The Hatz CB-1 Biplane Page
From the collected works of Stan Hall, 1997 (Reprinted by permission from EAA EXPERIMENTER, March, 1998. Article originally printed in SOARING magazine, March 1977.)
For as long as homebuilders have been building wooden aircraft, they have been frustrated by the unavailability of a low-temperature-cure glue. As a consequence, little building is done in the colder months in the colder climates. In fact, in the northern and eastern regions of the U.S., most construction simply goes into hibernation until things warm up.
Now we may have a breakthrough. The Chem Tech Company markets a two-part epoxy called T-88 Structural Adhesive. (Editor's Note: T-88 is now marketed by System Three, Inc. of Seattle, Washington.) It is claimed that T-88 will cure at temperatures down to 350F. I have little doubt as to the validity of this claim because I recently tested some samples which cured between 40º and 55ºF. And they showed good strength.
Although temperatures commonly get much colder than 35ºF in many parts of the country, 35º is a far cry from the 70ºF-minimum-cure temperature required by all the other epoxy brands I know. This fact alone should greatly enlarge the cure temperature "window" for homebuilders living in the colder climates.
AN INVITATION TO TEST T-88 But let me tell you about my experiences with T-88 which started with Harry Conner, a BG-1 2 rebuilder from Newport News, Virginia. He wrote that the boat builders were getting pretty heavily into T-88, noted the potential for a sailplane application, and suggested that I should give it a try. One reason the boat builders like it, aside from its low-temperature cure properties, is that it will cure on wet wood -and even cure under water! Also, T-88 mixes in a 1 -to-1 ratio with its hardener. This makes it much easier to measure out than, say, the 4-to-1 or 20-to-1 ratios common to other epoxies. I filed Harry's suggestion away for future reference (I'll get to it one day, I told myself.)
But, Harry Conner isn't the type to make a suggestion and then walk away from it. I soon received a phone call from Gerald Schindler, president of Chem Tech, stating that at Harry's urging Chem Tech was sending me a quantity of T-88 for "test and evaluation." If I was agreeable, that is. Although I've always been interested in the epoxies I've used, I had come to consider one epoxy to be quite like another. Thus, my acceptance of Chem Tech's offer was not spectacularly enthusiastic. When the shipment arrived, I simply let it repose on the bench for a week, wishing it would go away. But you can't simply stare at several plastic bottles of a new and (according to the manufacturer) potentially exciting epoxy for seven days in a row without being driven ultimately to find out if the stuff is really as good as the maker claims. Not in these Ralph Nader days, you don't. So, I dug out the old hydraulic testing device I made during the Cherokee 11 development days and proceeded to build and test 20 specimens assembled with T-88.
TESTING PROCEDURES The tests were of the tensile shear type and involved cementing 2024-T3 aluminum alloy to itself, 4130 chromoly steel to itself, 2024-T3 to Douglas fir, and 4130 to Douglas fir. (Since I'm on record as promoting Douglas fir as a substitute for Spruce, I can't use Spruce anymore, even if I love it. Bad form and all that.) I didn't build any wood-to-wood samples, because if the glue will bond metal to wood, it will certainly bond wood to wood.
T-88 is quite thick and viscous, although not as thick and gummy as Aerobond. And it gets even thicker at lower temperatures -which is why Chem Tech supplies tapered nozzle on its plastic squeeze bottles. You simply a squeeze out equal-length beads of the stuff as you do with hardware-store-type epoxy tube kits. I chose to weigh the material out on a laboratory scale and assemble the parts to be bonded at 70'F. Immediately after the parts were clamped (with clothespins), I put them outside in the 50ºF nighttime air. They stayed out there for seven days, and during that period the temperature cycled between 40 and 55ºF. Then I put them into the hydraulic tester at 700ºF. The test data are shown in Table 1 (see below).
CHEMICAL AND SANDBLAST CLEANING Table 1 shows the critical importance of the cleaning procedure. It would be difficult to overemphasize the proper cleaning of the metal parts before applying the adhesive. One fingerprint, and you're out of business! Chem-Tech prefers that metal parts be degreased by a vapor degreaser and then cleaned and treated with chemicals that most homebuilders would likely consider exotic and difficult to obtain - and maybe even a bit hazardous to use.
Since I don't own a vapor degreaser (do you?), and since my local alchemist doesn't carry Concentrated Orthophosphoric Acid (one of Chem Tech's preferred chemicals), I used Chem Tech's "alternative method." This requires cleaning with lacquer thinner (wipe first with a wet cloth, then immediately with a dry one -this is vital), and scrubbing with Ajax cleanser (or some equivalent abrasive detergent), followed by washing with distilled water and oven drying. I added one step at the beginning -I disc-sanded the metal down to its bright surface. (I can't convince myself that metal is clean unless it is at least bright.) Also, I omitted the distilled water treatment and used tap water instead. As shown in Table 1, I call this Cleaning Procedure No. 1. It was not a good procedure, as subsequent testing proved.
