Presence Sensing Devices — Reaching over sensing fields

This entry is part 2 of 7 in the series Guards and Guard­ing

I recent­ly heard about an appli­ca­tion ques­tion relat­ed to a light cur­tain where a small gap exist­ed at the top of the sens­ing field, between the last beam in the field and the sur­round­ing struc­ture of the machine. There was some con­cern raised about the gap, and whether or not addi­tion­al guard­ing might be need­ed to close the gap. To answer this ques­tion, we need to split it into a few small­er pieces that we can deal with using the tools in the stan­dards.

The first piece to con­sid­er is the gap at the top of the sens­ing field. For this part of the analy­sis, I’m going to assume that the light cur­tain is a fixed bar­ri­er guard, and we’ll analyse the gap based on that idea.

The sec­ond piece of the puz­zle is the place­ment of the light cur­tain, and we’ll look at that sep­a­rate­ly. Once we under­stand the two pieces, we’ll put them togeth­er to see if there are any oth­er issues that may need to be addressed.

The Application

For the pur­pose of this arti­cle, I’ve sketched up the fol­low­ing fig­ures to illus­trate the ideas in the arti­cle. These draw­ings don’t rep­re­sent any actu­al robot cell or appli­ca­tion. Note that the light cur­tain in the sketch is shown with zero safe­ty dis­tance to the robot enve­lope. This is NEVER per­mit­ted.

Cell Elevation View
Fig­ure 1 — Cell Ele­va­tion View show­ing Gap above Light Cur­tain


Cell Plan View
Fig­ure 2 — Cell Plan View

Analyzing The Gap

Light cur­tains are treat­ed the same way that mov­able guards are treat­ed, so the answer to this ques­tion starts with deter­min­ing the size of the gap. I’m going to ref­er­ence two sets of stan­dards in answer­ing this ques­tion: CSA and ISO.

Safety Distances for fingers reaching through an opening
Fig­ure 3 — Fin­ger-to-Knuck­le Reach­ing through a Reg­u­lar Open­ing [1, C.4]
Z432 Reaching Through Regular Openings
Fig­ure 4 — Arm-up-to-Shoul­der Reach­ing through Reg­u­lar Open­ing [1, C.4]
Ref­er­enced Stan­dards
CSA Z432 2004 [1] ISO 13857 2008 [2]
Table 3 — Min­i­mum dis­tance from haz­ard as a func­tion of bar­ri­er open­ing size Table 4 — Reach­ing through Reg­u­lar open­ings
Open­ing Size (e) Safe­ty Dis­tance (sr) Open­ing Size (e) Safe­ty Dis­tance (sr)

11.1– 16.0mm [0.376″–0.625″]

Slot­ted >= 89.0 mm [3.5″] Square >= 66 mm [2.6″] Slot
10 < e <=12 Square/Round
10 < e <=12
>= 100 mm >= 80 mm
49.1–132.0 mm [1.876–5.000″] Slotted/Square <= 915.0 mm [36.0″] Slot/Square/Round 40 < e <= 120 mm <= 850 mm

The first thing to notice is that CSA and ISO use slight­ly dif­fer­ent open­ing sizes (e) and safe­ty dis­tances (sr). These dif­fer­ences have their ori­gin in slight­ly dif­fer­ent anthro­po­met­ric data used to devel­op the tables. In both cas­es, the max­i­mum val­ue for e defines the largest open­ing per­mit­ted with­out addi­tion­al guard­ing.

Let’s look at the appli­ca­tion to see if the gap between the top-most beam and the edge of the phys­i­cal guard falls into the bands defined for e.

Cell Elevation Close Up
Fig­ure 5 — Cell Ele­va­tion Close-Up

Based on the sketch­es of the appli­ca­tion, we have a prob­lem: The gap shown above the light cur­tain is right at the edge of the robot enve­lope, i.e., the dan­ger zone. We are going to have to either, a) Move the fence back 915 mm to get the nec­es­sary safe­ty dis­tance or, b) close the gap off com­plete­ly, either with hard guard­ing or by extend­ing the light cur­tain to close the gap.

