Plate Power Supply building questions.

[Ed/KB1HYS]

Since I have made a 2500V 300mA plate tranny I need to start building a good supply.  I'm following the Orr and Arrl handbook sections.

I was reading in a Handbook, ('75 i think) that if you had a full wave transformer/rectifier system, you could use lower voltage rated chokes in a HV supply if you placed them in the return side (on the center tap, between the tranny and ground).  Caps are still the same from B+ to ground. 

Anyone ever try this?? any thoughts about if its a good idea??

Just thinking that since any chokes I get will probable the WYO (wind your own) brand, a lower voltage on the windings would be a plus. I have a stack of good size trannies that are not any handy voltage, but good raw materials.

Also on this note, any suggestions about testing HV supplies? Specifically putting a reasonably sized load on them and measuring the output voltage at the anticipated loading.  Should I just build up a bank of resistors to get the desired load?? (that would be a lot)

One last question, on a center tap transformer, how much difference in voltage between sides is acceptable.  I'v got a small imbalance in my tranny, (probably less than 200volts at full input voltage) and am just wondering if there's a recomended acceptable amount.

[K9ACT]

I don't have answers to your questions but I just want to offer a heartfelt congratulations on your homebrew spirit.   Winding transformers is up there with making tubes as far as I am concerned.  Way to go!

I do think the choke in the neg side thing is simply a safety trick and/or insurance against faulty insulation.  I do not think it is a design criteria you want to count on.

js

[Ed/KB1HYS]

Thanks, when you have some time and no money, homebrewing is alot more appealing.  Old oddball trannies are cheap or free...

The chokes would be designed to handle the full rated voltage, I was just thinking that I could get an extra margin of safety by using that technique, just incase. 

I may just try it and through a probe across the output and see what I get for ripple.  Takes time to cobble up and experiment though.

[WA1GFZ]

I did negative lead chokes in my 4CX3000A rig and it worked fine. Why have 6 KV on a part when you can have 1 or 2 kv across them.

[K9ACT]

I may just try it and through a probe across the output and see what I get for ripple.  Takes time to cobble up and experiment though.

Measuring ripple on HV supplies is tricky because if you use a divide by 1000 probe, you need a lots-of-digits volt meter to see the ripple.

Don KYV, splained a neat trick using a cap that I am too much of a coward to try.  Best he explain it.

js

[AB2EZ]

Ed

My comment/question would be as follows

1. The peak voltage across the choke will be the same... only the voltage between the choke's windings and ground (i.e., between the choke's windings and the case/core of the choke) would be reduced. However, reducing the voltage between the choke's windings and ground would be a good thing if the choke's case is mounted directly on a grounded chassis.

2. With the conventional design, the center tap of the transformer will be at ground potential. However, with the alternative design (choke connected to the center tap of the transformer), the center tap of the transformer will be at approximately:  [1.414 rms secondary voltage -rectifier voltage drop - B⁺] = (approximately) [1.414 rms secondary voltage - rectifier voltage drop - 0.9 rms secondary voltage] = (approximately)  0.5 rms secondary voltage.

Thus, you have to make sure that having about having about 50% of the rms secondary voltage between the center tap of the transformer's secondary winding and the case/core of the transformer (assuming the case is grounded) is not going to be a problem for your particular transformer design. Likewise for the voltage between the center tap of the secondary winding and the primary winding.

Best regards... and I agree that winding your own transformer is very impressive...

Stu

[The Slab Bacon]

Ed,
    I have used negative lead filtering on HV power supplies for years with no apparent problems. There are some wierd situations where it doesnt work, but they are few and far between. I have had negative lead filtering with the choke in the grounded lead on my 4X1 transmitter for years now with no problems. the choke is rated at something like 2500v and the power supply runs 3450v at full load.

With "negative lead" filtering the only real potential that the chokes insulation has to worry about is the DC voltage drop across the chokes internal dc resistance + the ripple voltage.

the lower potential across the chokes insulation makes it a lot easier to live without old or home made chokes zorching over.

                                                The Slab Bacon

[Jim, W5JO]

Negative lead filtering is used in the Globe King 500 B & C.  Be sure you have the right amount of henries and microfarads for it to filter.  There is a chart in your ARRL handbook that will give you the proper amounts of each to result in the amount of ripple to expect.  As I recall the design target for ripple in AM transmitters is 5% or less.

[AB2EZ]

There is a free application you can download here

http://www.linear.com/designtools/software/switchercad.jsp

called SwCAD III... which was pointed out to me by WA3KLR (Tom).

It is easy to learn to use, and it works great!

I use it all the time now to model power supplies... to see what the DC output voltage, residual ripple, peak currents through the diodes/rectifiers, etc. are.

As a rule of thumb, I try to get the peak-to-peak residual ripple at the output of the supply to be less than 2% of the dc output voltage. This implies that the hum on the output of (for example) a plate modulated transmitter will be more than 40dB down from the carrier.

For an example of how this application works... check out the bottom of this page of my web site, where I posted the simulation results for the 3000 volt power supply I built for my linear amplifier.

http://mysite.verizon.net/sdp2/id20.html

Stu

[k4kyv]

With "negative lead" filtering the only real potential that the chokes insulation has to worry about is the DC voltage drop across the chokes internal dc resistance + the ripple voltage.

But the ripple voltage can still be considerable, especially at the first or only choke in the filter.  Remember, the raw rectified a.c. coming out of the rectifier has ripple content in excess of the smoothed dc voltage from the filter, and this can (and has many times) been enough to break down the insulation of a choke.  I had it happen to me once, trying to use a choke far beyond its rated DC voltage.  So the choke  still needs to have enough insulation to safely carry the dc power supply voltage and more.  Also,  make sure the "start" end of the winding is the end that connects to ground.