TABLE 1 -T-88 TENSILE SHEAR TEST DATA TAKEN AT 70ºF (Cure at 40-55ºF for 7 days)
Materials Bonded Number of Tests Each Cleaning Procedure No. 1: Stress at Failure, psi Cleaning Procedure No. 2: Stress at Failure, psi 2024-T3 to 2024-T3 4130 to 4130
2024-T3 to fir
4130 to fir2
2
3
31560,1720
1720,1780
650,700,875(l)
950, 875, 700(2)1925,1925(5)
3100, 3200
1280,1280,1530(3),(4)
1450,1500,1550(3),(4)
NOTES:
(1) Joint failed in epoxy. No wood pulled.
(2)Amount of wood pulled varied from 20% to 80% of original bond area, mostly toward lower values.
(3) Amount of wood pulled close to 100% in all cases.
(4) ANC-18 (1951) shear allowable in Douglas fir, loaded parallel to grain (as in tests above) is 950 psi. Note that
Cleaning Procedure No. 2 gave much higher values. (ANC-18 data too conservative?)
(5) Chem-Tech allowable when using "alternate" cleaning methods (such as Procedure No. 2. above) is approximately 1800-2250 psi at 75ºF (no cure temperature given).
CLEANING PROCEDURES
No. 1
- (a) Disc-sand to bare metal.
- (b) Wipe clean with lacquer thinner.
- (c) Scrub with Ajax cleanser.
- (d) Wash off cleanser with tap water.
- (e) Oven-dry at 100-125º for 10 minutes.
No. 2.
- (a) Wipe clean with lacquer thinner.
- (b) Sandblast.
- (c) Scrub with Ajax cleanser.
- (d) Wash off cleanser with tap water.
- (e) Wash off tap water with distilled water.
- (f) Oven-dry at 100-125º for 10 minutes.
Chem Tech offers a couple of warnings about using the alternative cleaning method and has something to say also about using abrasive paper and paper towels for cleaning. The instructions stated that the lacquer thinner/Ajax cleanser route will yield about 60-75 percent the bond strength obtainable using their preferred method (the vapor degreaser/chemical method). Also, they advise against using acetone or methyl ethyl ketone (MEK) as cleaners; they dry so fast you don't get a chance to wipe off the residue they've collected from the surface. They note, too, that some abrasives are treated with stearates to reduce clogging. These stearates are waxy substances which can readily transfer to the metal and reduce adhesion. Chem Tech advises against using paper towels for wiping because they often contain chemicals which adversely influence adhesion. Apparently, when you hit it with lacquer thinner the chemical comes off unseen on the metal. One should use only clean rags. And stay away from commercial shop rags. They have chemicals in them, too.
Although I took note of the possibility that my abrasive paper contained stearate, I scrubbed extra diligently with Ajax cleanser to get rid of it. I don't think I was successful because, as Table 1 shows, these joints weren't as strong as those cleaned by Cleaning Procedure No. 2, which followed.
Procedure No. 2 involved sandblasting the part after first wiping clean with lacquer thinner, then scrubbing with Ajax, followed by washing with both tap water and distilled water, in that order. The joints turned out very well. In fact, they showed higher shear strengths in the wood than the wood is supposed to be capable of, according to ANC-18, "Design of Wood Aircraft Elements."
BOLT AND EPOXY COMBINED STRENGTH I performed three additional tests which, although not shown in the Table, may be of interest. It is normal practice in wooden aircraft to effect the attachment of metal fittings by means of bolts. However, I have long been interested in the strength of attachments where both bolts and epoxy are employed. The objective, of course, would be to reduce the number of bolts to save weight. It would be great if bolts could be eliminated entirely and full dependence placed on the epoxy alone. This is, in fact, a technique Don Mitchell is reported to have used successfully on his Nimbus series sailplanes in the 1950's.
Although I didn't have the fortitude to do this on the Ibex (glider), I did epoxy the main wing fittings in place. But I used bolts, too. I used a sufficient number of bolts to take the limit loads and depended on the epoxy to take the additional 50 percent required to handle the ultimate load. The tests reported here suggest that this was not too bad a move. In fact, I probably could have used fewer bolts than I did. The test series proceeded as follows:
1. The bearing strength of a quarter-inch bolt pulling parallel to the grain on a three-quarter-inch thick specimen of Douglas fir (with the bolt-end distance established in accordance with ANC- 18) was determined to be 1500 pounds. ANC- 18 gave about half that value -which is interesting in itself.
2. The shear strength of the wood itself, parallel to the grain, was determined by cementing a pair of 4130 steel fittings having a total of 1. 73 square inches glue area to the wood and applying a tensile load. Failure occurred at 2250 pounds. ANC- 18 gave 1640 pounds. Interesting again.
3. The strength of a joint combining the bolt and the epoxy (same dimensions as Tests 1 and 2) was determined to be 2800 pounds.
