What happens when Output transformer max voltage (slightly) exceeded?

[walangalam]

I'm planning a GM70 build, and am now looking at output transformers. The B+ voltage would be at 1000V (maybe 1050) from 800VAC.

I was thinking of getting a James JS6123, but I found a thread where the manufacturer says its not appropriate for 813 tube, even though its rated at 1000V max. In fact, they even said its only appropriate for 300B and similar tubes. Then I contacted Edcor and asked for a 1200V OT, but they said they cannot do > 1000V .

Of course there's ElectraPrint, Hammond 1628SEA, and the James 6135 which can do GM70 type tubes. And I will go for one of those for safety's sake.

But then there's Pete Millett's 813 , which uses a Tango XE20 which is also rated at 1000V and he uses it at least 800V, close to spec.

So for my enlightenment, what will happen if a 1000V OT is used at 1050V?

 

[Gilgy]

What happens to any component beyond its design parameters? It becomes unreliable and/or dangerous. In a transformer run right on the edge of its specification I wouldn't expect the winding insulation to last very long.

 

[walangalam]

I wouldn't expect the winding insulation to last very long.

Why? would it melt the insulation or break it down in some way?

 

[artosalo]

If a transformer is designed for 1kV operation voltage, then the safety margin is presumably 50...100%. I would say that 1050 V for a transformer specified for 1000 V is fully OK.

 

[walangalam]

So the max voltage rating is not "absolute maximum", such as in semiconductors or tubes, which will kill the device if exceeded?

 

[DF96]

Apply 51V to a 50V transistor and it won't necessarily kill the device. It might shorten its life. Same for a transformer. It will probably work fine. If it fails within a few days or months you have no warranty.

 

[FullRangeMan]

Why? would it melt the insulation or break it down in some way?

Overload is acceptable only for a few seconds, if it is on a steady situation a catrastrophe may occour.
It will do a short-circuit on output tubes and speakers voice coils and maybe a fire in the room.

transformer fire - YouTube.

 

[trobbins]

If you are going to go down to 5% differences then perhaps you should make an effort to estimate the minimum voltage across your tube from characteristic curves, and regulation of your power supply.

 

[JMFahey]

The B+ voltage would be at 1000V (maybe 1050) from 800VAC.

Sorry but no, it will be around 1130V, way over Max.

its rated at 1000V max.

Ok, max is MAX , not "operating voltage" where there may be say, a 10% or 20% safety margin built in.
In fact it's good practice to operate at least 10%, preferrably 20% below Max.

Pete Millett's 813 , which uses a Tango XE20 which is also rated at 1000V and he uses it at least 800V, close to spec.

"200V "below" max and 200V "above" are both mathematically "close to spec", but are not the same.

Example:
These guys are 4 yards "to the left" of the edge. Close.

Mathematically it would be the same if they were 4 yards "to the right", huh? Still close

 

[VaNarn]

The insulation of the transformer has to be capable of withstanding inductive spikes that may occur as a result of tube failure,clipping and mains fluctuations. L/C resonances can also be triggered which can cause excessive peak voltages.

 

[FoMoCo]

Apply 51V to a 50V transistor and it won't necessarily kill the device. It might shorten its life. Same for a transformer. It will probably work fine. If it fails within a few days or months you have no warranty.

True.

But, I'd feel much safer overvolting a transformer by 10% than a semiconductor. Some semiconductors will avalanche very close to the max rating. However, a transformer's insulation may be designed to withstand twice the rating for safety reasons.

Note that I'm not saying that it's okay to throw twice rated on a transformer.

 

[scholl]

IF the core is saturated the tube will see the DRC of the coil.

 

[walangalam]

Right, when the tubes fail, the OT will see 800*1.414.

What is DRC?

 

[Loren42]

You can contact the manufacture and ask them, but 5% is probably within the specifications.

I would contact the manufacture and see what they say.

 

[TheGimp]

If you are operating your tube at 1000V, you can expect to see almost 2X that on positive excursions. This should be figured into the transformer design, hence if a manufacturer states the transformer has a max operating voltage of 1000V they know it will see peaks near twice that.

So they are really designing for much more than the operating point voltage.

 

[walangalam]

I asked ElectraPrint about the max voltage and this is what he said:

"As for high voltage, the insulation used in these, is Hipot rated for 2KV
pulse test. If you unload, remove speaker, any of these high voltage types,
with music at high levels, they will pierce this insulation or short over
due to this condition generates more than 4KV. So ratings are safe for
normal use but not indestructible."

(! This is good to know. I regularly do this on my SS amp when I change components).

The Hammond 1628SEA also has the 2KV Hipot rating.

So maybe the "max voltage" is actually just the HIPOT pulse rating. What do you guys think?

 

[Chris Hornbeck]

Transformer insulation, especially after it's been machine wound to shape, doesn't have 10% tolerances. Large margins have to be built in before manufacture. If you're making this for yourself, take all appropriate safety precautions (keep your KiloJolts inside), use squelching diodes on the primary, maybe a Zobel, and don't worry so much.

All good fortune,
Chris

 

[Bone]

In a single ended amp with B+ at 1000v the transformer will experience twice that ie 2000v on signal peaks due to the inductive effect. The correct limit on a 1000v rated transformer is 500v.

 

[trobbins]

I asked ElectraPrint about the max voltage and this is what he said:

"As for high voltage, the insulation used in these, is Hipot rated for 2KV
pulse test. If you unload, remove speaker, any of these high voltage types,
with music at high levels, they will pierce this insulation or short over
due to this condition generates more than 4KV. So ratings are safe for
normal use but not indestructible."

(! This is good to know. I regularly do this on my SS amp when I change components).

The Hammond 1628SEA also has the 2KV Hipot rating.

So maybe the "max voltage" is actually just the HIPOT pulse rating. What do you guys think?

You may need to check if the Hipot test is rms or DC - it sounds like a 1 second application, rather than a 1 minute withstand test that is typical for power transformers. The insulation would likely be rated for at least 2kV, as insulation is not normally tested at its 'limit'. Sometimes the insulation is suitable for say 2.5kV hipot, but the creepage/clearance distances are designed for 1 or 2kV working levels (in which case moisture and pollution are then the likely breakdown paths if the windings aren't resin impregnated).

As suggested, there are ways to introduce over-voltage protection on the primary winding(s), and I would suggest that something to that end would need to be de rigour (let alone have B+ power supply over-current protection). I prefer to use a MOV across each half-primary for PP configuration.

 

[wa2ise]

A possible workaround: change the power supply in such a way to create a negative few hundred volts, and a few hundred positive volts. Run the output tube's cathode circuit off the negative supply, and the OPT off the positive few hundred volts. The tube will see the 1130 or so volts of supply, but the insulation in the OPT will only see around half that, well within the spec. To drive the output tube you'll need a higher than usual voltage rated coupling cap.