Know­ing the size of the gap, we can now decide if the gap should be reduced, or the light cur­tain moved or enlarged. Since light cur­tains run about $125/linear inch, adding an addi­tion­al plate to reduce the size of the gap is like­ly the most cost effec­tive choice. We also need to know the dis­tance from the top-most beam of the light cur­tain to the haz­ard behind the guard. If that dis­tance is less than 915/850 mm, then we have anoth­er prob­lem, since the guard­ing is already too close to the haz­ard.

Analyzing the Light Curtain

The light cur­tain posi­tion­ing is dri­ven by the stop­ping per­for­mance of the machine. Again, let’s ref­er­ence both CSA and ISO for the rel­e­vant cal­cu­la­tions.

Ref­er­enced Stan­dards
CSA Z432 2004 ISO 13855 2005 [3]
5.1 Over­all sys­tem stop­ping per­for­mance
The over­all sys­tem stop­ping per­for­mance com­pris­es at least two phases.Thetwophasesare linked by Equa­tion (1):

T = t1 + t2                             (1)

T is the over­all sys­tem stop­ping per­for­mance;
t1 is the max­i­mum time between the occur­rence of the actu­a­tion of the safe­guard and the out­put sig­nal achiev­ing the OFF-state;
t2 is the stop­ping time, which is the max­i­mum time required to ter­mi­nate the haz­ardous machine func­tion after the out­put sig­nal from the safe­guard achieves the OFF-state. The response time of the con­trol sys­tem of the machine shall be includ­ed in t2.

t1 and t2 are influ­enced by var­i­ous fac­tors, e.g. tem­per­a­ture, switch­ing time of valves, age­ing of com­po­nents.

t1 and t2 are func­tions of the safe­guard and the machine, respec­tive­ly, and are deter­mined by design and eval­u­at­ed by mea­sure­ment. The eval­u­a­tion of these two val­ues shall include the uncer­tain­ties result­ing from the mea­sure­ments, cal­cu­la­tions and/or con­struc­tion.

Clause 10.11 — Safe­guard­ing device safe­ty dis­tanceThe­cal­cu­la­tion­formin­i­mum safe dis­tance between a safe­guard­ing device and the dan­ger zone of a machine shall be as fol­lows:

S = [K (Ts + Tc + Tr + Tbm)] + Dpf

Ds = min­i­mum safe dis­tance between the safe­guard­ing device and the haz­ard

K = speed con­stant: 1.6 m/s (63 in/s) min­i­mum, based on the move­ment being the hand/arm only and the body being sta­tion­ary.
Note: A greater val­ue may be required in spe­cif­ic appli­ca­tions and when body motion must also be con­sid­ered.
Ts = worst stop­ping time of the machine/equipment

Tc = worst stop­ping time of the con­trol sys­tem

Tr = response time of the safe­guard­ing device, includ­ing its inter­face
Note: Tr for inter­locked bar­ri­er may include a delay due to actu­a­tion. This delay may result in Tr being a deduct (neg­a­tive val­ue).

Note: Ts + Tc + Tr are usu­al­ly mea­sured by a stop-time mea­sur­ing device if unknown.

Tbm = addi­tion­al stop­ping time allowed by the brake mon­i­tor before it detects stop-time dete­ri­o­ra­tion beyond the end users’ pre­de­ter­mined lim­its. (For part rev­o­lu­tion press­es only.)

Dpf = max­i­mum trav­el towards the haz­ard with­in the pres­ence-sens­ing safe­guard­ing device’s (PSSD) field that may occur before a stop is sig­naled. Depth pen­e­tra­tion fac­tors will change depend­ing on the type of device and appli­ca­tion. See Fig­ure 5 for spe­cif­ic val­ues. (If applic­a­ble, based on the style of safe­ty device.)

