Screen Tension Tutorial

Joe Clarke has spent the past 47 years in the lab and in the engineering department, in pre-press and on-press, as an R&D / technical researcher and as a manager of screen print production. Clarke has held executive positions as President of M&R Printing Equipment and as Vice-President at Wilflex [Poly One]. He has been granted a growing number of print-related patents, including one for High-Shear printing with Smilin'Jack - he is a member of the ASDPT, is an Associate Editor for NBM and an SGIA Fellow.

Clarke has presented hundreds of technical papers, written a couple books and published over 600 technical / management articles for which he has been awarded five Swormstedts; the international standard for excellence in technical writing.

Currently Joe Clarke is the President of CPR, a Chicago-based corporation which manufactures Synergy Inks including NexGen; environmentally & financially responsible T-Shirt inks. For more information on CPR, visit

I remember when I got my first tension meter. I was darn proud to give an average of the warp and weft tensions and worse yet… it was graded in PSI not N/cm²! But before you say “we’ve come a long way my friend,” I am here to tell you; optimal mesh performance can only be achieved through the application of independent elongation of the warp and weft yarns of any given mesh… and tension meters can’t tell the difference! Unless you are an anarchist, a zealot or a heretic and like the concept of “change” you’re likely to want proof of such a controversial claim, so let’s get started with a simple test you can (safely) run at home.

To get ink to transfer, you need about 35 N/cm² of tension at contact. One can either achieve this resistance to deflection in the screen room or on press. Most tension should be developed in the screen room and the balance on the press. (All images courtesy

Testing… one… two

No really, we’re going to do a test. Take enough of your favorite fabric to make two screens (the mesh will need to be scrapped when the test is completed). I will assume it is something in a scant, vintage mono-polyester that is plain woven to roughly between 30 and 500 counts per inch in either direction? Now put it in your retensionable frame on a stretch-and-glue frame, but only tension the mesh in one direction—either north/south (N/S) or east/west (E/W). It doesn’t matter which, but mark the mesh as to its orientation (for example, warp = N/S).

Next step is to pull it a lot. Don’t worry; it is very unlikely to rip. (Ripping in the screen room occurs due to isolated stresses imposed by applied forces which are both parallel and perpendicular to each other and we’re only going to stretch in one direction.) Please note and record just how far the stretcher moved in the stretching process in terms of notches on the roller, inches on the drawbar or clamp. 

Now, pull out your trusty meter and gently center it on the screen in the N/S direction. Read tension and record it. (For the sake of this argument, we’ll say you yanked the mesh up to 20 N/cm² in the N/S direction.) Next, gently position the same meter on the same screen but in the E/W direction—or, in the direction you didn’t stretch yet. You may or may not be surprised to see you have tension in the direction you haven’t stretched yet and its tension is probably more than half of the N/S stretch. Please record and contemplate the results while we conduct the next test.

The next test

Take another piece of the same fabric and put it on a second frame in the same manner as the first test piece. Once again, we are going to stretch one direction only, but this time, we want to stretch the fabric in the direction we haven’t done yet. So, if test one was warp = N/S; stretched in the N/S direction, then test two should be tensioned east/west (weft E/W) and will be stretched in the E/W direction. Conduct the test in the same manner as before, but you will notice one of two differences: Either the fabric will stretch a lot more or a lot less in the second orientation as compared to the first. 

The reason is the difference in the construction of the fabric so that warp (N/S) and weft (E/W) build resistance to stretch at different rates. Please record the data as noted in Table 1.0.

This table exemplifies the data you will want to record for tests one and two. You will note in test one that the direction of fabric that has not been stretched still registers a level of tension albeit somewhat less than the stretched direction. 

In the bottom row of test two is Tension* the actual tension levels will be higher than test one if you oriented Warp = N/S and lower than test one if you used Weft = N/S, perhaps to accommodate a whim or the length of fabric and the shape of the frame/stretcher. 

Today’s lesson

From a glance at the results of test one, we will see the venerable meter does not gauge elongation, only the result of elongation—that is, the incremental resistance of the fabric. But the meter does not isolate directional resistance. The operative word here is fabric, not thread; to achieve printable and stable tension levels, it is necessary to equalize the construction of the fabric but not to alter the molecular composition of the yarns.

Specifically, we need to achieve a warp-weft balance of dynamic or printing tension (static tension plus the gain in tension due to off-contact) between 25 and 35 N/cm² to screen print successfully. 

From a look at the results of test two; we can observe the resistance differential (the propensity to stretch) of the warp and weft at incremental amounts of elongation. The disparity is due to the weaving process. Here, a warp yarn is wrapped around half of a weft yarn, so the woven construction leaves warp yarns with a higher amplitude (they’re wavier) while the weft yarns exhibit lesser amplitude (they’re flatter). Because of this difference in construction—some of which is mitigated by the fabric-finishing process—to produce consistent tension and to achieve stability, the warp yarns will need to be elongated farther than the weft. 

