Technical Topics



Vacuum Tube Liquid Cooling System

This article describes an alternative to water as the cooling liquid for amateur radio vacuum tube power amplifiers. This cooling system is designed for adequate vacuum tube cooling, reasonable cost and very low maintenance.

Liquid cooling for the GS35B requires an anode coolant interface that clamps onto the vacuum tube. The GS35A version of this tube has a manufacturer's liquid cooler instead of the traditional air cooler of the GS35B. The liquid cooling jacket shown here was partially made from a damaged air-cooler by Mats Bengtsson, KD5FZX.

Choice of cooling fluid

All amplifiers previously described in the amateur literature have used water as the cooling fluid. This is probably because water has the highest specific heat capacity of all commonly available fluids and it is inexpensive and easy to get. The specific heat rating (the amount of heat needed to raise the temperature of one gram of a substance by one degree) of water is 1.0 and other cooling fluids considered here have lower ratings. However, a closer look at using water for the cooling fluid shows that there are tradeoffs necessary in the implementation.

Water can be conductive. This is a problem when using water to cool a tube anode that is energized at a very high voltage. Distilled or de-ionized (they are not the same) water helps, but the water must routinely be replaced because it becomes contaminated by ions leached from the cooling system hardware or from the atmosphere. Because of the molecular structure of water, it has a good ability to form strong bonds to dissolved ions.

A high voltage leakage current detector will indicate when the dissolved ion count becomes too high and it is time to change the water, but this requires additional hardware.

Water also supports life. Of course this means that things can grow in the water. The common green algae that is visible in aquariums is a familiar reminder of this. Using bleach or other simple chemicals as a remedy is a big mistake because they increase the conductivity of the water and make it unsuitable. Water is an efficient and inexpensive cooling fluid, but the continuous replacement and maintenance requirements justifies a look for a reasonable alternative.

Insulating liquids like transformer oil have very high dielectric strength. The tradeoff here is the specific heat of these types of fluids is lower than water with about a .4 to .5 value. However, transformer oil can sometimes be hard to obtain in small quantities of a couple of gallons and the increased viscosity over water implies it is harder to pump.

Mineral oil is not a good alternative because it can become contaminated like water.

A suitable alternative fluid that is easily obtained, inexpensive, has high dielectric strength, high flash point, and easy to use is common automotive automatic transmission fluid (ATF). The MSDS information for Pennzoil Dextron/Mercon ATF reveals these properties:

Dielectric strength: 35 kV
Flash point: 399 degrees F
Weight: 7.25 lbs/gal
Color: red

The biggest disadvantage of ATF is a lower specific heat rating when compared to water. This means it is a less efficient cooling fluid although this is not a hard problem to overcome. A more efficient cooler and/or larger pump can make up for the cooling efficiency difference of the two fluids.

ATF by design also contains friction additives that may degrade pump seals over the long term. This potential problem can be eliminated through the correct choice of pump type. A magnetic drive pump has no seals.

What about transformer oil?

PCB (polychlorinated biphenyl) was banned in the USA and Canada in 1976. This product was of course used in high voltage transformers, capacitors, and heat exchangers. Some modern replacements for PCB are fluids made with terphenyl. Their properties include high temperature and good thermal stability.

Therminol 66 transformer oil appears to be widely used and seems to be an excellent transformer/heat exchanger oil. I suggest that you avoid the LT version since it has a low flash point of about 134 degrees F. The 66 version has a rating of 363 degrees F. The specific heat rating is .409 which is similar to many insulating oils.

I would choose this fluid for an oil-cooled amateur tube amplifier or dummy load application provided it was reasonably priced and was easy to get.

Amplifier heat reduction requirements

To get an idea of the magnitude of heat involved, consider this example. Assume that two KW of heat is wasted by the amplifier and will be used to heat the ATF cooling liquid. For simplicity, we will not be concerned with the cooling ability of the radiator and fan. Just pretend that it does nothing to help the temperature reduction. Also assume that two gallons of ATF is used in the system.

If we place a two KW source into two gallons of ATF and wait one minute, what temperature does the ATF become? This equals the amount of cooling required per minute to maintain system equilibrium.

This may be expressed as:

Power BTU/min = (Mass rate) x (Specific heat) x (Temperature increase)

Note that Watts x .056884 = BTU/Min

Since 2 kW of power equals 113.8 BTU/minute, and the mass of two gallons of ATF is about 15 lbs, we can solve for temperature rise.

113.8 BTU/min = (15 lbs Mass) x (.5 Specific heat of ATF) x (Temp. increase in degrees F)

This means that after one-minute of key-down amplifier operation the two-gallons of ATF will rise by 15.16 degrees F. This indicates that the cooling radiator and fan needs to lower the temperature of the liquid about 15 degrees in one minute.

You may measure the effectiveness of a homemade cooling system without even turning on your amplifier. To do this, just heat the entire volume of cooling fluid in your system by about 40 degrees or so. Place the fluid in your system and circulate it enough to fill the radiator. Measure the temperature of the fluid in the cooling reservoir. After one minute of circulation again measure the temperature.

By rearrangement of the formula above you can determine how much power your system is capable of dissipating. Also remember that intermittent operation of your amplifier results in less heat generated.

Radiator and Fan Design

It is difficult to determine the effectiveness of another design since other radiators are made differently. The number of tubes, physical size, number of fins and surface area, and radiator material all affect the efficiency. The fan that you use may be different from this example too. By observing the effectiveness of this system's cooling you can get an idea of what is required for a similar system.

Generally, the larger the surface of the radiator and faster air flow will result in a lower system coolant temperature. For example, the GS-3A tube (only available as a liquid cooled version) specification indicates a maximum tube temperature of 150 degrees C or 302 degrees F. It is essential to keep below this number under all conditions. This is still 100 degrees below the flash point of ATF so a very comfortable fire safety margin exists.

The Lytron model MCS30 cooling system below was purchased brand new on eBay. This model is designed specifically for oil and will intermittently cool 2 KW. It is located external to the amplifier and doesn't occupy much space or make a lot of noise. Of course alternate home-constructed systems made of individual components could also be sucessfully used after first testing their efficiency.

Mats, KD5FZX has a page describing home made coolers.

You still need to air cool the GS-35B tube seals even with liquid cooling of the anode.

It is necessary to provide some air cooling for the tube seals. The heat generated from the filament and cathode is not conducted enough to the plate since the ceramic is a very poor conductor, and the internal heat conduction from these elements to the plate is too low . The GS35B generates 36W in the filament and up to 22W cathode-grid dissipation heat.

This heat is easy to remove with a small quiet muffin fan. You still need a small fan to remove about 60W.