Inexpensive toroidal transformer works as a modulation transformer!

[AB2EZ]

Hi

As an recap (I've been using the rigs described below for over 1 year... this is an old thread that has been resurrected):

I've been using this approach to modulate my Ranger (via the 9-pin connector of the back) for quite some time now. I use a pair of Hammond 10H, 300 mA, chokes, connected in series (for the Heising reactor); and a 25VA-rated Antek transformer with a pair of 115 volt primaries and a pair of 12 volt secondaries. The secondaries are in parallel, and the primaries are in series... so the step up ratio is 230: 12 = 19.2 . The transformer costs $9.95 + shipping. The Ranger has a modulation resistance of 5000 ohms (600 volts B+ / 120 mA of average plate current). With the 19.2 step up transformer, the audio amplifier sees a load impedance of 5000 / (19.2 x 19.2) ~ 14 ohms. I feed the transformer with the 8 ohm (nominal) output of a solid state audio amplifier. I use an audio amplifier capable of delivering 120 watts into 8 ohms... because solid state audio amplifiers are relatively inexpensive these days. [No need to use an amplifier that can only deliver the required modulating power, with no room to spare]. The 25VA Antek transformer has not had any problems handling the power.

On my high power (legal limit) plate modulated 2 x GS-35b transmitter, I use the following: An off-the-shelf Antek step-up transformer rated at ~ 1kVA, with a pair of 115 volt primaries and a pair of 800 volt secondaries. The primaries are in parallel, and the secondaries are in series... so the step up ratio is 115:1600. This wasn't quite enough of a setup ratio (and Antek didn't have anything with a higher step-up ratio available off-the-shelf at that time)... so I feed this with another Antek transformer, rated at around 1 kVA, with with a pair of 115 volt primaries and a pair of 62.5 volt secondaries. This gives me an additional 2:1 step up in front of the other transformer.. Even with two transformers in series: 62.5 volts => 115 volts => 1600 volts (a total step up ratio of 25.6:1), these transformers behave as if they are ideal transformers (the magnetizing inductance is so high that it does not affect the low frequency performance, and the winding capacitance is so low that it does not affect the high frequency performance) for frequencies between 30 Hz and 10 kHz. I believe that Antek now offers a single transformer that would give me the same step up ratio. I use a Peter Dahl 50H 300 mA Heising choke in this application. In this transmitter, the modulation resistance is 1700 volts B+ / 300 mA average plate current = 5667 ohms. With the 25.6:1 step up ratio, the audio amplifier sees a load impedance of 5667 ohms / (25.6 x 25.6) = 8.65 ohms. I feed the transformer an audio amplifier capable of delivering 600 watts into 8 ohms. Obviously, we are talking about very high voltages... so extreme caution is required in putting all of this together and using it. The output of the Antek step up transformer (driven by the audio amplifier) is more than 2000 volts peak. The voltage across the Heising capacitor is more than 3700 volts peak (1700 volts of B+ on one side, and less than -2000 volts peak on the other side). The 50H Heising choke extends the low frequency cutoff down to 30Hz, but one really doesn't need that much inductance. I.e., in the case of the Ranger, I am only using a 20H Heising reactor, even though the modulation resistance of the r.f. output stage is about the same as it is in my high power transmitter. It is, however, important that the Heising reactor can handle the very high voltage across it; and between it's winding and ground.

In both cases, the performance is excellent, in terms of bandwidth and linearity. I usually monitor the output of the audio chain, and the output of my r.f. sniffer, simultaneously, on a dual trace digital scope. I scale and shift the audio trace to sit right on top of the top portion (envelope) of the rf trace. They track each other perfectly.

If you hear me on the air, and if I am not using one of the above transmitters, I will be happy to switch over to one of them to let you hear it for yourself.

Best regards
Stu

[N4LTA]

You can purchase a toroid core from toroid corp of MD with no secondary for very low cost and wind your own secondary.  The largest is rated at 1400 Va and is about $95 . 

It is huge.The cores come with a 120 volt primary prewound and taped They have 5 or so sizes.  I have used the big one for a power transformer for a 700 watt SS RF amp and have a smaller one for an audio amp (modulator) power supply.

