build a high-voltage ham radio power supply capable of powering a variety of radios, including a hallicrafters hurricane
by John P Young, W3AFC
NOTE: All material on this web site is Copyrighted. You can use it for your own purpose or your club's, just don't sell it for profit or claim that you wrote it, or bad things will happen.
Ever since I was a teenager, I've wanted a Hallicrafters Hurricane so when I saw a mint-looking Hurricane on the Luther Auction site, I had to take a chance. The problem was, it came without the power supply. This is a common problem when buying any Amateur/HAM radio whether it's a Collins, Heathkit, Swan, Drake, or other, and it puts the new owner in the position of having to either build their own supply, search for a working supply at a HamFest, or wait for one to come up on an auction site. This was the conundrum that I faced.
I couldn't wait to find a P-2000 power supply or build my own, so I modified one of my Swan 117XC power supplies to make sure the receive worked. After using it as a receiver for awhile, I added transmit using the 800-volt tap in the 117XC. Everything worked, and it tuned up to around 200 watts PEP. I had hoped for more, but the lower output was expected since the Swan supply was heavily loaded down in transmit/Tune mode by the Hurricane's 8122 RF tubes. Nevertheless, I worked stations as far away as Japan and Oman with it with just an 87-ft inverted "L" antenna, and got lots of "Great Audio" reports. Once I determined that the radio was in good condition and would become a major part of my Shack, I started the process of researching cabinets and transformers, and acquiring the other parts needed. I decided from the beginning to build something that would power not only my new radio but also others in my shack. I wasn't interested in driving the Hurricane to a full 1000 watts PEP output for three reasons:
1. This is not the 1960's. Back then, homes were generally spaced farther apart, With today's congested neighborhoods, 1000 watts might
cause unwanted interference and complaints.
2. The early Hurricanes experienced heat-related problems due to the extreme high power, not the least of which was eventual destruction of the
main tank loading coil. A number of recalls and service bulletins were issued by Hallicrafters to address that issue.
3. I've worked stations as far away as Antarctica using my Swan 700 series radios, which generate 300-400 watts PEP. My goal was to have that
level of power in daily use, with the option of upping the power to 600 watts or more in high power mode.
One thing I quickly discovered was that such a goal could not be reached with a single transformer, unless I contacted someone like Hammond or Triad and asked that they build a custom unit which would be extremely expensive. That led to the designs that I show below. A lot of experimentation went into my supply. Basically, I created a prototype and schematics for two other units that are more Hurricane-friendly. All can be used for other radios, although they might be a little "hot" for the Drake TR-3 and 4's, which is why I will discuss mods in my next update to make them more adaptable to the Drakes. However, given the $750 + price tag for this supply, Drake owners might want to simply search for a factory power supply and rebuild it.
My model is literally two separate supplies in a single case, fed by a 12-gauge, 3-prong cord. Each transformer is a toroid made by Antek, with dual 115V primaries and dual full-voltage secondaries. The power switch for the high voltage section is a DPDT switch that switches the primaries between series and parallel, to switch between low (1000-1100 volts) and high voltage (2,000-2200 volts) output. Most of the time I use the Low Voltage mode, which yields a leisurely 300 watts PEP - more than enough to make contacts all over the world. High power (photo below) bounces the power meter up over 600 watts which allows me to break through pileups easily. It's not as aesthetically pleasing as the original power supply and it lacks the safety interlock feature and the prototype lacks the screen grid (G2) current metering circuit, but I have a Triplett external meter to monitor that. Two of my schematics DO have a screen current milliammeter.
So, here are some photos, and downloadable Excel files for the estimate.I'm being forced to upgrade to a Premium plan at the end of October, so maybe I'll be able to post icons for the Excel files and videos then.
UPDATE 9/16/24. The prototype power supply with the wafer switch is turning out to be the perfect supply for the Hurricane.
- The wafer switch design allows the Operator to tune up the radio with the high voltage enabled, and then switch the selector to
"Receive" mode while leaving the High Voltage enabled. I use the 2KV position, and set the VFO to the middle of the band.
- This kills G2 to the 8122's, effectively turning them off. With the RF tubes off, the radio runs lukewarm. You can browse around
the band until you hear a contact that you would like to make. Then you can turn the selector switch to "Send/Receive" and
make your call. Depending on how far you are from the center of the band, you might need to adjust the Preselector a bit.
- After you make the contact, you can switch back to the Receive position. The radio will immediately begin to cool down.
