Machinery Safety 101

Guarding Emergency Stop Devices

This entry is part 3 of 16 in the series Emer­gency Stop

This art­icle was ori­gin­ally pub­lished in 2010 and was updated in 2015. Addi­tion­al updates are in the works and will be pub­lished dur­ing the month of June 2019.

Much con­fu­sion exists when it comes to Emer­gency Stop sys­tems, and cli­ents often ask me if it is ‘OK’ to guard emer­gency stop devices like e‑stop but­tons, foot ped­als, pull-cords, etc. Without get­ting into a ton of reg­u­lat­ory details, this art­icle will look at the require­ments in for emer­gency stop devices in three key jur­is­dic­tions: Canada, the USA and the European Uni­on.

If you need inform­a­tion on the func­tion­al aspects of emer­gency stop sys­tems, see “Emer­gency Stop – What’s so con­fus­ing about that?

Why Guard an Emergency Stop?

Gen­er­ally, emer­gency stop devices, or e‑stop devices as they’re often called, need to be pro­tec­ted from unin­ten­tion­al activ­a­tion. This prob­lem occurs because e‑stop devices have to be loc­ated close to where people work in order to be use­ful. An e‑stop you can­’t reach when you need it may as well not be there in the first place, so emer­gency stops are loc­ated at ‘nor­mal oper­at­or sta­tions’. This often means they are loc­ated under the edge of a machine table, or on an oper­at­or con­trol bar like that used on power presses, put­ting the e‑stop with­in reach, but also in the ‘line-of-fire’ when it comes to the oper­at­or’s nor­mal move­ments.

To pre­vent unin­ten­ded oper­a­tion, people often want to put rings, col­lars, or worse – cov­ers – on or around the e‑stop device to keep people from bump­ing the device. Some of these can be done and should be done, and oth­ers are nev­er per­mit­ted for good reas­on.

Regulatory Requirements

Let’s take a look at the key require­ments from the reg­u­la­tions world­wide:

  1. Emer­gency Stop devices must be clearly iden­ti­fied. The tech­nic­al stand­ards require that emer­gency stop devices be col­oured RED with a YELLOW back­ground [1].
  2. They must be loc­ated with­in easy reach of the oper­at­or. This applies to all nor­mal work­sta­tions where oper­at­ors inter­act with the machine. For main­ten­ance and ser­vice activ­it­ies where work­ers may be in loc­a­tions oth­er than nor­mal work­sta­tions, a pendant or oth­er port­able con­trol must be used to cause machine motion. This device must include an emer­gency stop con­trol along with oth­er com­ple­ment­ary safe­guard­ing devices such as enabling devices and hold-to-run con­trols. Where access is only allowed under lock­out con­di­tions, this is not required [2], [3].
  3. But­tons must be palm or mush­room-shaped devices.
  4. Devices must require manu­al reset­ting. This means that the device must latch in the oper­ated pos­i­tion and require a delib­er­ate action to reset the device. This includes actions such as: pulling put a pressed but­ton, twist­ing a but­ton to release the latched con­di­tion, press­ing a reset but­ton on a pull-cord to reset the tripped con­di­tion, etc [1].
  5. Unguarded. This means that easy access to the device may not be impeded, con­sid­er­ing the per­son­al pro­tect­ive equip­ment (PPE) that work­ers are required to wear. Devices that would be con­sidered to be guards would include:
  • Close fit­ting rings or col­lars that require a work­er to insert a fin­ger inside the ring or col­lar to reach the device and activ­ate it,
  • cov­ers that close over the device to pre­vent access,
  • a lock­ing device that pre­vents access to the device, etc.

So, con­sid­er­ing point 5 above, isn’t this the end of the dis­cus­sion? Not at all! There are a few factors to con­sider first.

One import­ant con­sid­er­a­tion is any poten­tial for acci­dent­al oper­a­tion. Depend­ing on the machine or pro­cess, unin­ten­tion­al oper­a­tion of emer­gency stop devices may res­ult in sig­ni­fic­ant lost pro­duc­tion and/or dam­age to equip­ment, includ­ing dam­age to the emer­gency stop device itself. In cases like this, it is reas­on­able to pro­tect the device from inad­vert­ent oper­a­tion as long as the meas­ures taken to pro­tect the device do not impede the oper­a­tion of the device in emer­gency con­di­tions.

