer secondary and headphone connector.
2. Experiment with value of said resistor until volume is satisfactory at receiver mid gain setting.
3. Enjoy receiver.
4. Goto 2 if substituting headphones for different type.

Hope this helps a bit.

Frank.

[Station X]

Could be due to differences in either the expected and actual impedances of the 'phones, or different power sensitivities for a given sound pressure level. Some modern headphones are amazingly sensitive.

I am using the correct Second World War headphones which have a resistance of 120 ohms and an impedance of 800 ohms at 800 Hz. I would have thought that using the incorrect headphones would reduce the volume due to mismatch. I accept what you say abut modern headphones being more sensitive.

1. Connect resistor between transformer secondary and headphone connector.
2. Experiment with value of said resistor until volume is satisfactory at receiver mid gain setting.
3. Enjoy receiver.
4. Goto 2 if substituting headphones for different type.

I could try this, or I could use another tap on the transformer deliberately introducing a mismatch. Another possibility is to use a "T" attenuator in in the output which would maintain a correct impedance match.

Devising a fix is not really a problem, but I would really like to know what the problem is. ie is it due to a more efficient transformer. I think I shall have to carry out some sensitivity measurements at full volume and see how the results compare with the spec, which says 1 to 3 microvolts RF input for 10 milliwatts output. Max output is quoted as 50 milliwatts.

Graham.

[JHGibson]

Just add various value resistors across the transformer primary until you get the volume you desire. This will have an added benefit of flattening the audio response somewhat.
You could make the resistor variable if you want a audio volume control.
John.

[YC-156]

Originally Posted by G4ILN

Devising a fix is not really a problem, but I would really like to know what the problem is. ie is it due to a more efficient transformer.

My thinking is that if the difference you observe is due to a difference in losses, then the simple fix I suggest will be as good as any other I can think of when it comes to properly match the impedances involved. Losses in a transformer can be thought of as a perfect component with a series or parallel (loss) resistance.

You already know that the transformation ratio is OK, so if your new transformer has too low an inductance, the result should be a lower volume, not a higher one.

If the original transformer had a very high internal, primary resistance, then the anode voltage might have been a bit lower originally, something which might also have contributed to the lower, original volume.

Final option: Do you know for what use the set was originally intended? If it was to be operated inside an armoured vehicle or in battle, the volume might actually be right on...

Not sure what else to suggest.

Frank.

[YC-156]

Originally Posted by G4ILN

I think I shall have to carry out some sensitivity measurements at full volume and see how the results compare with the spec, which says 1 to 3 microvolts RF input for 10 milliwatts output. Max output is quoted as 50 milliwatts.

For comparison: 10mW into a pair of modern, 300 ohm Sennheiser HD600 headphones would result in a Sound Pressure Level (SPL) of 108dB.

It is not all vintage headphones, which are inefficient, so it could be that the set was designed for delivering ear splitting volumes.

Frank.