First time I've written a non-computer topic on this blog, but at least it'll be somewhat techical in a different field

Summary
This is an attempt to quantify the performance of various ballheads. When someone asks how a ballhead performs, people generally say “well *mine* doesn’t creep”, “my pan lock is tight *enough*”, etc. w/o saying under what conditions it was done in, how they determined it didn’t creep, etc. In an attempt to test a cross section of ballheads, I bought/borrowed 3 ballheads for this test: the latest Acratech GV2, the Kirk BH-1, and an old Arca-Swiss B1.

Figure 1: Kirk BH-1

Figure 2: Arca Swiss B1

Figure 3: Acratech GV2

Ballhead Tension/Smoothness
Ballhead tension is directly related to how much they creep, so this has to be addressed first. There are basically two types of ballheads on the market: a) ones that use a single tension knob with a minimum tension limiter and b) ones that have separate tension and lockdown knobs. I’ll call these one brake and two brake systems. The Arca Swiss B1 (and the Markins heads which are copies of the B1 with bigger/improved brakes and spherical balls) are the one brake variety; these have a single lockdown knob and there’s a minimum tension limiter that you can set so you don’t inadvertently loosen it too much so it “flops”. The two brake type basically pushes a brake against the ball from two different positions via the lockdown knob and tension knob; with these systems, you set the tension so that the ballhead doesn’t flop over but you use the lockdown knob to fully lock it down once you have composed the image. The GV2 is a two brake system because it tightens the brake on two different corners; the main negative about this design is that tension is not truly independent of lockdown (i.e., if you loosen the lockdown knob enough, the camera will flop). The BH-1 is a two brake system that tightens the brake in the middle, but on opposite sides. The RRS BH-55 is also a two brake system but it behaves more like a one brake system because the lockdown/tension knobs push on the same brake end.

To complicate things, ballheads have roughly 3 different “zones” where a tension setting works best. The zones are roughly from 0-15 degrees (zone 1: roughly straight up), 15-30 (zone 2: somewhat downward), and 30-45 (zone 3: nearly to the ball collar); these degrees are measured by looking at the ball step where 0 is straight up and flopped against the collar of the ballhead is 45 degrees (not the side notch which is 90 degrees). Most ballheads are smoothest in zone 1 because not that much brake force is needed. In zone 2, the ballheads with bigger brakes are smoother (the smaller brakes have to be set tighter). In zone 3, again, ballheads with bigger braking area are the smoothest. In general, you can’t set a ballhead’s tension so it’s smooth in all zones. E.g., if you set it so it’s smooth in zone 1, it’ll flop over if you let go in zone 2 or 3. Generally, the best thing to do is to set it so it’s smooth in zone 2 but only set it for smoothness in zone 3 if you plan to use it for a while there (e.g., for macro work).

The Arca-Swiss’ elliptical head has pros/cons (all other ballheads have perfectly spherical balls because Arca Swiss has a patent on this). It effectively feels like there’s a larger “brake” in zone 3, so you’ll almost never have camera flop because even if you leave the tension loose, the camera will slow down when it hits zone 3; it also feels smoothest of the bunch in zone 3. The negative is that this isn’t intuitive to users who may feel that it should get looser when going into zone 3 instead of tighter. Also, if you push the B1 into zone 3 after having setting tension for zone 2, the elliptical ball will “stretch” the brake so that when you bring it back up to zone 1 or 2, you’ll have to tighten then loosen the lockdown knob to get it back to your minimum tension setting; the same is true if you use the side notch.

For zone 1, all ballheads were smooth; for zone 2, the BH-1 performed best followed by the B1 and then the GV2. For zone 3, the GV2 was fairly jerky in positioning. The BH-1 was somewhat jerky but it was also difficult to find a tension position where it would prevent creep without locking up the head so it didn’t move. The B1 was least jerky in zone 3, but you had to set tension for it since this zone requires the least amount of tension because of the elliptical ball.

The BH-1 gets special mention for having the smoothest lockdown knob. The B1 gets special mention for having numbered tension settings (the other ballheads in this test do not) so that it’s easy to know what value to use for lockdown and tension.

Creep Testing
In general, creep is annoying if you do a lot of product macros or if you do architectural work where framing so particular flat features line up are important. In most other cases (e.g., landscapes, people photography), creep is not that important because you can crop the image later and most viewfinders only show you 95% of the actual image.

The goal of this creep testing is to quantify the amount of creep reliably (reproducibly). This can be done by using a 1:1 macro lens and measuring using the same camera sensor size; since Sony’s 10mpix CCD is used in so many cameras, it’s a good sensor to standardize testing with. A 1:1 macro lens will only be in focus if you’re a specific distance away from the object (thanks to William Quan for this suggestion) and the Kiron 105/2.5 lens is also available in many mounts, so that was used. In addition, the Kiron is a fairly heavy lens (23 oz or nearly 1.5 lbs), so it’s a good stress test while in the angled zones where creep is important. Because the lens is 1:1 (the image is the same size as the sensor), we can simply choose a feature in the image and compare distances at the pixel level.