The proximity of the blade to the frame is never equal with respect to warp and weft. The printing tension is higher nearer the edges of the frame than in the center—which is why screens tend to rip at the edges, and stick in the center and not the perimeter. It is also why dot gain is highest at the edges and dot height is tallest in the center. 

Now we know the meter tells us what’s true—near average resistance to deflection—but not what’s relevant—that is, how the warp and weft contribute to the near average—and it is obvious the warp and weft need to be elongated to different lengths.

In the past, when we have unwittingly used a tension meter to guide fabric tension, the odds are that we have pulled the weft yarns too far and the warp yarns not far enough. This imbalance in the directional forces along with the meter-centric approach leads to image shift, tension loss and ripped mesh. For optimal tension, both the warp and weft threads must make equal contributions to the resistance to deflection. Meters don’t recognize that fact.

Not the same count

In almost all cases, yarns of the same composition and dimension are used to weave both warp and weft directions. The warp yarns are tied to a warp beam so, if your 305 fabric is 60" wide, some human with superb eye-to-hand coordination but far greater patience, attached more than 18,300 yarns to the beam before they began weaving. That is to say, the actual warp count is close to the published count. 

Conversely, the weft yarns are projected between two warp yarns. The warp yarns are then wrapped around the weft, which is held at both ends under tension, and the weaving process continues. Hence, the weft yarns are not often close to the published count. If you want to make long-lasting, high-quality mesh, you need the two counts to be as close as possible. 

For optimal ink transfer, the mesh periods (one thread plus one knuckle) should be square. If the fabric begins square (equal counts warp and weft) but the elongation-to-resistance or stress-strain ratio differs, the final count must be different as well. Specifically, let’s say the warp yarns require an elongation of 6 percent and the weft yarns only 4 percent; our ideal 305 mesh would have 305 warp threads and 311 weft threads.

In practice, as one direction is elongated, the count in the opposite direction is reduced. So when our “305” warp is elongated 6 percent, the weft count drops to 292. The weft is then elongated 4 percent and the warp count is reduced to 292. The final product is a square mesh opening with equal contributions to tension from both warp and weft yarns. The load is equally distributed, resulting in superior ink transfer, more stable tension and much longer screen life.

To equalize printing tension, predict the change from static to dynamic resistance and create resistance levels which result in equal resistance at all points on the screen—a process called inverse tensioning.

Author publically apologizes to meters

Just in case, let’s pull the meter back out of the case (or out of the trash). We’re going to need it. To get ink to transfer, you need about 35 N/cm² of tension at contact. One can either achieve this resistance to deflection in the screen room or on press. I have found, without exception, most tension should be developed in the screen room and the balance on the press.

Screen mesh has been perennially referred to as “uniformly elastic” but this expression is not only inaccurate, but very misleading on-press—the proximity of the blade to the frame is never equal with respect to warp and weft. 

As any point of the blade nears the frame, the resistance increases. In other words, the printing tension is higher nearer the edges of the frame than in the center. You may make note; this is why the screens tend to rip at the edges (even with rounded ends), why the screen sticks in the center and not the perimeter, and why the dot gain is highest at the edges and the dot height is tallest in the center. So, to equalize printing tension, we must compensate for this change from center to edge. The process is called inverse tensioning, wherein we predict the change from static to dynamic resistance and create resistance levels which result in equal resistance at all points on the screen.

Let’s say the off-contact gap is preset at 1/8".  Our dynamic or printing tension must be 35N/cm² and the stress-strain of the 305 mesh equates to 1N/cm²/0.015". The center static tension, then, must be a minimum of 27N/cm² to transfer ink at the center of the screen. However, at the edges of the image, where the mesh-resistance-to-deflection is higher, the same 305 may build at a rate of 2/N/cm²/0.015" (or two times as fast). This means at our 1/8" gap, the edge printing tension will be 16N/cm² greater. In order to achieve consistent 35N/cm² printing tension, the static tension at the corners of the image must not exceed 19N/cm². Our screen of 27N/cm in the center and 19N/cm² at all four corners will print at a consistent 35n/cm² at all points. 

Closing mesh remarks

For optimal screen performance and extended screen life; recognize the warp must be elongated slightly more than the weft. Stretch the mesh by its prescribed elongations in order to equalize the strain on both warp and weft directions. The ideal mesh count(s) would have a slightly higher weft count. So, once properly stretched, the openings of the mesh will be square. The percentage difference in weft-to-warp count should equal the percentage difference in warp to weft elongation.

Finally, always stretch so the glossier side of the mesh is the stencil side. Use a tension meter to control inverse tensioning so the printing tension—not the static tension—is equal at all points on the image.