Collins 30k-5 plate transformer question

Discussion in 'Amplitude Modulation' started by KN4CTD, Apr 21, 2019.

ad: L-HROutlet
ad: l-rl
ad: FBNews-1
ad: L-MFJ
ad: Left-3
ad: Left-2
ad: Subscribe
  1. KN4CTD

    KN4CTD Ham Member QRZ Page

    Hi all. I picked up a plate transformer at a hamfest today. It is a Chicago Transformer P2520. A google search said this was the plate transformer from a Collins 30k-5 commercial AM transmitter. It’s only rated @ 1kva and shows dual 115v primary taps and secondary taps @ 2000 and 2500v. I’m hoping someone in this AM forum may be familiar with this transmitter and may be able to give me more insight regarding the possibility of using this plate transformer in an hv power supply for tube amplifier project. My main questions are, why is this thing so big and heavy yet only rated @ 300-400ma ccs on the secondary? And is the gonna be suitable for a 1.5kw pep SSB amplifier? Were these old transformers underrated for their size and really good for more current handling capabilities or is that just how they were built back then? Thanks for any and all feedback/help!
    KD2ACO likes this.
  2. K5UJ

    K5UJ Ham Member QRZ Page

    The short answer is: Yes it's how they were built back then. That's real iron, as in the old style core and not a more modern light weight hypersil job. But more important, it is probably not suitable for your needs. The reason is that transformers like that one were made in a time when power supplies used choke input filters and full wave rectifiers. It probably has a center tap on the secondary. I'll guess that you want to make a plate supply that has a lot of dynamic regulation because your load is going to be continually varying. You probably want to use a full wave bridge but if not, you'll want a lot of capacitance and no choke. That transformer was made for a choke input and not a lot of capacitance, maybe as little as 4 mfd. Collins also used a "resonant choke" on the 30Ks. A FWB rectifier and cap input filter on that old iron will burn it up. The secondary won't handle having the entire winding continually under load. I advise you sell it to a 30K owner and look for a more modern transformer for your project.

    K4PF, AG5CK and KN4CTD like this.
  3. KN4CTD

    KN4CTD Ham Member QRZ Page

    Ok, but just so I understand better......what is the determining factor for this? Is it because of the type of iron used in the laminations? The physical size/type of wire the secondary was wound with? Combination of the two? Just trying to better understand. And yes, my plan was to use it in a full wave bridge capacitor input configuration. Thanks.
  4. W8KHK

    W8KHK Ham Member QRZ Page

    I believe Rob (K5UJ) has provided a very accurate and comprehensive analysis of the issue, and I agree theoretically with his statements. But I would like to offer a contrasting, and somewhat conservative point of view, based upon my own personal experiences.

    First, to address your specific question about what is different in the old vs. new transformers. As Rob stated, the older transformers used a center-tapped full-wave rectifier, with one or more choke input filters. In some cases, a resonant filter (choke in parallel with a capacitor) is used to provide a higher impedance load on the rectifier. This worked well when we had oil capacitors at high voltage, low capacity, on the order of 2 to 4 microfarads for the entire filter. The load of a plate modulated AM transformer was less dynamic than an SSB linear, in that the class C amplifier provided a fixed load impedance, in parallel with the smaller varying load of the class B modulator. Today, the linear amplifier load varies from almost no load to peak load at a syllabic rate, and the linearity of the amplifier is greatly dependent upon the stability of the high-voltage supply. Non-linearity results from a sagging HV supply on voice peaks. The larger capacitors available today address this issue elegantly. The significant difference in the newer transformer is the lower voltage, higher current secondary winding, made possible by using larger gauge wire, having a smaller overall resistance, and higher current capacity, due to the fewer turns required. This is important because the filter capacitors are charged for a smaller percentage of the input AC waveform, whereas the choke input filters were able to draw a lower current during a larger portion of the AC waveform.

    So today, instead of a center-tapped secondary, with many turns of finer, higher resistance wire, we typically see a lower voltage, higher current secondary, and a voltage doubler rectifier circuit to achieve the desired high voltage required. This lower resistance secondary has less voltage drop when voice peaks demand high current to keep the filter capacitors charged. And we typically avoid a filter choke, and use large microfarad filter capacitors to provide a large storage repository to meet the high current demands of voice peaks without the voltage sagging, resulting in non-linearity. Hipersil core material is used to make the core more efficient, and save weight and space.

    So the goal in building a power supply for an SSB linear is to have good regulation with a constantly varying load, which requires a low resistance in the transformer windings, especially the secondary. In contrast to Rob's analysis, I would like to suggest that we determine whether the secondary resistance of your transformer may meet the intermittent load requirements (ICAS) you demand, since the transformer was designed for continuous commercial service (CCS).