Here’s an image of the test setup:

The top magnet was the straight ahead test target and the bottom magnet was the angled down test target which was nearly 45 degrees down for the lens/camera. The tripod is a Gitzo 3541LS with minimal leg extension and 2sec MLU was triggered wirelessly so the only touch was to the lockdown knob.

For the Arca Swiss B1, the optimal tension for this lens/camera combination was 6. When tightened to 9, the lockdown knob could no longer be turned. When tightened to 8, you could move the tripod by pushing the camera around.

For the BH-1/B1, there wasn’t much difference in amount of movement when the camera was positioned facing the notch versus perpendicular to it (though most people will use it perpendicular to it to be able to put the camera into portrait mode). With the GV2, there was a noticeable difference in performance when using it facing the slot versus not facing the slot, so it’s recommended that you always use the GV2 perpendicular to the slot; in addition, using tension knob to lock down the ball instead of the lockdown head produced significantly less movement but this isn’t that usable because there’s no minimum tension adjustment so it’ll flop when you loosen it.

The tension on the BH-1 was surprisingly hard to set in zone 3 to prevent creeping; it’s a fairly large head with a large brake so it shouldn’t have this issue. The lockdown was extremely smooth, but unfortunately there are no markings so I’m not sure how far the optimal lockdown point was, so it actually had more creep than the GV2 did; this could also be a factor of the lockdown knob’s slow thread so a similar amount of torque or fingertip pressure will tighten it more than you would tighten the other heads.

In the following table, “straight” means the camera was pointed straight ahead (zone 1) and “down” means the camera was pointed downwards until it was nearly touching the lip of the ball enclosure (nearly a 45 degree angle, i.e. zone 3). Three attempts were made to measure lockdown creep in each position with the camera reset and repositioned each time. The delta is the difference in pixels between the position where you let go of the camera and the position after you lock down the head until the camera can’t move. The delta includes x-axis difference and y-axis difference. A positive x-axis delta means it went right. A positive y-axis delta means it went up. E.g., 5x37 means the image moved to the right 5 pixels and up 37 pixels when the lockdown knob was tightened down. For the side notch, “facing slot” means the slot was straight ahead (the lens is facing the slot) and “not facing slot” means it was to the side (perpendicular to the lens).

Ballhead	Delta 1	Delta 2	Delta 3
Arca Swiss B1 (straight)	5x37 (tighten to 9, not facing slot)	0x1 (tighten to 8)	1x-1 (tighten to 8)
Arca Swiss B1 (down)	0x2 (tighten to 8)	0x3 (tighten to 8)	2x1 (tighten to 8)
Arcatech GV2 (straight)	43x-55 (facing slot)	68x-48 (facing slot)	64x-43(facing slot)
Arcatech GV2 (down)	49x-59 (facing slot)	59x16 (not facing slot)	61x12 (not facing slot) 32x12 (not facing slot) 14x3 (tension only)
Kirk BH-1 (straight)	50x-193 (not facing slot)	39x-133 (not facing slot)	35x-132 (not facing slot)
Kirk BH-1 (down)	18x-166 (not facing slot)	12x-146 (not facing slot)	10x-135 (not facing slot)

Kirk’s BH-1 instructions describe an unusual method of setting the tension. You tighten the lockdown knob all the way, then you tighten the tension knob, then you loosen the lockdown knob all the way, then you set tension by loosening the tension knob. I also tested the GV2 this way for completeness, but didn’t test the B1 because it did so well already. Here are the results:

Ballhead	Delta 1	Delta 2	Delta 3
Arcatech GV2 (straight, not facing slot)	-31x55	-31x15	-15x10
Arcatech GV2 (down, not facing slot)	-12x13	-6x4	-10x10
Arcatech GV2 (straight, facing slot)	-15x11	-22x19	-22x12
Acratech GV2 (down, not facing slot)	-12x8	-9x16	-12x10
Kirk BH-1 (straight, not facing slot)	20x155	0x111	-5x68
Kirk BH-1 (down, not facing slot)	-15x47	-6x20	-13x29
Kirk BH-1 (straight, facing slot)	-2x52	-1x40	-3x71
Kirk BH-1 (down, facing slot)	4x15	4x23	5x16

The different tightening technique seems to help both ballheads. For the Kirk, I was trying less pressure on the lockdown knob and that seems to help it (learning the pressure is the variation in the first 3 Kirk numbers). For the GV2, I was also trying to minimize the amount of pressure needed (it needs roughly one notch of the lockdown knob turn); unfortunately, for the GV2, if you use this technique, you increase the risk of a flop. Both are hard to do this consistently with because there is no labelling on the knobs.