Pat
N4LTA

[WA1GFZ]

Stu,
I bet you could add more turns using say #22 teflon wire to increase the output voltage at the expense of a bit more c between turns.

[w4bfs]

BUMP for WB4BFS

Thanks John ... I enjoyed our qso on 40 mtr today ... your signal peaked at 40 over ... the TMC is very strapping here .... tnx fer all the dope about it

Hi Stu ... how are you doing? ... on the dual gs35 tx, have any winding insulation issues surfaced ?  ... are you using spark gaps or some form of neg peak limiter to prevent arcing ?    I am considering using some of these toroids on several tx projects

[w3jn]

Indeed, was FB!

What brought this up was a nice QSO today with KA1KAQ and WB4BFS, who mentioned the thread and said he couldn't find it.  I said I'd bump it to the top for him.

[AB2EZ]

John

Happy New Year!

No problems with arcing etc. on the 2 x GS-35B rig

I'm not using any spark gaps.

Note: The solid state amplifier acts as a voltage source... provided the current through it doesn't exceed the limits set by its protection circuitry. On negative peaks... when the modulated B+ goes through zero volts.... the audio amp is already sinking all of the current flowing in the Heising reactor. If the output of the audio amp goes further negative... it sees a high impedance (differential impedance) ... but it doesn't care. I decided that a negative peak limiter would serve no useful purpose in this configuration.

Best regards
Stu

[w4bfs]

ok Stu ... this is good stuff ... I just visited the Toroid co of Md website and their toroid kits look interesting ( thanks Pat ) ... might just be the ticket for a WELL insulated application as a modulator ... is the shipping tape (both clear and colored) we all use for shipping, mylar tape and would it be suitable for hv insulation ?

[WU2D]

Great thread and real possibilities for compact rigs and mobiles! I use a backwards 6.3 V 5A filament transformer driven by P-P Mosfets in my ARC-5 mobile. It was not so flat! I employed modified Hiesing and feedback to straighten the response. I am sure that this kind of approach would help here too.

Mike WU2D

[N4LTA]

They (Toroid of Md) supply 3/4" wide thin mylar tape with the cores - it has no glue. They warn about winding too many turns and high voltage - so to go higher in voltage than a couple hundred volts you need more insulation probably.

I have used Scotch HV tape - it also has no glue - used for taping high voltage electrical power connections -  on RF toroids to 2000 Watts in tank circuits. It should work well and is like fiberglass cloth. I also used teflon tubing and inserted the magnet wire into the teflon tube. You can buy it on the net for exactly that purpose.

Pat
N4LTA

[KM1H]

The toroids Ive seen look far from precision wound which brings up the question...

Can stranded Teflon insulated wire be used? It is readily available in scrap yards for next to nothing, has silver plated wire (thats just a chemical reaction reason) and sizes 20 to 28 seem very common often in full reels.

Trying to stuff a half mile of solid wire down a teflon tube doesnt appeal to me....Im lazy

Mylar tape comes in different widths and thickness so a single layer of sufficient voltage standoff with a proper overlap should be OK. Ive used this in regular transformers when I had access to a small local transformer company. Then the owner had the audacity to have a massive stroke and die.

Carl
KM1H

[WA1GFZ]

Carl stranded wire works fine as long as you have plenty of room to fit the Teflon insulation. EI core you quickly run out of room. It is also easier to bend around a core when you are using large size wire.

[w1vtp]

Therefore, I think it is helpful to go back to the theory, rather than relying on the design rules that were based on assumptions which might not be valid for our objectives (which are to have fun).

I was going back to the theory. It shows that using a larger cap produces overshoot. It has little or nothing to do with frequency response requirements.

Steve

What you said about overshoot was revealed during Paul, WB2SKC's PSpice modeling for my EICO 720 experiment.  We opted for a compromise of 2 MFD in conjunction with the Hammond 20 Hy choke -  using a 4 MFD revealed a big overshoot at the bottom end.

Al

[w4bfs]

This thread has been most thought provoking, thanks to all who have responded ...