    Looking at the manual for the transmitter in question ( we can see that the power supply is rated at an output of 2500 volts. It has a two-stage choke input filter after a pair of 866 mercury vapor rectifiers. One of the filters uses a resonant LC section, as Rob stated. A total of 4 microfarads of filter capacitance is used in the two-stage filter. Reading more details in the manual indicates that the duty cycle is 50 percent, 15 minutes on within a 30 minute period, at either 150 MA final plate current (plus modulator plate current) in AM mode, or 200 MA plate current in CW mode. At 2500 volts, we are looking at 500 watts continuous DC input for 15 minutes at a stretch, 50 percent duty cycle. The manual also states that 100 percent duty cycle is allowable with the addition of air circulation for cooling. This air is probably more important for the tube circuits, and not likely required for the transformers in the bottom of the rack.

    In order to determine whether this transformer could be used in an SSB linear amplifier, we must first look at whether it can maintain a well-regulated output voltage under dynamic load, and we must determine whether copper losses (winding resistance) will cause undue heating within the transformer. Simple resistance measurements can be performed on the secondary winding to get an idea of the voltage drop and transformer heating under load. You might consider applying a fixed load (a resistor, or string of light bulbs) that will determine secondary voltage drop under a constant load, simulating worst-case power requirements.

    It will also be necessary to determine what voltage and current will be required for your linear amplifier, and this is a function of what tubes and load impedance you plan for your RF stage. With some testing and research, this transformer might fit your needs, even though it was designed for a rather steady load with choke input filtering circuits.

    A case in point I would like to share, from my own personal experience: I built a linear amplifier in 1967, while I was in the US Air Force. Assembly was performed in my barracks at Davis Monthan AFB in Tucson, AZ. I built it, for the most part, using parts available from the junk box that my father and I collected over the years. I ran a pair of Eimac 4-400 tubes, in standard grounded-cathode configuration, using about 4000 volts on the plates. I was able to tune the amplifier to 2 KW DC input, 4000 volts at 500 milliamperes, using the power supply at hand. And I must say it was much more "wimpy" than what you might build with your Collins transformer. The amplifier still lives today, with the original UTC power transformer. In use for over 50 years and never a failure.

    The transformer is a UTC S-48 special series transformer, see page 49 in this document. (

    The transformer is rated at 1300 volts DC at 500 milliamperes using a center-tapped full-wave choke-input supply, intermittent duty (ICAS). I probably broke all the rules in the supply design, but operated it modestly. The supply used a full-wave bridge of solid-state diodes, into a group of four 14 uF 2000 volt GE pyranol (PCB filled) oil capacitors in series parallel, providing 14 uF at 4000 volts. Yes, I ran them right up to the limit of 4000 volts, and the capacitors still live. I probably would not design it that way today, a bit more conservative if you will. I was able to use the power supply in this fashion due to the short duty-cycle of the SSB amplifier, and the larger capacity of the filter.

    I used a variac to lower the plate voltage when tuning, and I used a 250 watt photo-flood bulb as a series resistance with the 115 volt primary for initial tuning. I did not have any step-start circuit. The transformer center tap was of course elevated, not grounded. I had very good regulation and a clean, splatter-free signal.

    Based upon my experience, I suggest that you evaluate the transformer, looking at winding resistance, and the potential for decent regulation in SSB mode. I would certainly not try to use that transformer with capacitor-input for continuous duty AM mode, especially for long transmissions. But considering the CCS rating and the fact that SSB is short duty cycle, if you can achieve decent voltage regulation without undue losses and heating in the transformer, it might just work for you. Look and see if you can design an amplifier to run with around 3000 volts on the plates, using a center-tapped full-wave rectifier and decent capacitance filtering. But if you want to go all-out with maximum power and have the ability for 100 percent power with unlimited key-down time, then follow Rob's suggestion and obtain a newer transformer designed for higher current, dynamic load with minimal voltage drop on current peaks. Everything Rob said is accurate, but since you already have the transformer, why not evaluate it carefully and learn a bit in the process?