I prefer a more reliable method, which is to keep the choke on the HV side, but if I am dealing with more than a fraction of the rated voltage of the iron, I always mount my filter chokes, modulation transformers and modulation reactors on insulation to isolate the case from ground.  No point in putting unnecessary stress on the internal insulation, which is usually nothing but paper - especially in older transformers, by grounding the core while there is high voltage on the winding.

But with that set-up, one must make it a point never to touch the case of the choke or transformer while the HV is on.  If there were a  short from winding to core, the case would be at full HV potential.  Also, even with a few megohms of  leakage, there may be enough high voltage on the case to give you a jolt.  If there is  no  leakage, the winding and core can hold a charge like a capacitor and you can still get a nasty jolt. It might not be a bad idea to post a warning notice in a conspicuous place just in case someone unfamiliar with the transmitter might some day be working on it while the HV is on.

By observing all the above precautions, you can safely and greatly extend the life expectancy of irreplaceable transformers and reactors.

[Ed/KB1HYS]

Good advise, thanks!

I had planned on encasing the finished 'formers and chokes in nonconductive cases, to isolate the cores from possible human contact. 

Don,

what was the ripple measuring technique above??  place a suitable HV cap on the end of your probe and AC couple the ripple through that cap to the scope? 

[AB2EZ]

Here are some simulations of a (sample) HV supply... using SwCAD III

1. The schematic of a choke input supply using a 1414 VAC center tapped transformer, a 10 Henry choke, and a 50 uF filter capacitor. The load is set at 5000 ohms. The series resistance of each half of the secondary plus the associated rectifier/diode is assumed to be 50 ohms. Also shown is a 100:1 capacitive voltage divider for observing the residual AC ripple at the output of the supply.

2. A simulation that shows the residual ripple at the output (multiply what is shown on the plot by 100 to get the actual ripple... to take into account the 100:1 capacitive voltage divider). The peak-to-peak ripple at the output of the supply... as calculated... is 30 mV x 100 = 3 volts p-p

3. A simulation that shows the output voltage and the voltage on the input side of the 10H choke (the difference of these is the voltage across the choke). As expected, the DC output of the supply is approximately 0.9 x the rms voltage between either side of the transformer and the center tap = 0.9 x 707 volts = 636 volts. The peak voltage across the 10H choke in one direction is 1000 volts - 636 volts = 364 volts. Yhe peak voltage across the 10H choke in the other direction is 636 volts - 0 volts = 636 volts. The peak voltage between the 10H choke and ground is 1000 volts.

Stu

[Steve - WB3HUZ]

And other parts of the country too.

http://mysite.verizon.net/sdp2/id20.html

Good choice on the plate iron Stu. Tom Hand/W4WDS transformer/choke products can be found in AM rigs all over the South & Southwest.

Mack

[AB2EZ]

Here are the schematic and simulations for the alternative configuration

1. Schematic
2. Simulation of the voltage at output of the supply, and on the left side of the choke (the right side of the choke is at ground potential).

As discussed, the peak voltage across the choke (left side to ground) is the same as in the conventional configuration: 636 volts

The peak voltage from choke to ground is reduced from 1000 volts (in the conventional configuration) to 636 volts (in the alternative configuration).

The voltage on the center tap of the transformer rises to as much as 636 volts above ground in the alternative configuration (v. tied to ground in the conventional configuration)

Stu

[AB2EZ]

Ed

To illustrate the effect of imbalance in the secondary with respect to the "center tap"... I simulated the power supply I have been using in these posts with 950 volts peak (not rms) on one side of center tap and 1050 volts peak on the other side of center tap. Thus, I simulated the effect of a 10% imbalance.

The results are shown below:

Slide2.JPG (repeated from an earlier post in this thread) shows the residual ripple at the output of the supply when it has a balanced transformer (1000 volts peak on each side of center tap). Don't forget to multiply the ripple voltage shown on trace by 100 to take into account the 100:1 capacitive voltage divider I am using to measure the ripple in my simulated supply.

HV Supply Simulation 3.JPG shows the residual ripple with the unbalanced transformer (950 volts peak on one side of CT and 1050 volts peak on the other side). Again, don't forget to multiply the ripple voltage shown on the trace by 100.

As you can see, the peak amplitude of the ripple isn't much bigger...  but it does contain some 60 Hz component when the transformer is not balanced. [With a balanced transformer, the ripple contains only 120 Hz and small amounts of harmonics of 120 Hz].

In the unbalanced case, the output voltage is essentially the same as in the balanced case.

[k4kyv]

Don,

what was the ripple measuring technique above??  place a suitable HV cap on the end of your probe and AC couple the ripple through that cap to the scope? 

I'm not sure, but I think we discussed putting a HV blocking cap in  series with the a.c. voltmeter to read the ripple voltage directly.  You have to make sure the blocking cap is fully charged to the HV potential before inserting the a.c. meter; otherwise the discharged cap would act just like a dead short when you inserted the ac meter  - the equivalent to measuring HV dc with a low voltage a.c. voltmeter, not too good for the movement.  It would probably be safer to use a scope.  But I have done it many times and never blown up a meter.

OTOH I tend to blow up meters by forgetting to change settings on a multi-meter.  I once blew up a nice Fluke DVM when I tried to measure 1000 volts DC with it, forgetting to move the probe from the milliamp to the +HV receptacle.  It literally exploded like a firecracker.  I have never had a VOM to last me for more than 4-5 years, no matter how carefully I try to avoid making a mistake.  Inevitably, sooner or later I'll try to measure something like amperes with it in the microamp position, or 110 volts a.c. using the milliamp scale.