Jump to Navigation

Taking the Mystery Out Of Cap Tubes

Everyone in the refrigeration industry has heard of and has probably had some experience with the cap tube.  It seems to be one of the most feared, condemned, and misunderstood devices in the industry.  One big reason for this is that it is not an “off the shelf” replacement item.  It is difficult to replace. Cap tubes have become popular with manufacturers because you can’t beat the economics of the cap tube.  It is cheap, and will outlast all the other components of a system.  It is self-equalizing, saving the cost of starting components, and cap tube systems use much less refrigerant than any other type of expansion device.


Two laws govern the flow rate of refrigerant through a cap tube.


1.  Liquid flows faster than gas.

2.  The colder the liquid, the faster the flow rate.


If we have a cap tube with water flowing through it, we would find that the pressure drop is linear; in other words, for every foot of tubing, we have an equal pressure drop.  The flow rate through the tube is also linear; that is, as pressure increases, the flow rate increases in direct proportions.  With refrigerant flow through a cap tube, something happens that destroys this linearity.  This resulting non-linearity separates the cap tube from a simple pressure-reducing device to being a refrigerant control.  That something is flash gas.


As subcooled refrigerant enters a cap tube, it acts like water.  For the first number of feet, the temperature remains the same and the pressure drop is linear; for every foot of tubing, we have equal pressure drop.  As the refrigerant continues, the pressure is further reduced until the pressure is below the saturation point of the refrigerant.  At this point, the refrigerant flashes (boils) and bubbles form in the cap tube.  Where the first bubble occurs is called the bubble point.  These bubbles restrict the flow of liquid.  This restriction causes pressure to drop at a higher rate, causing more flashing, more restriction, etc.  The main reason for the bubble point occurring where it does is due to subcooling.  The more the refrigerant is subcooled entering the cap tube, the longer the liquid length in the cap tube before the bubble point.  After the bubble point is called the two-phase length.  The bubble point and two phase length gives the necessary control of the refrigerant flow.  A cap tube does have some load adjustability; nowhere near what a TXV can do, but as the ambient changes, the degree of subcooling will change in an air-cooled condenser, and as a result, the bubble point and two phase length will change.  Although head pressure has an effect on flow rate, it’s the changing two-phase length that counters this action.  Subcooling the refrigerant has the greatest effect on flow rate. It is very difficult to field apply a cap tube.  The manufacturers’ design engineers have to go through many tests before establishing an optimum bore and length.  It may change many times before they achieve the most efficient system.  Manufacturers take weeks, even months to finalize the proper cap tube(s) for a particular system.


While difficult, field replacement of clogged or restricted cap tubes can be accomplished.  First, get the cap tube from the manufacturer of the equipment.  They have done all the testing for that system and have determined the correct bore and length.  You should not have to experiment when using their cap tube.  If this is impossible, you will need to experiment in replacing the cap tube.  If you are replacing a cap tube on a system that has multiple cap tubes distributing refrigerant to the evaporator, try and ascertain the length and bore of the cap tube.  Measuring the length of the cap tube is easy, but finding the diameter of the bore can be difficult.  The OD of a cap tube does not necessarily indicate its I.D.  The I.D., or bore, is the most important measure.  The I.D. is measured in thousandths, such as .049 or .036, etc.  Use a cap tube gauge to determine the bore.  As small an increase of .005 inches in a cap tube of equal length can double its flow rate.  The OD is only important to the extent the cap tube will fit conveniently where the old one was.  Always cut cap tubes with a file.  There are cap tube files available.  Simply notch around the tube.  Do not file down to the I.D.  Snap off the tube by bending at the notch.  Clean out any burrs.  Always hold the tube downward so small particles don’t enter the cap tube.  Deburring is very important.  Any burr left will cause a restriction.


If the exact bore and length of the cap tube cannot be determined on a multiple tube coil, you will have to replace them all.  If all the cap tubes are not exactly alike, you will get unequal flow through the evaporator and experience all the problems of unequal loading of evaporator circuits.


There are charts available to show recommended cap tubes for different applications.  These charts are approximate, and due to varying conditions in the field, you may have to experiment when replacing cap tubes.  Remember that, by varying the refrigerant charge to change the amount of subcooling, you can possibly achieve the design parameters.  Naturally, don’t push the subcooling temperatures too far from normal or other problems will result.