This is exactly what I sought to accomplish when I built this supply. I used to use my Yaesu FT-101E or my Kenwood TS-830S hybrids as receivers until I heard a desirable contact. Then I switched on the RF tubes. The trouble was, by the time the RF tubes warmed up, there were a hundred other Ops screaming out their call signs. This supply setup reduces that likelihood to a minimum.
I couldn't wait to find a P-2000 power supply or build my own, so I modified one of my Swan 117XC power supplies to make sure the receive worked. After using it as a receiver for awhile, I added transmit using the 800-volt tap in the 117XC. Everything worked, and it tuned up to around 200 watts PEP. I had hoped for more, but the lower output was expected since the Swan supply was heavily loaded down in transmit/Tune mode by the Hurricane's 8122 RF tubes. Nevertheless, I worked stations as far away as Japan and Oman with it with just an 87-ft inverted "L" antenna, and got lots of "Great Audio" reports. Once I determined that the radio was in good condition and would become a major part of my Shack, I started the process of researching cabinets and transformers, and acquiring the other parts needed. I decided from the beginning to build something that would power not only my new radio but also others in my shack. I wasn't interested in driving the Hurricane to a full 1000 watts PEP output for three reasons:
1. This is not the 1960's. Back then, homes were generally spaced farther apart, With today's congested neighborhoods, 1000 watts might
cause unwanted interference and complaints.
2. The early Hurricanes experienced heat-related problems due to the extreme high power, not the least of which was eventual destruction of the
main tank loading coil. A number of recalls and service bulletins were issued by Hallicrafters to address that issue.
3. I've worked stations as far away as Antarctica using my Swan 700 series radios, which generate 300-400 watts PEP. My goal was to have that
level of power in daily use, with the option of upping the power to 600 watts or more in high power mode.
One thing I quickly discovered was that such a goal could not be reached with a single transformer, unless I contacted someone like Hammond or Triad and asked that they build a custom unit which would be extremely expensive. That led to the designs that I show below. A lot of experimentation went into my supply. Basically, I created a prototype and schematics for two other units that are more Hurricane-friendly. All can be used for other radios, although they might be a little "hot" for the Drake TR-3 and 4's, which is why I will discuss mods in my next update to make them more adaptable to the Drakes. However, given the $750 + price tag for this supply, Drake owners might want to simply search for a factory power supply and rebuild it.
My model is literally two separate supplies in a single case, fed by a 12-gauge, 3-prong cord. Each transformer is a toroid made by Antek, with dual 115V primaries and dual full-voltage secondaries. The power switch for the high voltage section is a DPDT switch that switches the primaries between series and parallel, to switch between low (1000-1100 volts) and high voltage (2,000-2200 volts) output. Most of the time I use the Low Voltage mode, which yields a leisurely 300 watts PEP - more than enough to make contacts all over the world. High power (photo below) bounces the power meter up over 600 watts which allows me to break through pileups easily. It's not as aesthetically pleasing as the original power supply and it lacks the safety interlock feature and the prototype lacks the screen grid (G2) current metering circuit, but I have a Triplett external meter to monitor that. Two of my schematics DO have a screen current milliammeter.
So, here are some photos, and downloadable Excel files for the estimate.I'm being forced to upgrade to a Premium plan at the end of October, so maybe I'll be able to post icons for the Excel files and videos then.
UPDATE 9/16/24. The prototype power supply with the wafer switch is turning out to be the perfect supply for the Hurricane.
- The wafer switch design allows the Operator to tune up the radio with the high voltage enabled, and then switch the selector to
"Receive" mode while leaving the High Voltage enabled. I use the 2KV position, and set the VFO to the middle of the band.
- This kills G2 to the 8122's, effectively turning them off. With the RF tubes off, the radio runs lukewarm. You can browse around
the band until you hear a contact that you would like to make. Then you can turn the selector switch to "Send/Receive" and
make your call. Depending on how far you are from the center of the band, you might need to adjust the Preselector a bit.
- After you make the contact, you can switch back to the Receive position. The radio will immediately begin to cool down.
This is exactly what I sought to accomplish when I built this supply. I used to use my Yaesu FT-101E or my Kenwood TS-830S hybrids as receivers until I heard a desirable contact. Then I switched on the RF tubes. The trouble was, by the time the RF tubes warmed up, there were a hundred other Ops screaming out their call signs. This supply setup reduces that likelihood to a minimum.