ISO 13850 [4] sup­ports this idea in Clause 4.4 Emer­gency stop device:

4.4.2 An emer­gency stop device shall be loc­ated at each oper­at­or con­trol sta­tion, except where the risk assess­ment indic­ates that this is not neces­sary, as well as at oth­er loc­a­tions, as determ­ined by the risk assess­ment. It shall be posi­tioned such that it is read­ily access­ible and cap­able of non-haz­ard­ous actu­ation by the oper­at­or and oth­ers who could need to actu­ate it. Meas­ures against inad­vert­ent actu­ation should not impair its access­ib­il­ity.


Summing Up

The key dif­fer­ence between North Amer­ic­an think­ing and International/EU think­ing is in the term “unguarded” as used in the North Amer­ic­an stand­ards, versus [4, § 4.2.2], where the design­er is reminded, “Meas­ures against inad­vert­ent actu­ation should not impair its access­ib­il­ity.”

In my opin­ion, it is reas­on­able to pro­tect an emer­gency stop device from inad­vert­ent oper­a­tion by pla­cing a ring or oth­er sim­il­ar struc­ture around an emer­gency stop device as long as the struc­ture does not impair easy access to the device by the oper­at­or.

I know this opin­ion appears ini­tially to go against the estab­lished North Amer­ic­an stand­ards, how­ever it can be logic­ally argued, based on the defin­i­tion of the word “guard”.

A guard is a device that pre­vents access to some­thing, usu­ally a haz­ard. Con­sid­er­ing that we are talk­ing about a con­trol that is designed to reduce or lim­it harm, any struc­ture that does not pre­vent access to the emer­gency stop device asso­ci­ated with the struc­ture should be con­sidered to be accept­able.

That said, devices like:

  • hinged cov­ers;
  • doors;
  • lock­ing devices;
  • nar­row col­lars; and
  • any oth­er device or struc­ture

that unduly lim­its access to the emer­gency stop device can­not be con­sidered accept­able.

Effects of PPE

The phrase ‘unduly lim­its access’ has spe­cif­ic mean­ing here. If work­ers are expec­ted to be wear­ing PPE on the body part used to activ­ate the emer­gency stop device, such as gloves or boots for example, then the design of the struc­ture placed around the emer­gency stop device must take into account the added dimen­sions of the PPE, the reduc­tion in tact­ile cap­ab­il­ity that may occur (e.g. heavy work gloves make it hard to feel things eas­ily), and must com­pensate for the effects of the PPE. Big gloves/boots = Big open­ing in the struc­ture.

Light­ing and pro­tect­ive eye­wear can also play a part. You may need to use reflect­ive or lumin­es­cent paint, or illu­min­ated e‑stop devices, to high­light the loc­a­tion of the device in low light envir­on­ments or where very dark eye­wear is required, like that needed by weld­ers or used by work­ers around some infrared lasers with open beam paths.

Effects of State-of-Mind

It’s also import­ant to con­sider the likely state-of-mind of a work­er need­ing to use an emer­gency stop device. They are either urgently try­ing to stop the machine because,

  1. anoth­er safe­guard has failed and someone is involved with a haz­ard, includ­ing them­selves, or
  2. the machine is dam­aging itself or the product and they need to lim­it the dam­age.

Both scen­ari­os have a high level of urgency attached to them. The human mind tends to miss obvi­ous things includ­ing train­ing when placed under high levels of stress. Struc­tures placed around emer­gency stop devices, such as cov­ers, that com­pletely block access, even though they may be eas­ily opened, may be enough to pre­vent access in an emer­gency.

The answer you’ve been waiting for!

So in the end, can you put a struc­ture around an emer­gency stop to reduce inad­vert­ent oper­a­tion of the device:


Just make sure that you con­sider all the factors that may affect it’s use, doc­u­ment your ana­lys­is, and don’t unduly restrict access to the device.