Clause 6.2.3 — Elec­tro-sen­si­tive pro­tec­tive equip­ment employ­ing active opto-elec­tron­ic pro­tec­tive devices with a sen­sor detec­tion capa­bil­i­ty of  < 40 mm  in diam­e­ter Cal­cu­la­tion

The min­i­mum dis­tance, S, in mil­lime­tres, from the detec­tion zone to the haz­ard zone shall not be less than that cal­cu­lat­ed using Equa­tion (2):

S = (K x T ) + C                             (2)


K = 2 000 mm/s;

C = 8 (d — 14), but not less than 0;

d is the sen­sor detec­tion capa­bil­i­ty of the device, in mil­lime­tres (mm).

[Author’s Note — T comes from 5.1, above]


S = (2 000 x T ) + 8(d-14)               (3)

Equa­tion (3) applies to all min­i­mum dis­tances of S up to and includ­ing 500 mm. The min­i­mum val­ue of S shall be 100 mm.

Where the val­ues for S, cal­cu­lat­ed using Equa­tion (3), exceed 500 mm, Equa­tion (4) can be used. In this case, the min­i­mum val­ue of S shall be 500 mm.

S = (K x T ) + C                          (2)


K = 1 600 mm/s;

C = 8 (d - 14), but not less than 0;

d is the sen­sor detec­tion capa­bil­i­ty of the device, in mil­lime­tres (mm).


S = (1 600 x T ) + 8(d — 14)

ISO 13855 Fig. 3 a) Normal Approach
Fig­ure 6 — ISO 13855 Fig. 3 a) Nor­mal Approach


1 haz­ard zone

2 detec­tion zone

3 fixed guard

S min­i­mum dis­tance

a Direc­tion of approach

The two cal­cu­la­tion meth­ods shown above are essen­tial­ly the same, with the pri­ma­ry dif­fer­ence being the val­ue of K, the “hand-speed con­stant”. ISO uses a high­er val­ue of K for light cur­tain instal­la­tions where the field is ver­ti­cal or angled as low as 45º. If the cal­cu­lat­ed val­ue of S is >500 mm, then the val­ue of K is reduced to 1600 mm/s. Using the high­er val­ue of K for a North Amer­i­can instal­la­tion is not wrong, and will result in a more con­ser­v­a­tive instal­la­tion result. Use of 1 600 mm/s for machines going into inter­na­tion­al mar­kets is wrong if S is <500 mm when cal­cu­lat­ed using 2 000 mm/s.

Let’s assume some val­ues so we can do a rep­re­sen­ta­tive cal­cu­la­tion:

Stop­ping Time of the sys­tem (T) = 265 ms [0.265 s]

Light cur­tain res­o­lu­tion (d) = 30 mm [1.2″]

Cal­cu­lat­ing Dpf

Dpf = 8 x (d — 14) = 8 x (30 — 14) = 128

Using K = 2 000 mm/s

S = (2000 x 0.265) + 128 = 658 mm

Since appli­ca­tions where S > 500 mm can be recal­cu­lat­ed using K = 1600 mm/s

S = (1 600 x 0.265) + 128 = 552 mm

So, from the above cal­cu­la­tion, we can see that the dis­tance from the plane of the light cur­tain to the edge of the robot enve­lope (i.e., the dan­ger zone) must be at least 552 mm [21.75″]. That dis­tance is enough that some peo­ple might be able to stand between the light cur­tain field and the fix­ture in the cell, so we should prob­a­bly add a hor­i­zon­tal light cur­tain to pro­tect against that pos­si­bil­i­ty. See Fig­ure 7.

Figure 7 - Vertical Light Curtain with Horizontal segment
Fig­ure 7 — Ver­ti­cal Light Cur­tain with Hor­i­zon­tal seg­ment [1, Fig. B.15 ©]
Anoth­er alter­na­tive to adding a hor­i­zon­tal sec­tion is to slope the light cur­tain field, so that the plane of the light cur­tain is at 45 degrees above the hor­i­zon­tal, with the high­est beam as far away from the haz­ard as pos­si­ble. See Fig­ure 8.