    I hope this information gives you a better understanding of the issues relating to the older and newer technology regarding the high voltage power supplies for amateur use.
    Last edited: Apr 21, 2019
    W2NBC and W2VW like this.
  5. W8KHK

    W8KHK Ham Member QRZ Page

    PS: If you decide to obtain power supply components for the ultimate in QRO performance, the go-to guy for this resource is Dennis, W7TFO, on this board. He knows where all the best deals on heavy metal may be found!
  6. KN4CTD

    KN4CTD Ham Member QRZ Page

    Thank you for taking the time to better educate me on these transformers. I enjoy learning about and tinkering with vacuum tube linear amplifiers. Most of my experience (which is still limited btw), comes from using smaller, more modern plate transformers out of known applications where it is easier to just say for instance “well, this transformer was able to run 2 3-500z at legal limit out pep at a given plate current, so it should work in this application.” Not a very technical approach, I understand! I have access to other smaller, more modern transformers to use when I actually decide to do something else. I’ve never messed with a transformer this old and I guess, in my mind, I thought, given its size, it would be more capable even though the spec plate on it clearly states it was only rated for 300 milliamperes continuous. I was hoping that they were overbuilt and underrated for their capabilities. I didn’t realize they were actually wound with smaller gauge wire thus making the resistance higher. What you’ve said makes good sense though. I have no plans to build any kind of “brick on the key” supply using large tubes or anything like that. I have a small supply of Russian triodes (Gi7b and gi-46b) to choose from and I only operate ssb voice. One day, I would like to find or have wound a good plate transformer with multiple output taps so that I may use it to power different rf decks by only changing transformer taps. Thus eliminating the need to fit a transformer, filter/rectifier board into every rf deck and making them easier and cheaper to build.
  7. KM1H

    KM1H Ham Member QRZ Page

    A resonant choke input filter has been used for decades by companies such as Henry for SSB and I suggest starting with their manuals on:

    Stick to the original standard full wave rectifier the transformer was designed for and followed by as much C as needed plus a resistive ~100 ma bleeder load to get reasonable regulation. Some of the current amp offerings on the market are no better than ~15 %
    While SSB is dynamic and regulation varies it never reaches what you would measure at full key down in real world use. The LCR time constant of the PS mitigates this rather well and since I own and have used a spectrum analyzer for several decades I can see that the effect on IMD is far from as bad as some without that experience would have you believe. Done right while keeping in mind the limits you have to work with it will perform just fine.

    Use a pair of 3B28's for rectifiers (not 866A's) which were common in commercial and military use to provide a softer 120 Hz load on the transformer than instantaneous switching SS diodes.

    I learned all of this the hard way in the 70-80's when building 4-1000A amps for others using vintage iron and smoking a few in the process. Done with sufficient thought they will be just fine BUT I would start with using a Variac along with sufficient R to run a reasonable starting load and bringing that transformer up slowly while monitoring Input and Output current. No rectifier circuit wanted at this stage, just seeing if there is excessive moisture in the windings that might have to be baked out using a different method.

    The hardest part may be finding the proper choke and resonating cap; they do show up on here and elsewhere. Another option is to use a regular choke of the minimum inductance required which is covered in the various ARRL and other handbooks well into the 70's. The reason for the resonant choke was simply MONEY as lots of Henries was expensive and they are not that reliable long term as many have found out as the caps develop leakage which changes the value and things go BANG very fast.

    Ham since 1955
    Built first KW AM amp in 1957 while in HS using WW2 surplus
  8. K5UJ

    K5UJ Ham Member QRZ Page

    Congratulations on taking the initiative to build something. Building, restoring and repair is really more about ham radio than operating. Operating alone can get boring and boredom is the main reason fellows quit the hobby. You have options: Sell the 30K transformer, and build with something else, buy a 30K1 ($$$$) and become an AM operator, or ditch the SSB leenyar plan and instead, homebrew a ~250 watt phone transmitter with the 30K transformer. Ambitious yes, but at a minimum, I endorse the idea of making it available to a 30K owner because their replacement options are fewer, whereas you have a much wider set of options for your amplifier plate supply.
    KN4CTD likes this.
  9. KN4CTD

    KN4CTD Ham Member QRZ Page

    I was basically just confused about this when I picked it up yesterday. I’m more interested in finding a transformer that will handle more current with multiple output voltages so as could be used with multiple rf decks using a full wave bridge with modern electrolytics. I don’t have much in this transformer and I’ll probably pass on even using it. I would make it available to anyone who could use it as I have several smaller, more compact transformers from which to choose from already.
  10. N2EY

    N2EY Premium Subscriber QRZ Page

    That sounds like the best plan.

    What you have is a transformer designed for a completely different technology than what you want to use. It could be used if you were willing to do what the Ancient Ones did, but it's clear your not into that. Plus there are probably 30K-x owners out there who would love to have an exact replacement plate transformer.

    The reasons it's so big and heavy for so few VA are:

    - the iron they had back then wasn't the iron we have today
    - the end bells
    - the insulation they had then isn't the insulation we have today (both on the wire and between layers)
    - the secondary has an enormous number of turns of quite fine wire.

    73 es GL de Jim, N2EY

Share This Page

ad: SDRKits-1