Need more help? Feel free to email me!


IEC – Inter­na­tion­al Elec­tro­tech­nic­al Com­mis­sion

ISO – Inter­na­tion­al Organ­iz­a­tion for Stand­ard­iz­a­tion

[1]  Safety of machinery – Elec­tric­al equip­ment of machines – Part 1: Gen­er­al require­ments, IEC 60204 – 1, 2005

[2]  Con­trol of Haz­ard­ous Energy ­– Lock­out and Oth­er Meth­ods, CSA Z460, 2005.

[3]  Con­trol of Haz­ard­ous Energy – Lockout/Tagout and Altern­at­ive Meth­ods, ANSI ASSE Z244.1, 2003.

[4]  Safety of machinery — Emer­gency stop — Prin­ciples for design, ISO 13850, 2006.

Series Nav­ig­a­tionCheck­ing Emer­gency Stop Sys­temsEmer­gency Stop Cat­egor­ies

8 thoughts on “Guarding Emergency Stop Devices

  1. There was a study done in con­junc­tion with a Mas­ter­’s Thes­is by Patrica Zar­ate at Ore­gon State Uni­ver­sity in 1995 – 96 time frame that looked at a num­ber of human factors regard­ing the effects of e‑stop actu­at­or guard­ing, ori­ent­a­tion, and human response time to activ­ate e‑stops.

    I was the site Ergo­nom­ist at HP’s then largest and most com­plex site in Cor­val­lis, OR with sig­ni­fic­ant amounts of auto­mated equip­ment pro­du­cing inkjet cart­ridges. I presen­ted the idea of a Mas­ter­’s stu­dent to review some of the com­plex­it­ies and com­pli­ance issues around this top­ic to OSU’s Indus­tri­al Engin­eer­ing pro­fess­or Dr. Ken­neth Funk, PhD. He had a stu­dent that was inter­ested in this as a MS thes­is and the rest was his­tory. It was a real pleas­ure to work with Patri­cia on the study and to help nav­ig­ate what would be help­ful study outout to me as an Envir­on­ment­al Health and Safety pro­fes­sion­al, who was con­stantly bat­tling the “you can­not guard” e‑stops mind­set.

    I have a PDF copy of the study and there were a num­ber of inter­est­ing res­ults but guard­ing did not in this study present sig­ni­fic­ant degrad­a­tion of e‑stop actu­ation. By the way, the palm was very rarely used to actu­ate but thumb and first couple digits were primary modes to actu­ate. Reac­tion time was the biggest sur­prise, much longer reac­tion times than the car brake actu­ation of 3/4 of a second used in most driv­ing schools, some­thing you don’t devel­op muscle and men­tal memory for. The study, like all good stud­ies, raised a num­ber of oth­er ques­tions, trans­lated to mean there is “free project/thesis mater­i­al” for future human factors MS stu­dents! Also SEMI S2 stand­ard makes allow­ances for pro­tect­ing against acci­dent­al trip­ping of e‑stop/EMO. I prefer the terms either shroud­ing or shield­ing and there are spe­cif­ic anthro­po­met­ric attrib­utes that a shield must have in my opin­ion to ensure safe and timely actu­ation of an e‑stop device.

    1. Rob,

      Very inter­est­ing! I’m glad to hear that there is some more spe­cif­ic research that sup­ports these ideas. There are cer­tainly instances where the palm of the hand is used, but I think you are cor­rect – thumb or the first couple of digits are more com­mon. The oth­er big issue comes in envir­on­ments where bulky gloves can make access to small but­tons or closely shrouded but­tons chal­len­ging.

      If you would be will­ing to share the thes­is PDF, I would be inter­ested in hav­ing a look at the res­ults.

  2. Pingback: Mark French
  3. Pingback: Amazing Machinery
  4. Pingback: Ed Ng
  5. Pingback: MachinerySafety
  6. Pingback: MachinerySafety
  7. Pingback: MachinerySafety

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

All original content on these pages is fingerprinted and certified by Digiprove