Figure 8 - Sloped light curtain installation [1, CSA Z432 Fig B.15 (c)]
Fig­ure 8 — Sloped light cur­tain instal­la­tion [1, CSA Z432 Fig B.15 ©]
This type of instal­la­tion avoids the need to replace the exist­ing light cur­tain, as long as the field depth is enough to meet the cal­cu­lat­ed Ds.

The field could also be laid hor­i­zon­tal­ly, with no ver­ti­cal com­po­nent. This will change the Dpf cal­cu­la­tion as high­light­ed by the note in Fig­ure 8. Dpf for a hor­i­zon­tal field is cal­cu­lat­ed using the fol­low­ing equa­tion:

Dpf = 1 200 mm [48″]


S = (1 600 x 0.265) + 1200 = 1 624 mm

Note also that there is a height restric­tion placed on hor­i­zon­tal devices based on the object res­o­lu­tion as well, so the 0.3 m max­i­mum height may not apply to an exclu­sive­ly hor­i­zon­tal appli­ca­tion. Note that ISO 13855 allows H a max­i­mum val­ue of 1 000 mm, rather than cut­ting the val­ue off at 990 mm as done in CSA Z432. Using either the 14 mm or the 30 mm res­o­lu­tion cur­tains yields a min­i­mum height of 0 mm and a max­i­mum of 990 mm (CSA) or 1 000 mm (ISO). Note that the 3rd Edi­tion of CSA Z432 is like­ly to har­mo­nize these dis­tances with the ISO cal­cu­la­tions, elim­i­nat­ing these dif­fer­ences.

Also, note that field heights where H > 300 mm may require addi­tion­al safe­guards in con­junc­tion with the Pres­ence-Sens­ing Safe­guard­ing Device (PSSD) field.

Figure 8 - Calculating "H" [1, Fig. B.15 (g)]
Fig­ure 8 — Cal­cu­lat­ing “H” [1, Fig. B.15 (g)]
Going back to our orig­i­nal ver­ti­cal field instal­la­tion, there is one more option that could be con­sid­ered: Reduce the object res­o­lu­tion of the light cur­tain. If we go down to the small­est object res­o­lu­tion typ­i­cal­ly avail­able, 14 mm, the cal­cu­la­tion looks like this:

Dpf = 8 x (14–14) = 0

S = (2 000 x 0.265) + 0 = 530 mm

Since S > 500,

S = (1 600 x 0.265) + 0 = 424 mm [16.7″]

While we have sub­stan­tial­ly reduced the safe­ty dis­tance, it looks like we will still need the hor­i­zon­tal light cur­tain to ensure that no one can stand behind the cur­tain with­out being detect­ed.

If the design of the machin­ery allows, it might be pos­si­ble to reduce the stop­ping time of the machine. If you can reduce the stop­ping time, you will be able to short­en the safe­ty dis­tance required. Note that the safe­ty dis­tance can nev­er go to zero, and can nev­er be less than that deter­mined by the object res­o­lu­tion applied to the reach­ing-through tables. In this case, a 14 mm open­ing results in an 89 mm [3.5″] min­i­mum safe­ty dis­tance (CSA). Since the stop­ping time of the machine can nev­er be zero, 89 mm works out to a stop­ping time of 44.5 ms using K=2 000 mm/s, or 55.6 ms if K= 1 600 mm/s. Very few machines can stop this quick­ly.

The cal­cu­lat­ed safe­ty dis­tance is about half of the safe­ty dis­tance required for the gap, at 915 mm. Clear­ly, clos­ing the gap with the light cur­tain or hard guard­ing will be prefer­able to mov­ing the fence away from the dan­ger zone by 915 mm.

Here’s one more fig­ure to help illus­trate these ideas.