Pan Lock
Some reviewers have claimed Markins heads had bad pan locks so it wasn’t possible to lock it down to take it off the tripod, so I came up with three criteria for “the pan lock sucks”:

You should be able to hand tighten it by hand with the head off the tripod so that you cannot twist the base by hand

You should be able to tighten it by hand enough so you can spin it tight enough on the tripod that the base won’t get loose when using the ballhead with it tensioned for a lens at the 30 (for the B1’s elliptical ball) or 45 degree angle; you should be able to spin the camera around without loosening the ballhead in case you want to put the camera into the side slot

You should be able to tighten it by hand to remove the ballhead for travel

All ballheads passed the pan lock tests. The GV-2 was easiest to lock down, followed by the BH-1 and then the B1 because the BH-1 and GV2 both had bigger knobs with the GV2 pan knob being the largest.

In terms of smoothness of the pan mechanism, the B1 is easily the smoothest, almost feeling like fluid dampening. The GV2 was next in smoothness and the BH-1 was last because it was tighter than it should be.

Side Notch Play
If you’re using the ballhead’s side notch to pan up/down like a gimbel head, it’s important to have minimal side to side play. The B1’s brake was showing above the bottom of the notch, so it kept the head stem from going to the bottom of the notch and there was nearly 1/4” of side-to-side play. The BH-1 had a minimal 1/8” of side-to-side play. The GV2 had no side play because of its “gimbel” feature which consists of a bushing on the bottom of the ball that falls into a notch on the bottom of the brake arm. The GV2 was the best, followed by the BH-1 and then the B1.

QR Clamps
Another ergonomic feature of each of these heads is the QR clamp. Both the GV2 and BH-1 have more modern rounded/smoothed QR Arca-Swiss style clamps with screw locks and spirit levels. The GV2 improves on this by using a double speed thread (the BH-1 also has a fast thread though it feels slower than the GV2’s but that may be because it opens wider) so it’s much quicker opening the clamp to put the camera in and it also has a detent pin to keep plates with hollowed out bottoms from slipping out if the screw lock is inadvertently loosened. The BH-1 has all the spring hardware hidden so it’s harder to get dirt caught in the mechanism.

Figure 4: Kirk Clamp

Figure 5: Arca Swiss Fliplock Clamp

Figure 6: Acratech Leveling Clamp

The B1 had the old-style (lots of sharp edges) B1 flip lock which has a lot of safety features to keep it from opening and dropping your camera. It has open, half-open, and closed positions. In the closed position, you can’t catch it on a branch and have it pop open by mistake; you need to pull a switch to the outside and then it only opens enough to slide the camera. Unfortunately, it’s missing a detent pin and retaining screw notch, so sliding is enough for the camera to fall out; the Kirk plate has a retaining screw so it can only slide in one direction (you leave this screw sitting outside the clamping area), but it’s enough for the camera to fall off if you loosen it while it’s tilted in the wrong direction. In order to open it fully, you have to use your fingernail to pull on another bar next to the lever. It’s definitely not a glove-friendly clamp, though it’s probably the safest of the lever style clamps on the market. Liao has written a good review of this clamp.

The light GV2 clamp with the fast screw lock threads is the best of the screw lock clamps tested. I wish Arca Swiss would modernize their fliplock clamp with safety notches and a detent pin, but even with that, it’s not glove friendly (both screw clamps can be easily operated with gloves on).

Ballhead Lockup
The GV2 has another feature that make it unique for hiking and dirty conditions. The ball head is out in the open so it’s easy to clean (there have never been reports of it locking up in the field while other ballheads have reportedly locked up).

The Arca Swiss B1 was infamous for locking up in the field because of a bad batch of brake pads; this has been fixed, but if you carry the camera/tripod over your shoulder, this can still happen a bit because of the elliptical head getting tighter as you bounce it.

The BH-1 should only be vulnerable to the typical dust/water getting into enclosed ballheads to cause lockup in the field.

The GV2 wins in this category. The GV2 is also easily the lightest at 15.4oz including the clamp. The BH-1 is 30oz. The B1 with fliplock is 25.4oz.

Fit and Finish
The BH-1 easily trounced the others in fit and finish. The knobs all feel highly precise and the gap tolerances were the smallest. The GV2 was 2^nd best in fit/finish because of slight markings on the finish, a badly painted spirit level circle and bad machining of the brake (though this is factory 2^nd that shipped mistakenly to me). The B1 was in worst shape because it’s only available used and the gaps varied and the one I bought was fairly beaten up.

Conclusion
Unfortunately, with all this testing, I’ve found there is no perfect ballhead; to further complicate this, most DSLRs only display 95% of the image in the viewfinder, so even without creep, you most likely need to do some cropping. Each ballhead has a set of tradeoffs depending on its design. For those of you with a priority on minimizing lockdown creep, the Arca Swiss heads should be evaluated. For hiking, the GV2 is a good compromise with the gimbel feature and ease of cleaning, despite the connected tension/lockdown controls. Prioritize your tradeoffs and learn to use your chosen tools well...