The Antec 3500 volt between primary and secondary windings, no initiate any spark specification appears to be in response to UL / CSA voltage isolation specifications which led to insulated split bobbin EI lamination construction some years ago ... the question that is still most bothersome (to me, anyway) is the insulation between the secondary windings.

I am assuming that these dual winding secondaries are wound simultaneously in a bifilar fashion in a single pass (?) with some type of shuttle passing machine ... if this is so, then putting windings is series would tend to put high voltage stress on the wire insulation ... it seems to be a reliability issue to me  ... if the secondaries are single wound with attention paid to voltage gradient (with additional insulation ) then my 'objection' becomes a non-issue.

[AB2EZ]

On my high power transmitter, I am using an Antek transformer (as a step up / modulation transformer) that is rated at 120 or 240 volts AC in, and 800 or 1600 volts AC out.

In my application, with less than 1800 VDC for B+, I need less than 2250 volts (peak) output from the transformer to obtain 125% modulation on positive modulation peaks (i.e. B+ x 1.25).

1600 volts AC corresponds to 1600 x 1.414 volts peak, which corresponds to 2262 volts peak.

So... I'm below the rated output voltage of the transformer if I keep the modulation peaks to less than 125%

In my Ranger configuration, I'm using a transformer (backward-connected) rated at 120 or 240 volts AC in, and 12 or 24 volts AC out.

240 volts AC corresponds to 240 volts x 1.414 = 339 volts peak.

I'm running the Ranger with 500 volts B+ on the plate of the 6146 (when I run it in its full power mode). Therefore I need 625 volts from the Antec transformer to achieve 125% modulaton peaks.
In the case of the Ranger, I am exceeding the rating of the transformer by a significant factor. It is working fine, but... to be on the safe side (and also to obtain less saturation of the core at low frequencies)... one might choose a transformer with a 450 volt AC (or greater) rating on the output winding, while still having a sufficient step-up ratio (i.e., ~20:1).

Stu

[flintstone mop]

Velleman makes a kit amplifier using toroids for the output. Nice looking stuff
Fred

[w4bfs]

John

Happy New Year!

No problems with arcing etc. on the 2 x GS-35B rig

I'm not using any spark gaps.

Note: The solid state amplifier acts as a voltage source... provided the current through it doesn't exceed the limits set by its protection circuitry. On negative peaks... when the modulated B+ goes through zero volts.... the audio amp is already sinking all of the current flowing in the Heising reactor. If the output of the audio amp goes further negative... it sees a high impedance (differential impedance) ... but it doesn't care. I decided that a negative peak limiter would serve no useful purpose in this configuration.

Best regards
Stu

so it seems that the solid state modulator output stage acting thru the step up transformer then cap coupled to the modulation reactor can 'snub' some voltage peaks that it didn't generate .... this sounds like damping factor to me ... can be a good thing

been thinking about this, some .... Tom, K1JJ also found out in a related thread that his solid state power amp could generate some large spikes at the end of transmitting to receiving mode .... this tended to fire mod xfmr spark gaps .... I believe he had to relocate his solid state amp away from the rf pa to eliminate the problem ... sounds like a line power transient or rf bombard

could be like the bucolic aphorism 'closing the barn door after the horse got out'

consider the inductor carrying current (ex: dc relay coil) ... when the current is shut off, a large magnitude reverse polarity voltage 'spike' is generated which is snubbed with a reverse connected diode ....end of problem

perhaps a series reverse connected resistor/diode combination (ala reverse peak limiter) across the modulation reactor does several good things .... helping prevent splatter (with the keep alive circuitry) and negative spike snubbing when at end of transmission .... this is likely more important with tube modulators since they tend to have lower damping factors

the connection to this thread is my attempt to increase system reliability by reducing stress on the modulator system

[AB2EZ]

Beefus

Let's review the two "rainy day scenarios" that would cause arcing of a modulation transformer (or its associated spark gaps):

Scenario A: Overmodulation

The output of the modulation transformer drives the modulated B+ on the plate of the rf output tube to zero volts (relative to the cathode of the tube). The rf output tube cannot conduct in the reverse direction (it acts like diode with respect to current flowing from plate to cathode). If the modulator tries to drive the modulated B+ below zero, the load on the modulator will be higher (because the r.f. tube will not conduct in the reverse direction). The higher load on the modulator means that each additional mA of current produces a higher change in the voltage. Thus the modulator now has the ability to produce a higher voltage across the input and the output of the modulation transformer than it would if the rf tube were still conducting.