Intuitively, one would be tempted to assume the strength of the joint to be the sum of the allowable bolt bearing strength and the wood shear strength. If so, failure would be expected to occur at 1500 plus 2250, or 3750 pounds. Obviously, the strength was considerably less than that, by about 25 percent.
Now, this is not information you can use in design. You can't simply assume that a joint comprising both bolts and epoxy will be 25 percent less strong than the sum of their individual strengths. There is undoubtedly some relationship between bolt bearing area and glue area. (However, one cannot determine just what that relationship is on the basis of one test only; each bolt/epoxy pattern will have its own properties.) And one needs also to know which tends to fail first, the bolt in bearing or the glue (the wood) in shear. Whichever fails first, the other picks up the load. Just how much is hard to say.
About all one can say is, a joint combining bolts and epoxy will be stronger by some as yet unknown amount than bolts or epoxy alone. I say, "as yet" in the hopes that some of the professionals among our readership might like to take a whack at the problem.
THE EXPERIENCE OF OTHER HOMEBUILDERS But back to the T-88. To supplement my own testing and to further enlarge my understanding of T-88 in general, I wrote for the comments of two other T-88 using homebuilders whose opinions I respect. One was Tony Bingelis, highly regarded author of the (formerly) monthly column, "The Sportplane Builder" in the EAA's SPORTAVIATION. The other was Wayne Ison, then well-known designer and proprietor of PDQ Aircraft of Elkhart, Indiana (now TEAM, Inc. of Bradyville, Tennessee).
Tony had this to say: "I find that T-88 will make a perfect joint every time. Apparently the 50/50 mix is not critical. If the material being glued is clean wood, the glue joint will not fail, and the wood fibers will separate first. Even if you lay a glue stick on the dirty work bench, a perfect joint results and you will have to chisel it away the next day.
"Chem Tech is not pressure-sensitive, and a perfect joint is not a prerequisite as it is with other types of glue. Little or no pressure is required and this is important in a wood structure where it is difficult to apply much pressure. Often a few staples will serve to hold the material in contact, with excellent results.
"I don't know what kind of service life T-88 has, but I've used it in four of the five airplanes I have built and am building. My present project, another Emeraude, is going together with T-88.
"As you know, epoxies cannot withstand heat very well and are said to soften at temperatures above 2000F. I have no experience in this regard. However, all my airplanes are painted white.
"My garden gatepost is my test lab, where several samples of plywood and plexiglass are glued together with T-88. Four years in the sun, wind, rain, and other Texas weather extremes have apparently done no harm to the glued joints. The plywood is heavily weathered but the joints look fine. One day soon I will test them for strength.
"Do I have anything bad to say about T-88? Not really, except that the uncured T-88, like most epoxies, can cause skin irritation. Most individuals eventually become sensitized to epoxy. However, the remedy is simple; use a skin cream or wear throw-away gloves."
And here is what Wayne lson wrote: "We wanted a semi-flexible resin that would not become brittle or crack with age as many resins do. In aircraft such as those being developed at PDQ (TEAM)), there are a multitude of materials requiring bonding together and which could give problems due to various rates of expansion, shrinkage, vibration, etc. The characteristics of T-88 have given very good results in this area. As of this date no failures Of bonds on any of my projects have been detected, nor have I received any indications of this from the field.
"In the beginning I ran the usual types of tests to verify Chem Tech's claims, and found them to be as stated. We still have several test sections stored outdoors, some of which are about four years old. To date, no problems.
"T-88 also makes a good wood sealer. Simply thin the resin with a good grade of lacquer thinner to brushing consistency (or even spraying consistency) and coat the area. Although thinning the material increases the set-up and cure time considerably, this is usually not important. The sealing of wood by epoxy will not age-crack as most varnishes seem to do in time"
In summary, Chem Tech's T-88 seems like the answer to a number of important bonding problems, particularly because of its low temperature curing properties.
Editor's Note: As we prepared this material for this issue of Technical Counselor News, we checked with TEAM, Inc. President Wayne lson to verify their continued use of T-88 on their miniMAX aircraft. Ison confirmed that T-88 is still their epoxy glue of choice. lson says the only problems they have with T-88 occur when builders in northern climates "heat their shops while they're working on their project, then shut the heaters off and go in the house for the rest of the evening, causing the garage or workshop to return to near freezing temperatures. In those conditions, the glue still takes forever to set up.” TEAM Inc. sells T-88, as do a number of aircraft so suppliers such as Wicks, Aircraft Spruce and Specialty and others. TEAM's prices are $45.00 for a one-half gallon kit, $28.00 for a one quart kit, both of which include equal parts of glue and hardener.
Copyright © 1999, 2000, The Hatz Homepage, All Rights Reserved
Webpage design by Douglas MacBeth, Wee Beastie Biplane WorksComments, suggestions always welcome at HatzGuy@weebeastie.com