Z432 Figure B.14 a
Fig­ure 9 — CSA Z432 Fig­ure B.15 a)

Fig­ure 9 shows the dif­fer­ence between the reach­ing-through or reach­ing-over light cur­tain appli­ca­tions. Notice that with­out a restrict­ing guard above the cur­tain as we have in our exam­ple, the Dpf val­ue goes out to 1 200 mm [48″], rather than the 915 mm val­ue used in our exam­ple.

The low­er fig­ures show light fence appli­ca­tions, where two or three beams are used, rather than a full cov­er­age light cur­tain.


Here are some of the more impor­tant con­sid­er­a­tions:
1) Is the field of the light cur­tain placed cor­rect­ly, based on the stop­ping per­for­mance of the machine?
2) What is the object res­o­lu­tion of the sens­ing field? This dimen­sion may be used to assess the size of the “open­ings” in the field if this becomes rel­e­vant.
3) What is the height of the low­est and high­est beams or the edges of the sens­ing field?
4) What are the dimen­sions of the gap above the field of the cur­tain, and the dis­tance from the open­ing to the clos­est haz­ard?

ed. note: This arti­cle was reviewed and updat­ed 28-Aug-17.


I’d like to acknowl­edge my col­league, Chris­t­ian Bid­ner, who sug­gest­ed the idea for this arti­cle based on a real-world appli­ca­tion he had seen. Chris­t­ian works for OMRON/STI in their Toron­to office.


[1]     Safe­guard­ing of Machin­ery. CSA Z432. Cana­di­an Stan­dards Asso­ci­a­tion (CSA).  Toron­to. 2004.

[2]     Safe­ty of machin­ery — Safe­ty dis­tances to pre­vent haz­ard zones being reached by upper and low­er limbs. ISO 13857.International Orga­ni­za­tion for Stan­dard­iza­tion (ISO). Gene­va. 2008.

[3]     Safe­ty of machin­ery — Posi­tion­ing of safe­guards with respect to the approach speeds of parts of the human body. ISO 13855. Inter­na­tion­al Orga­ni­za­tion for Stan­dard­iza­tion (ISO). Gene­va. 2010.

Digiprove sealCopy­right secured by Digiprove © 2015–2017
Acknowl­edge­ments: Fig­ures from CSA Z432, Cal­cu­la­tions f more…
Some Rights Reserved

Interlocked gate testing

Did you know that inter­locked gates require stop­ping per­for­mance test­ing?

Machin­ery needs to be able to stop in the time it takes a per­son to open the guard and reach the haz­ard. If the dis­tance from the guard open­ing to the haz­ard is short enough that a per­son can reach the dan­ger point before the haz­ard can be con­trolled, the guard is use­less. The result­ing sit­u­a­tion may be worse

Did you know that inter­locked gates require stop­ping per­for­mance test­ing?

Machin­ery needs to be able to stop in the time it takes a per­son to open the guard and reach the haz­ard. If the dis­tance from the guard open­ing to the haz­ard is short enough that a per­son can reach the dan­ger point before the haz­ard can be con­trolled, the guard is use­less. The result­ing sit­u­a­tion may be worse than not hav­ing a guard because it’s pres­ence leads to a false sense of secu­ri­ty in users.

Test the stop­ping time of guard­ed haz­ards and make sure that guards are far enough away from the dan­ger zone to be effec­tive. For more on stop­ping per­for­mance require­ments, see CSA Z434, EN 999 (soon to be replaced by EN 13855:2010), and in the USA, 29 CFR 1910.217(h)(9)(v).

Down­load ISO Stan­dards
Down­load IEC stan­dards, Inter­na­tion­al Elec­trotech­ni­cal Com­mis­sion stan­dards.
Down­load BSI Stan­dards (British Stan­dards Insti­tu­tion)
Down­load ANSI stan­dards

Need help with stop­ping per­for­mance test­ing? Con­tact us!