Transformer and choke protection circuit

[AB2EZ]

I have a question about the viability of a particular approach for protecting a modulation transformer, in modified-Heising configuration, from voltage surges caused by the sudden interruption of average plate current. This sudden interruption in average plate current would occur when the transmitter is switched from operate to standby.

As shown in the attached .jpeg, an additional low voltage winding of the modulation transformer (could also be the the primary, in some situations) is used to allow the stored magnetic field in the Heising choke to "discharge" in a controlled fashion... by diverting that current into the modulation transformer, and using an appropriately selected transient protection device in series with a resistor. The use of a low voltage winding on the modulation transformer trades current through the transient protection device for a lower breakdown voltage. The energy that is stored in the magnetic field of the Heising choke is dissipated in the resistor and the transient protection device, in a controlled fashion.

Example: consider a modulation transformer used in a transmitter with a 1500 volt B+. Suppose we want to keep the voltage between the main secondary and the primary below 3000 volts peak. Since the main secondary is grounded on one side, we want to keep the voltage across the main secondary below 3000 volts. If the plate current is suddenly interrupted, the current in the Heising choke will be diverted to the secondary of the modulation transformer (until the energy stored in the Heising choke's magnetic field is dissipated). The sudden, fast ramp-up of current in the modulation transformer's main secondary will cause a voltage spike to rapidly ramp up across all of the windings of the modulation transformer (secondaries and primaries). When the transient suppressor starts to conduct, current will flow through the extra secondary... which will allow the diverted current from the choke to flow without further increasing the voltage across the main secondary.

Comments?

Stu

[WA1GFZ]

Stu,
I have measured transient suppressor voltages as current transients up.
The breakdown voltage takes off with current so the TVS would need to be sized to handle the energy. I would prefer an orderly shutdown with tubes biased off before the antenna relay changes state. Killing the high voltage ahead of the TR relay will also work. Your idea sounds cool for transients while operating though.

[AB2EZ]

Frank

I think I understand what you are saying... but just to clarify:

Disconnecting the AC to the B+ supply (as is done in the KW-1), and allowing the final(s) to draw down the voltage across the B+ supply's filter capacitor will work... because the unbalanced current through the secondary of the mod transformer (in non-Heising configuration) or the choke (in Heising configuration) will decrease relatively slowly (100's of milliseconds), as the B+ decreases.

However, simply removing the B+ with a relay will not work (big spark across the relay, followed by sparks across the Heising inductor and/or modulation transformer) unless you include a capacitor (in series with a resistor) across the relay (or from the opposite side of the relay to ground)... that is large enough to sufficiently slow down the rate of change of the unbalanced current through the transformer or the Heising choke. For example, one might place a 25uF capacitor in series with a 3000 ohm resistor from the tube side of the relay to ground.

See attached .jpeg

[Steve - WB3HUZ]

Wouldn't the cold side of the secondary need to be grounded?

[WA1GFZ]

There is a lot of stored energy so agree a snubber may help. A relay in the HV lead sounds hard to do because the choke voltage is going up very fast when the current stops. I had the same problem in the viking 2 PDM rig and ended up letting the power supply go down while killing the digital drive to the switch tubes. (I have a damper diode so easier in my case)
The more I think about this the harder it is.

[AB2EZ]

As an addition to my first post in this thread...

If you assume that the capacitance to ground (not shown in my original .jpg) at the bottom of the RFC (plate choke) is .001uF; and if you assume that the average plate current, just before the tube is turned off, is 300 mA...

Then, when the tube is turned off,  the voltage at that point (where the bottom of the plate choke, the Heising choke, etc. come together) will ramp up at:

.3 mA x 1/(.000000001 F) = 300 million volts per second =  300 volts per microsecond.

.... unless the current has somewhere else to go (see further discussion below)

Thus, in this example, the protection device needs to turn on in less than 10 microseconds if the maximum allowed voltage on the main secondary of the transformer is 3000 volts, unless there is an alternative path for the diverted plate current (see further discussion below).

If you consider a rig like a Ranger (stock configuration, except for removing a few modulated B+ bypass capacitors in order to open up the high end), where the transition to standby is done by grounding the screen of the 6146 (and, also, un-keying the oscillator)... and where the average plate current is around 125 mA...

Then the voltage across the modulation transformer's secondary ramps up at around 125 volts per microsecond (assuming .001 uF of remaining capacitance, left between the modulated B+ line and ground).

In the stock Ranger, there can be a corresponding change in the current flowing through the modulator tubes (i.e., on the primary side of the modulation transformer) that helps to limit the ramp-up of voltage across the modulation transformer secondary and primary when the Ranger is switched from transmit to standby. [In a Ranger, the plate resistance of the modulator tubes, with a fixed screen voltage, is very high... so the beneficial impact of this is not as great as it would be with triode modulator tubes]  In that case, most of the energy stored in the magnetic field (about 1 Joule) will be dissipated by the modulator tubes. But, if you add a modification to turn off the modulator tubes when you switch to standby (i.e. grounding their screens)... then, depending upon the sequencing... the ramp up in voltage across the modulation transformer secondary and primary could reach much larger values than in the stock configuration.