In this scenario, there is no sudden interruption of current (whether in Heising configuration or not). There is only an increase in the load impedance on the modulator when the plate current decreases to zero. This increased load on the modulator may result in excessive voltage being produced across the modulation transformer if the modulation being applied (in the negative direction) is large enough.

Scenario B: Loss of rf drive to the rf output tube.

If the rf drive to the rf output tube suddenly disappears when the transmitter is "on", then the output rf tube will not conduct current. Any current that was flowing from plate to cathode in the output rf tube, at that instant, will be interrupted. Because of the presence of large inductors (the Heising choke and/or the modulation transformers input and output windings)... this interrupted plate current will need to be diverted to somewhere. It cannot just stop, instantaneously. The interrupted plate current will be diverted into the modulation transformer... and, from there, into the output side of the modulator. If the modulator has a large output impedance (as in the case of a typical tetrode-based, class B modulator), the diverted current will produce a large voltage (much larger than the normal output voltage of the modulator)... which will cause a spark to occur across the modulation transformer's windings, or across a spark gap (if present), or across the modulator's output tube, or wherever the weakest point is for the diverted plate current to flow.


Now, getting back to the issues you are raising.

If the modulator has an output impedance that is intentionally matched to the modulation impedance of the rf stage... [which is the standard engineering practice in many traditional vacuum tube transmitters; because it maximizes the output power that can be delivered by the modulator, and therefore (for a given modulation power requirement) would minimize the cost of the modulator tubes, and the cost of the power supply for the modulator]... then scenario A will result in a doubling of the load on the modulator during excessive negative modulation peaks... and therefore twice the change in voltage across the modulator and the input and output of the modulation transformer during excessive negative peaks. If the components are marginal, in their ability to handle the voltages they need to handle... then damage to the modulation transformer might occur. [Note, in a tetrode-based modulator, without any feedback between the output and the input of the modulator, the peak output voltage swing is limited by the plate voltage; but the output impedance is very high] If we are using a design in which the output impedance of the modulator (after taking into account the turns ratio of the modulation transformer) is significantly lower than the modulation impedance of the rf stage, then the effect of the loss of the modulation resistance of the rf output stage, on excessive negative peaks, will be small. You still need to be concerned, independent of the diode effect produced by the r.f. stage, as to whether your modulator is putting out excessive current, and therefore, excessive voltage across the modulation transformer and across the modulated rf tube, when you modulate more than 100% in either direction. In my legal limit transmitter, the solid state audio amplifier (modulator) has a source impedance (after taking into account the step-up transformer) that is much lower than modulation resistance of the tube: less than 4 ohms x 25 x 25 modulator source impedance v. 1800 volts/300mA rf stage modulation resistance => less than 2500 ohms modulator source impedance v. 6000 ohms rf stage modulation resistance. On 100% positive peaks, the modulated B+ voltage across the tube is 3600 volts. If I were to push this above 4000 volts, I would be entering the territory where the tube is known to suffer catastrophic failures. This has, of course, nothing to do with negative peaks.

With respect to loss of rf drive to the rf output stage (scenario B), one has to look carefully at the output resistance of the modulator. As mentioned, in a tetrode based Class B modulator (e.g., a Johnson Ranger) the interrupted plate current of the rf output tube has no place else to go... unless it generates an arc. Increasing the plate voltage on the tetrode modulator tubes (caused by the inductors trying to maintain the interrupted current) has too small an effect on the plate current through that tube.  Therefore, scenario B implies a serious problem, that cannot be mitigated by a negative peak limiter. You need to ensure (or hope that) the drive signal does not turn off while there is plate current flowing. It is possible that the bias on the grid of the 6146 (in the Ranger) will increase (less negative) fast enough (if the rf drive is lost) to keep the plate current flowing.  

If the modulator has a reasonably low output resistance (1/ the change in modulator current that results from a change in modulator output voltage) then it can absorb the interrupted (rf stage) plate current that occurs in scenario B, without producing a large voltage change at its output.

