Re: Let's talk about tube amps

Eric wrote:

The amp is a mesa boogie tipple rectifier and some fool
had added a resistor to the bias supply increasing the
cathode current. He must have had some interesting
sounds as the power transformer primary shorted and B+
went up and up,

In what way does "up and up" differ from "up"? What do
you mean by shorted? I can understand that you want to
join in with all the fun we have here, so I don't wish to
dampen your ardour, but you need to think and write more
clearly. Try proper paragraphs.

It might be useful, for the purpose of grasping how
transformers work, to consider what happens if just some
of the primary windings become shorted. Would that lead
to an increase in secondary voltage? There is an obvious
theoretical reason to think that it might, and a rather
more complicated reason, to realise that it most probably
won't. The best assumption is that the manufacturer has
used as few turns as possible.

Your shorted primary hypothesis is not best tested by
plugging it into the mains.

But it made me think of a nice example of a "reductio ad
absurdam". If you continue to short primary turns, then
secondary voltage should keep rising because the turns
ratio increases. In that case, when you have no turns,
then you should have a very high secondary voltage
indeed.

it still does this but no output tubes were in place as
I tested it.

Still does what, exactly? HT voltage will normally be
higher with valves removed, but by how much depends on
the details of the circuit, which few here will be
familiar with.

lots of nasty stuff under the power trans that is some
kind of potting compound. So it needs that replaced but
I am not sure if the output trans is good or not. The
primary reads 21 and 23 ohms from the center tap.

Likely to be normal. The difference is because the two
halves will have the same number of turns, but can't
occupy the same space, so one will have a different
average coil diameter, and so a different length of wire.

This is a 160 watt amp with 6-6l6GC's. The output trans
is good and rusty so I am sure it got overheated as
well.
Also there was lots of glass chips from the output tubes
under the chassis, could the tubes have exploded? Any
thoughts?

Explosions are scarce, I should imagine, although you
could have fun thinking up a possible scenario. I wonder
how much of the contents would have to vaporise before
sufficient positive pressure resulted?

Implosions are more likely, because of the vacuum.

Ian
I meant that when first powered up the voltage was normal
for a low load condition and over a period of a few
minutes the voltage increased about 12 percent. The
transformer also buzzed more as it warmed and this was the
case with a Marshal I repaired a few years ago that needed
a power transformer.

I checked the line voltage and it was constant at 117
volts. My area has an old power system. So I was thinking
only a few windings had shorted before I discontinued the
test. The leaking potting material also lead me to think
the transformer was likely to be defective.

"The drinking man's favourite model." Yes this is the
most probable explanation for the glass fragments. Perhaps
a collision between power tubes and beer bottle lead to
the amps present condition.

Is it not likely that a test not using line voltage would
not heat the transformer and therefor not show a short?

I did fail to provide enough information in my first post
but would be pleased to let the blame fall on the US
educational system for the sentence structure.

Shorted primary isn't the same thing in most ppls heads as
shorted primary turns.

At bottom-left here:

http://www.ivesonaudio.pwp.blueyonder.co.uk/TransformerIllustration.GIF

is the usual and *very* useful low-frequency model of a
transformer. At high frequencies it becomes necessary to
include winding capacitance and leakage inductance, but for
your purpose the simple example shown is OK.

From left to right, there's the primary winding resistance
in series with the primary inductance. In parallel with the
primary inductance is the load, made up of the secondary
winding resistance in series with the resistance presented
by your amp, both of which should be multiplied by the
square of the turns ratio, Tprimary/Tsecondary.

Notice that the primary winding resistance carries the
current attributable to the load, *and* that which flows
through the primary inductance. If you short turns on the
primary, you reduce the turns ratio *and* the primary
inductance. Both changes result in extra primary current.
Considering also that the inductance falls by the square of
the proportion of turns lost, the increase in current is
much greater than the reduction in primary winding
resistance, so the power it dissipates rises steeply, so it
gets hotter. If it was already hot enough to cause a fault,
then hotter still is likely to lead to catastrophic failure.

If you must test using the mains, then you should at least
work out beforehand what you need to measure during the
test, and what you expect the measurements to be if your
theory is correct. You should also eliminate as much
extraneous circuitry as possible...easily done in your case
by disconnecting the transformer from the amp and using a
suitable resistor instead. Then you can do your test quickly
and effectively and just once, thus minimising the risk of
causing further damage.

If you measure the AC primary and secondary voltages with no
load, and divide the one by the other, you get the turns
ratio, hopefully in the no-fault condition.

If you measure primary current and secondary voltage under
load, then wait for the fault to occur and measure them
again, you will have all the data you need. A little algebra
will then tell you whether your theory fits the facts.
Remember that inductance is proportional to the square of
the number of turns. Once you've calculated the change in
turns ratio, you can infer the proportional reduction in
primary inductance. Then you can work out the expected
increase in current due to load and inductance, and compare
that to your measurement. If you are right you will be
seeing something that Phil has never seen before, so it's
probably very rare and an experience to be savoured.
Certainly it's not a condition that's likely to last for
very long.

Theory is likely to be of academic interest only, as it
sounds like you need a replacement transformer anyway.

Ian


.

Relevant Pages

  • Re: Modeling high current power transformer in LTSpice
    ... The transformer I want to model consists of a 240 VAC primary toroid, with a single turn of bus bar through the hole. ... I added 10 mOhms primary resistance and 2 uOhms secondary resistance, and I also added a non-coupled inductance of 200 nH to the output. ... I can measure the primary resistance by using a DC current source, and I can measure the inductance by using AC voltage below saturation. ...
    (sci.electronics.design)
  • Re: Transformer drawing 1 amp
    ... When there is no secondary load, the inductance of th ... precipitously, twice a cycle, if the core flux reaches ... This occurs if you apply excessive voltage to ... The transformer limits current to the secondary by the ...
    (sci.electronics.basics)
  • Very handy Gadget-AC Continuity Tester
    ... "Ultimate" Continuity Tester Hides Many Tricks Up Its Sleeve ... has a resistance generally below 1 . ... The secondary of a 50/60-Hz transformer will generally have a dc ... for a specific charge connector. ...
    (rec.audio.opinion)
  • Re: Transformer drawing 1 amp
    ... That inductance drops precipitously, twice a cycle, if the core flux reaches saturation. ... The transformer limits current to the secondary by the resistance of both primary and secondary windings (that waste some of the available voltage, so less is there to drive current through a given load resistance). ...
    (sci.electronics.basics)
  • Re: Transformer Question
    ... > interested in understanding "magnetizing inductance". ... > an ideal impedance meter and measure the primary impedance with the ... if I were to model this transformer and refer all of the inductance ...
    (sci.electronics.design)
Loading