See the attached .jpg files

When I externally modulate my Ranger in modified Heising configuration, using a step-up transformer driven by an 4 ohm output audio amplifier.... I now believe that most of the energy stored in the Heising choke (again, about 1 Joule) ends up being dissipated in the output stage of the 4-ohm audio amplifier. Fortunately, my audio amplifier has fast protection circuitry at its output!


Stu

[flintstone mop]

Would a better solution be to regulate the HV a little better? It seems that there are excessive load-no-load excursions?
I hope I picked up on the problem from your first post, that this goes back to the HV supply.

Fred

[AB2EZ]

Fred

The issue I am concerned about is how to safely transition from transmit to standby. HV regulation is not the issue here.

Stu

[WA1GFZ]

Yup,
I agree the power is absorbed by the modulator. This morning I started to write a suggestion to bias the modulator off first then it hit me that would be even worse because there was no place for the energy to go. You might be lucky in the Ranger if the transformer saturates in the reverse direction reflecting less power/voltage back to the SS amp. risky business though.
Maybe the thing to do is put a power suppply cap between the relay and choke so the choke dumps into the cap when the relay opens. It would also need a bleeder across it. Voltage rating nice and high. Still 1 J is a good belt of energy.
 

[flintstone mop]

Hi Stu
It seems that numerous boatanchors go through the key-up un-key process many times during its life time and no serious issues with mod transformers being destroyed. I might not be up to speed, is there a failure issue from PTT?
The transitions would be very bad in CW operation. That's why the secondary is shorted, or I should say, it is not in series with the final RF plate.
My .01 cents worth.
Fred

[AB2EZ]

Fred

The issue is always there when you have the plate current passing through a large inductance (i.e. the magnetizing inductance of a modulation transformer or a Heising reactor)*.

The approaches taken to deal with the issue vary from:

1. Ignore the issue, and hope that something about the design will take care of it (often results in blown modulation transformers)

2. Provide an alternate path for the current to flow... e.g. a diode protection circuit that turns on when the voltage across the secondary of the modulation transformer and/or the Heising reactor exceeds a safe value. The energy stored in the inductance is dissipated by the protection circuity.

3. Design the circuitry on the primary side of the transformer, so that an offsetting current will flow through the primary of the transformer to support the magnetic field (in the transformer or the Heising inductor) until its associated energy safely dissipates (e.g., the modulator on the primary side of the transformer absorbs the stored magnetic energy in a safe manner)

4. Turn down the B+ (and the associated average plate current) slowly, before turning off the RF tubes.

Yes, bypassing/shorting the modulation transformer secondary or the Heising reactor is an essential thing to do when operating CW... so as to not create the magnetic field in the large inductance.


*The inductance was usually selected as a compromise between cost and performance (as is always the case in a commercial product)... with the reactance targeted to be equal to the modulation resistance of the rf deck (B+/average plate current) at around 50 Hz.

For a Ranger, this comes out to around 16 Henrys, but the actual magnetizing inductance of a stock Ranger modulation transformer is only about 8 Henrys. So, the EF Johnson folks used 100Hz as the target for the low end 3dB down point.

If you try to abruptly interrupt the current through the inductor, the collapsing magnetic field will produce a large voltage across the inductor (v= L di/dt).

 

[k4kyv]

I think a better solution would be to kill the HV by shutting off the primary voltage to the plate transformer.  The filter caps in the power supply and the bleeder resistors will remove the high voltage gradually enough to avoid the kind of transient you are worried about.

Also, it would be a good idea to place spark gaps across the primary of the modulation transformer and across the mod reactor, and adjust them a little beyond the point where they arc over under normal modulation peaks.  In addition, especially with older transformers and reactors in which the insulation may have deteriorated due to decades of ageing, I always mount mod reactors, mod transformers and power supply filter chokes on insulation to isolate the core from ground.  Grounding the core doesn't serve any useful purpose and places needless stress on the insulation between the winding and the core.

One safety precaution: Treat the transformer or reactor as if the whole thing were at full H.V. potential, and never touch the case or the laminations while the HV is on, in case the insulation ever did happen to fail.  Even with the insulation fully intact, the capacitance between the core and winding may cause enough of a static charge to build-up to give a nasty shock if you touch the transformer while the voltage is on.

[flintstone mop]

I externally modulated a T368 with a pair of 810's running about 2500 volts on the plates and to reduce the amount of control cricuitry I just biased the 810's to cutoff with the PTT circuit. The B+ was always on the 810's. There was never any failures with these high voltages. Nicky Badwires has that same transformer in another transmitter with no problems.
I can't see what problem would arise with a B+ of 600VDC in a Ranger or Dixie 100.
I know you are preserving something that is getting harder to replace, but there is always a work-around. I still love my power transformer mod tranny in the AF-67. It is outboarded and I added a Heising choke, and the audio looks beautiful down to 40 hz. Nice smooth sine wave on the carrier.

fred