Faulty power transformer winding "bodge"

[Colin Boggis]

I'm faced with the unenviable option of trying to repair a 90 year old mains transformer on which one half of the HT secondary is O/C, I have come up with a possible solution. Not sure if it's been done before so I don't claim originality for this idea.
I plan to disconnect the centre tap from ground and reconnect it to the valve recifier anode that's currently connected to the faulty winding. Then add 2 1N4007 from the valve anodes to ground, thus forming a bridge.
See attached circuit.
I can't see any problem with this - any comments appreciated!

[ms660]

Transformer current.

Lawrence.

[Electronpusher0]

Agreed, normally the secondary current is shared between the two halves of the secondary, with the proposed scheme all the current for the HT is supplied by one half.

Peter

[theredhouseinn]

Hi Colin.
Yes, it will work, but you will only have half the current available. i.e. if the full winding is rated at 100mA, then you will only be able to draw a max. 50mA.
John.

The failed winding won't be generating heat so I think this is a good idea. If you do all the rectification with silicon then the heater power required by the valve rectifier will also not heat the transformer.

[snowman_al]

Err yes but, the remaining half is now working 100% of the time where it used to 'work' 50% of the time. So the heat will be the same if not a little hotter... when you over current it.

Only a wee bit more.

[Herald1360]

If it's the PSU for a basic receiver, you could bias back the output valve to around half its normal anode current. Output level will drop, but for most purposes not enough to be noticeable.

[vinrads]

I think the answer is suck it and see ,what have got to loose , Mick.

[kalee20]

The secondary-half that you use, will be generating twice as much heat as normal.

The O/C secondary, on the other hand, will be generating zilch. Overall, it's the same amount of heat, but concentrated in a winding portion rather then being spread out - so you'll have more of a hot spot.

It'll almost certainly work! Personally, if the transformer is 90 years old, I'd like to get a thermal fuse and strap it to the windings - ancient enamel has a tendency to fail with inter-turn shorts, rapid overheating, smoke and eventually, flames - rather than failing open.

[Colin Boggis]

I'm not sure about the current argument. With a centre tap, surely each half supplies an equal current as each half conducts alternately. What will change is the duty cycle - the "good" half now being the sole source on every cycle. So yes it may warm up a bit but I think I'll still have same current available. As Vinrads suggests, suck it & see probably the best bet. I wanted to keep the valve in use to ensure HT comes up slowly rather than being instant - if all diodes are solid state HT will rise to an abnormally high level. Lowering the output valve anode current seems a good idea too if the transformer overheats. I will set it up on the bench with appropriate dummy loads and test the idea.
Updates to follow.

[snowman_al]

Another thing to consider,
The bridge circuit draws more AC current from the primary to produce the same DC current (Idc = 0.62 Iac)
Whereas in the 2 diode bridge Idc = Iac.

[kalee20]

I can't see that - the primary current will be the same.

Each half-cycle, primary current flows to balance the amp-turns in the secondaries. The primary doesn't 'know' which secondary it is - if it's the same one each alternate half-cycle, or a different one, the primary will absorb current the same.

[snowman_al]

Here you go see the Hammond Transformer pdf.
I seem to remember it is the current needed to keep the smoothing capacitor charged?

[Herald1360]

Originally Posted by snowman_al

Another thing to consider,
The bridge circuit draws more AC current from the primary to produce the same DC current (Idc = 0.62 Iac)
Whereas in the 2 diode bridge Idc = Iac.

Hmm...., isn't the overall transformer utilisation factor for a bridge better than for two diodes fullwave? Two more thermionic diodes and their holders cost more than a slightly bigger transformer though! Not so with solid state.

[valves4amps]

If you're going to suck it and see then I would use a full silicon bridge but feed the + into the 2 anodes of the rectifier strapped together. That way you can retain the same slow warm up time and same volt drop the valve would drop.

Dave

[Colin Boggis]

Thanks for all the comments and advice - I particularly like the idea of full silicon bridge fed through the valve as this will accomplish my desire for slow warm up and the bridge is balanced better. Regards the Idc being 0.62 Iac. That doesn't take into account the field coil (choke). According to the Hammond chart, with this configuration Idc = 0.94 Iac which is still worse than the full wave circuit with choke where Idc = 1.54ac but not so bad.
Anyway as the transformer is "shot" the suck it & see method is worth a try. Also I found out that the output transformer primary & one IF secondary is also o/c - great fun this restoration!

[ms660]

Originally Posted by Colin Boggis

Thanks for all the comments and advice - I particularly like the idea of full silicon bridge fed through the valve as this will accomplish my desire for slow warm up and the bridge is balanced better. Regards the Idc being 0.62 Iac. That doesn't take into account the field coil (choke). According to the Hammond chart, with this configuration Idc = 0.94 Iac which is still worse than the full wave circuit with choke where Idc = 1.54ac but not so bad.
Anyway as the transformer is "shot" the suck it & see method is worth a try. Also I found out that the output transformer primary & one IF secondary is also o/c - great fun this restoration!

What's the filter input, capacitor (reservoir capacitor connected between the cathode of the rectifier and HT-ve) or choke (no reservoir capacitor connected between the cathode of the rectifier and HT-ve) ?

Lawrence.

[Colin Boggis]

I got that wrong didn't I!. Of course there's a reservoir cap so the previous comments regards Idc = 0.62 Iac is correct. Or does the fact that the choke and smoothing follows make some difference?

[Herald1360]

No, if there's a cap reservoir up front what follows doesn't affect what the rectifier "sees", only the degree of ripple filtering.