Some contractors have successfully retrofitted troublesome cap tube systems with a TXV.  If you want to mount a TXV on a cap tube system, keep the following in mind.  On multiple circuited coils, you’ll need a distributor.  Make the distributor tubes to the coil as close to being of equal length and size as possible to get equal distribution of the refrigerant.  You will need to add hard start components to the compressor if it does not already have them, since you will be starting against an unequalized system.  If you use an Alco take-a-part TXV, it is possible to drill a bleed hole in the valve body to equalize on the off cycle.  Consult Climatic Control Company for the proper size of this hole.  Do not add a receiver to the system with a TXV that has a bleed port.  Remember, the refrigerant will bleed through to the low side of the system on the off cycle, just like a cap tube does, so we want as little refrigerant as possible in the system to minimize flooded starts.


A final word about cap tubes.  A manufacturer painstakingly engineered a cap tube system.  In comparison to other systems, it is very critical with very little tolerances.  Cap tube systems cannot afford to lose much refrigerant and will not tolerate an overcharge.  The cap tube itself cannot change its flow rate and a restriction throws it off balance.  It is a factory pre-set system with no adjustments.  Cap tubes have no moving parts, and don’t wear out.  A well-engineered cap tube system is an excellent system and a service technician can become proficient in servicing them.


There are two types of R4795’s.  The R4795A is recycling and the R4795D is non-recycling.  (Honeywell has informed us that the R4795D will no longer be available by about the end of 1995.  (Note: R4795A are still available as of 5-5-00.)  The R4795’s were superseded by the R7795 series (discussed later) and now the RM7895 series is the latest upgrade. 
You may be asked how long purge timing should be for the plug-in purge card.  Purges are supposed to be long enough to make four air changes before lighting the burner.  If you knew the CFM rating of the blower, and the volume of air in the boiler and chimney, you could calculate the purge timing needed to charge the air four times.  Since most of the R4795’s will be replacements, use the same timing as the existing card. 
Low voltage controllers cannot be used with R4795’s.  The T & T terminals on an RA890 are now 6 & 7, where an airflow switch is connected.
Purge timing does not start counting until the airflow switch closes these contacts.  Once the purge has timed out, the lighting sequence is the same as the RA890’s.  The amplifier circuit is energized during purge so we have safe start check.  If a flame simulating condition is present during purge, the flame relay coil, 2K, will energize preventing ignition, but the burner motor will continue to run.  This will give continuous purge, a “safe” failure condition.  The relay will not “lock-out”.  If the flame simulating condition, or real flame, goes out, the start-up will proceed.  If a purge card fails or is not installed correctly, the burner motor, on a call for heat, will run but pre-purge cannot be completed so ignition cannot occur resulting in a continuous purge.
If the airflow switch doesn’t close, or opens during pre-purge, the purge cannot be completed, and once again, the burner motor will run but no ignition can take place.
If the airflow switch opens during the run period, terminals 3, 4, and 5 will be de-energized, dropping out the main valve, pilot valve, and ignition.  Terminal 8 will remain energized so the burner motor will continue to run.  If the airflow switch closes, the purge timing will start and the start-up sequence will begin again.  Note that no lockouts have occurred which have to be manually reset.  Lockout requiring manual reset happens when no flame is detected after purge.  Flame relay, 2K, will not energize and the safety switch will heat and lockout the control in about 15 seconds. If there is a flame failure during run, terminals 3, 4, and 5 are de-energized; pilot, ignition, and main valve.  If airflow is still proven, an R4795A will begin purge timing and attempt to re-light.  It will make only one try.  An R4795D will not recycle.  An R4795D will lockout on flame failure during run.
An R4795D differs from the A series in safe start check.  If a flame is detected during pre-purge (2K relay energizes), the purge will stop and safety lock-out will occur in about 15 seconds—the time it takes the safety switch to heat up.  These two things are the only differences between R4795A and D.