[w4bfs]

great stuff, Stu ... if I am following along does not scenario B occur at each end of transmission if supply B+ is present unless some significant turn off sequencing is involved ? 

[AB2EZ]

Beefus

Scenario B only occurs if there is current already flowing in the Heising reactor  (a large inductor with lot's of energy stored in its magnetic field) or (alternatively) unbalanced DC flowing in the secondary of the modulation transformer (a large inductor with lot's of energy stored in its magnetic field)... which is suddenly given no place to go (except to flow into the output impedance of the modulator or to arc across the easiest arcing path that allows the current to continue to flow).

If the transmitter is off, there will be no significant current flowing in the Heising reactor or the secondary of the modulation transformer (unbalanced DC or otherwise). Therefore there is no sudden disruption of existing current associated with the turn-on transient process. On turn-on, the current in the Heising reactor or the secondary of the mod transformer will slowly build up.

In the case of the Heising design, when the rf tube first begins to conduct, at turn-on, it will draw current from the Heising capacitor (if the other end of the Heising capacitor can deliver any current), or at least attempt to do so. This will cause the voltage on the tube side of the Heising reactor to drop below the B+ value. This will create a voltage across the Heising reactor that will cause the current in the Heising reactor to start building up to its steady state value, in an orderly way.

Violent effects should only occur if you try to suddenly discharge the energy stored in the electric field of a capacitor (not relevant here); or if you try to discharge the energy stored in the magnetic field of an inductor (by turning off the current that is flowing in that inductor).

Best regards
Stu

[w4bfs]

hi Stu ... I didn't know you taught and from the discussion I assume its engineering level courses ....so should I call you doctor Stu ? .... I have been thinking about this discussion and perhaps we are saying the same thing with different words .... I fully agree that no snubbing is required with high damping factor modulators (most solid state) unless enough energy were reflected back to exceed ratings or abilities of the ssa .... after talking with Reed W2CQH , he suggests the use of diode / resistor snubbing would likely be useful with a vacuum tube modulator, especially pentodes , due to much lower damping factors and might be able to obviate the need for spark gaps to protect mod iron...

wish I had time for more but I am an income tax preparer and am busy now ....73 ....John

[w4bfs]

with apologies to those who may not be interested in re-surrecting this thread ONE MORE TIME ... finally got around to doing some hi pot testing on a small Antek toroid (dual 120 V to dual 12 V @ 25 VA ) ... the construction of this unit appears to be dual primary wound on first then layer of mylar insulation then dual secondary .... seems to be adequate insulation between primary and secondary and is spec'd so ..... the hi pot test (not spec'd by Antek) between the primary windings (one end to one end) showed leakage starting at about 1200 vdc ....this seems to indicate formvar( or enamel ) insulation only .... the construction method does not indicate a varnish dip of completed transformer .... this is probably ok for a 600 V class modulated 6146 and is the application suggested so far

I wonder how the higher voltage transformers are assembled ... are the paired windings wound together or one at a time with additional insulation ? ... assuming this is so is dangerous and could lead to premature meltdown

[AB2EZ]

John

Good input! I'm not surprised that the two, unconnected, 120VAC windings of this 25VA transformer are showing leakage current starting to flow between them... when tested with 1200 volts between one winding and the other. When the windings are placed in series, the transformer is rated at 240 VAC end-to-end... but the voltage between the two windings would be 120 VAC. This corresponds to 170 volts peak. It is, however, good to know, specifically what the limitations of this transformer are. The larger 800VA transformer that I am using with my legal limit, plate modulated transmitter is rated at 800 VAC across each of the two separate output windings (1600 VAC when the windings are placed in series). That corresponds to a difference between the two windings of 1/2 x 2263 volts peak. I would assume that a hi pot test of the same kind that you conducted with the small transformer would again confirm a significant margin between the peak voltage that would be present between the secondaries in series, and the voltage at which leakage current flowing between them would occur. Antek now offers a toroid rated at 2000 VAC (output windings in series), which corresponds to difference between the windings of 1/2 x 2828 volts peak.

Stu