Almost all of the components are mounted on a 0.250-in thick aluminum plate The advantage of this is that the plate was easy to drill and tap for the mounting screws, and it serves as a common source of multiple continuous (single point) grounds.
Here are three schematics, along with the Excel files. The cost for the first two units is essentially the same, which is about $750 if you have to buy all the parts. I had some already in my shack. The 2400 VDC version is simply the second one using a huge, 36-lb 1500 VA AS-15T950 transformer from Antek. All of the components that I picked for the first two supplies will handle that high voltage, but I haven't found a case yet that will accomodate the size and weight of the AS-15T950. The 2400 Volt version (2,375 actually) will cost around $900. I'll frequently update this page.
I tried the large TRIAD transformer shown in two of the photos. It puts out an impressive 1780 VAC, but at a max of 250 mA. I may play around with it someday in a different power supply. It has a single non-tapped secondary.
These schematics show only two indicator lamps. My original plan was to have two red HV indicators: The lower one would light up when I switched on the 1,000 plate bias, and the upper one would light up when the PS was switched to 2,000 volts. But the series/parallel primary scheme made that difficult without additional electronics. My prototype has two reds and one yellow.
The Antek transformers get only lukewarm so I didn't need to install a fan. The rectifier bridges also remain cool because they are mounted on a 1/4 inch thick aluminum plate, and they are handling a load that's well below their maximum PIV rated values.
I tried the large TRIAD transformer shown in two of the photos. It puts out an impressive 1780 VAC, but at a max of 250 mA. I may play around with it someday in a different power supply. It has a single non-tapped secondary.
These schematics show only two indicator lamps. My original plan was to have two red HV indicators: The lower one would light up when I switched on the 1,000 plate bias, and the upper one would light up when the PS was switched to 2,000 volts. But the series/parallel primary scheme made that difficult without additional electronics. My prototype has two reds and one yellow.
The Antek transformers get only lukewarm so I didn't need to install a fan. The rectifier bridges also remain cool because they are mounted on a 1/4 inch thick aluminum plate, and they are handling a load that's well below their maximum PIV rated values.
high-voltage power supply - improved design
When I built my universal Ham radio/Hurricane power supply, my goal was to keep it Old School. Transformers, Bridge or full-wave rectifiers, and resistive voltage dividers. And it’s worked well. But the more I used it, the more I grew tired of the heat and voltage drop losses in the 280-volt General Bias circuit. It was a waste of energy, so I decided to replace the entire fixed divider with a smooth, variable, heavy-duty solution.
I chose a high-power IGBT, something that I’ve had experience with. The one I chose - a Fairchild HGTG12N60A4 - was a whopping $8.00 for three. Rated at 600VDC at 23-54 amps depending on the temp it barely gets warm because V=350 VDC and I=150 mA or lower. Mine works beautifully. Instead of a General Bias level that starts at 300-310VDC and then drifts down to 260-270VDC as the Hurricane’s tubes warm up, I get whatever I set the control potentiometer to. If I set it to 280 VDC when the radio is fully warmed up, at the most it might start at 283 VDC but it settles at 280 VDC within seconds. Everything runs lukewarm, and the voltages on Rectifier Stack 1 (low/medium voltage supply) don’t sag anymore, resulting in a higher PEP output. Even better, I can set the general Bias control at voltages from 200 VDC to 300 VDC. So 250 VDC for the Drake TR rigs is there.
The current is only 4.54 mA, and the power consumption is only 1.6 watts, allowing me to use a 4-watt potentiometer, and 5-watt resistors in the voltage divider circuit. Connected to the IGBT, I can have more power than I'll ever need for HAM radio equipment.
I chose a high-power IGBT, something that I’ve had experience with. The one I chose - a Fairchild HGTG12N60A4 - was a whopping $8.00 for three. Rated at 600VDC at 23-54 amps depending on the temp it barely gets warm because V=350 VDC and I=150 mA or lower. Mine works beautifully. Instead of a General Bias level that starts at 300-310VDC and then drifts down to 260-270VDC as the Hurricane’s tubes warm up, I get whatever I set the control potentiometer to. If I set it to 280 VDC when the radio is fully warmed up, at the most it might start at 283 VDC but it settles at 280 VDC within seconds. Everything runs lukewarm, and the voltages on Rectifier Stack 1 (low/medium voltage supply) don’t sag anymore, resulting in a higher PEP output. Even better, I can set the general Bias control at voltages from 200 VDC to 300 VDC. So 250 VDC for the Drake TR rigs is there.