The next upgrade of the R4795’s was the R7795 series.  The R7795 series used more solid-state technology.  The R7795’s still used plug-in purge timers, ST795A’s, but the amplifier is not plug-in or interchangeable.  Therefore, an R7795 has to be selected with the correct amplifier to match the scanner.  R7795A’s are used with UV detectors and B’s are flame rectification.  A’s and B’s are intermittent pilot models.  R7795C’s and D’s are interrupted pilot models, the C’s for UV detectors, the D’s with flame rectification detectors.  R7795’s require a Q795 sub-base.  Their operation is the same as the R4795’s.
In light of the RM7800 series, do not upgrade a customer from an R4795 to an R7795.  Always upgrade to the 7800 series.  Honeywell is only keeping the R7795 available due to O.E.M. demand.  To an O.E.M., the R7795 is less expensive than the 7800 series and OEMs are very, very price conscious.  With the demise of the R4795 series, the 7800 series will be the service industry’s control of choice.
To select an RM7895 system to replace an R4795 system, some decisions have to be made.  All R4795’s were intermittent pilot.  We can now choose intermittent pilot, the RM7895A or B, or interrupted pilot, the RM7895C or D.  Intermittent pilot means the pilot is on during the run period.  Interrupted pilot means the pilot is shut off during the run period.  All RM7895’s have an initiate sequence that lasts at least 10 seconds on initial powering of the relay.  During this ten seconds, the relay is checking that the line voltage is within 132 VAC and 102 VAC and line frequency is within plus or minus 10%, or 66 HZ and 54 HZ.  If any of these tolerances are not met, the 10 second initiate sequence will go into a hold condition until the tolerances are met, and if not met the RM7895 will lock-out in four minutes.  If, at any time during this hold period the tolerances are met, the 10-second initiate sequence will restart checking voltage and frequency again.
After passing the initiate sequence, the relay goes into stand-by.  Stand-by can be any length of time.  Stand-by simply means the control is waiting for a call for heat.  On a call for heat, terminal 4 is powered; the blower motor and pre-purge begins.  Pre-purge timing is whatever ST7800A plug-in card was selected, from 2 seconds to 30 minutes.  The airflow switch (AFS), installed between terminals 6 and 7, must close within the timing of the short timing purge cards, 2, 7, or 10 seconds, or within 10 seconds for longer timing purge cards.  The purge timing does not start to count until the AFS closes.  Should the AFS not close within the specified time or 10 seconds, whichever is shorter, the control will recycle or lock-out, depending on jumper 3 being intact; recycle or cut; lock-out.
All RM7895’s have three configuration jumpers.  Jumper number 3 is the jumper that governs what happens if there is AFS failure.  If the AFS opens at any time after it has been made, that is in pre-purge, ignition trials, or during run, the RM7895 will recycle if jumper number 3 is left intact or if the jumper is cut the control will lock-out.
All RM7895’s have three jumpers that can be cut or left alone.  They are labeled JR1, JR2, and JR3.  Cutting a jumper enhances the level of safety.  Cutting a jumper never makes the control inoperative!  Jumper number 1 configures the PFEP (Pilot Flame Establishing Period).  Left intact, terminal10 will be powered for 10 seconds, the terminal that the ignition transformer is connected to.  If this jumper is cut, terminal 10 is powered for only 4 seconds.  Jumper number 2 configures the control to be a recycle or lockout control.  If left intact, the control will recycle on flame failure.  If cut, the control will lockout on flame failure.  Just like the RM7890, this jumper must be cut if an amplifier with 3 second flame response timing is used.  Jumper number 3 has been discussed.
The RM7895B and D have a feature the A and C series do not; an air flow switch check.  What this means is that on a call for heat or in stand-by, the control checks for a closed circuit between terminals 6 and 7.  If this circuit is closed, the RM7895B or D will lockout in 2 minutes.  Remember this: In the “old days”, to check R4795 nuisance shutdowns, we often jumped out the AFS for a while to see if bouncing contacts in the AFS were causing the problem.  Obviously, you can’t do this when dealing with the RM7895B or D.
The block diagram, Fig. 6 on page 11 of Honeywell’s form 65-0086 on the RM7895 has an error.  The “Airflow Interlock” is shown as a closed circuit.  It should be shown as an open circuit.  Another error is on the top of page 4.  For the RM7895B under “Flame Establishing Period” “main” it says “yes”.  This should be “no”.  Under “AFSC”, it says “no”.  This should be “yes”.
Attached is the Gordon Piatt diagram that shows the results of converting from the T3 or T4 timer system to the R4795
Product Categories:
Feature this resource?: 

Main menu 2