The current is only 4.54 mA, and the power consumption is only 1.6 watts, allowing me to use a 4-watt potentiometer, and 5-watt resistors in the voltage divider circuit. Connected to the IGBT, I can have more power than I'll ever need for HAM radio equipment.
There was one fly in the ointment, as they say. IGBT's are current amplifiers, and since my power supply relies on bridge rectifiers and large electrolytic capacitor values, a very small amount of ripple existed along the line. It wasn't enough to be heard on my Hurricane, but the IGBT amplified it to a level that was noticeable.This irritated me and I was concerned that it might come though on my transmitted signal, so I set about the task of eliminating it.
The hum issue was an easy fix and now my Hurricane receiver is quiet once again. A 50mH (or higher) iron-core filter choke followed by a decent high-value electrolytic restored peace and quiet. I had planned to use a filter-choke-filter setup early on but space became an issue.
UPDATE: A second fly appeared. IGBT's have switching frequencies between 20 & 50kHz, which is a bad thing in a receiver, so I needed to come up a second filter, all of which I put on a little perf board on standoffs above the Stack 1 bridges. I'm running out of space.
A simple CLC low-pass filter, consisting of two 0.1uF 630V Metallized Polypropylene Film caps and a 660uH toroidal choke solved the noise issue. This combination rolls everything off at around 13kHZ. I didn't have a 660uH toroid so I connected two 330uH's in series on a terminal strip. It's ugly but it works!
The choke and electrolytic cap are for hum. It's a filter-choke-filter setup where the Panasonic 330uF capacitor from bridge rectifier #1 is filter #1. Number 2 is a 180uF 450-volt. The filter choke can be anything over 50-100uH, with a 350mA or greater rating.Here's a schematic for the IGBT setup and the filter board.
The hum issue was an easy fix and now my Hurricane receiver is quiet once again. A 50mH (or higher) iron-core filter choke followed by a decent high-value electrolytic restored peace and quiet. I had planned to use a filter-choke-filter setup early on but space became an issue.
UPDATE: A second fly appeared. IGBT's have switching frequencies between 20 & 50kHz, which is a bad thing in a receiver, so I needed to come up a second filter, all of which I put on a little perf board on standoffs above the Stack 1 bridges. I'm running out of space.
A simple CLC low-pass filter, consisting of two 0.1uF 630V Metallized Polypropylene Film caps and a 660uH toroidal choke solved the noise issue. This combination rolls everything off at around 13kHZ. I didn't have a 660uH toroid so I connected two 330uH's in series on a terminal strip. It's ugly but it works!
The choke and electrolytic cap are for hum. It's a filter-choke-filter setup where the Panasonic 330uF capacitor from bridge rectifier #1 is filter #1. Number 2 is a 180uF 450-volt. The filter choke can be anything over 50-100uH, with a 350mA or greater rating.Here's a schematic for the IGBT setup and the filter board.
MODIFIED SWAN 117 XC
I received a request for information on how I modified a Swan 117XC to make it work with the Hurricane, so here are some pictures, along with the schematic.
The yellow text in the photo at the end of this gallery states that I hoped to get 350 Watts PEP out of the Swan 117XC supply, but I was assuming a 45/55% input/output ratio, similar to sweep tubes. The 8122's are much more rugged, but their efficiency is lower.
Since the radio was so pristine, I was very cautious while I assembled the plug, verifying every connection with my DMM before applying power. The first thing I connected was was the filament feed, then the blower, negative bias, main bias, and so forth. The 800V cable was done last, after I was convinced that nothing was wrong inside the Hurricane.
I used an RCA pin plug cable and went right to the built-in Swan speaker, which yields very good audio.
The yellow text in the photo at the end of this gallery states that I hoped to get 350 Watts PEP out of the Swan 117XC supply, but I was assuming a 45/55% input/output ratio, similar to sweep tubes. The 8122's are much more rugged, but their efficiency is lower.
Since the radio was so pristine, I was very cautious while I assembled the plug, verifying every connection with my DMM before applying power. The first thing I connected was was the filament feed, then the blower, negative bias, main bias, and so forth. The 800V cable was done last, after I was convinced that nothing was wrong inside the Hurricane.
I used an RCA pin plug cable and went right to the built-in Swan speaker, which yields very good audio.
Any questions? Feel free to email me at [email protected].