Zebra DS3678 Rugged Cordless Scanner: Specs and Guide

Most people ordering a rugged cordless barcode scanner treat the model number as the decision. They pick the DS3678, sign off the quantity, and assume the letters after it are packaging or trim. They are not. They are the single most consequential choice in the whole order, and the one most buyers make by accident.

The Zebra DS3678 is a handheld cordless rugged 2D area imaging scanner, supplied in four scan range variants: SR, HP, ER and DPM. Each of those is a different optical engine built for a different reading job. Ordering the wrong one does not give you a scanner that works slightly less well. It gives you a scanner that physically cannot read your codes.

SR, HP, ER and DPM are four different scanners

The temptation is to read the four variants as a ladder, with SR at the bottom and DPM at the top, and to assume that paying for the higher letters buys a scanner that does everything the lower ones do plus more. That is not how imaging engines work. Each engine has an optical design tuned to a particular combination of code size, working distance and surface type, and tuning for one of those costs you performance in the others.

SR is standard range. This is the engine for ordinary printed labels read at arm's length: carton labels, pick labels, shelf labels, shipping labels, anything printed on paper or synthetic stock and presented to the operator at normal handling distance. For the large majority of despatch, goods-in and retail back-office work, SR is the correct engine and the other three are solving problems you do not have.

HP is high performance. This engine is built for small, dense codes: high-density linear symbols with very narrow bars, and small 2D codes such as Data Matrix printed at close to the minimum practical cell size. Electronics manufacturing, medical device labelling, laboratory sample tubes and component bins are the usual homes for it. HP buys you resolution at short range, and that is the trade: an engine optimised to resolve tiny cells is not the engine you want reading a label across an aisle.

ER is extended range. This is the racking engine. It is built to read labels at distance, which in a warehouse means location labels on upper racking beams, dock door signage and pallet labels read from the floor or from a truck cab without the operator having to get closer. If your operators are climbing, leaning out of a forklift or using a ladder to get a scanner close enough to a label, ER is the answer and no amount of retraining on an SR unit will substitute for it.

DPM is direct part marking. This is a different problem entirely. A DPM code is not printed on a label. It is marked into the surface of the part itself: laser etched, chemically etched or dot peened into metal or moulded into plastic. There is no ink and often very little contrast. What the imager is reading is the difference between marked and unmarked surface under controlled illumination, which is why DPM engines use different lighting geometry rather than just a different lens. An SR, HP or ER unit pointed at a dot-peened code on a machined aluminium part will usually read nothing at all.

The part that catches people out is that this is not a setting. You cannot reconfigure an SR unit into an ER unit in software, and there is no firmware update that adds DPM capability. The engine is the hardware. If you order the wrong variant, the remedy is returning the units and ordering again, which is why it is worth spending twenty minutes on the survey below before anyone raises a purchase order. Our guides on identifying barcode scanner variants and barcode scanner depth of field and range cover the underlying optics in more detail.

How to work out which variant you need

The method is simple and almost nobody does it: measure the worst case, not the typical case. A fleet specified around the label your team scans most often will fail on the label they scan least often, and that failure will be daily, visible and expensive in lost time.

Walk the process with a ruler and a notebook and record three things.

The smallest code you have to read. Not the average. Find the tightest, densest, physically smallest symbol anywhere in the workflow, including the ones that only appear on returns, on supplier packaging or on a single product line. If that smallest code is a dense 2D symbol on a small item, you are in HP territory even if ninety per cent of your scanning is ordinary carton labels.

The furthest distance you have to read from. Again, the worst case. Stand where the operator actually stands and measure to the label they actually have to read, including the top beam of the racking and the labels on the far side of a conveyor. If any routine read is beyond comfortable arm's length, you are looking at ER.

Whether any code is marked into a surface rather than printed on a label. This is the question that overrides the other two. If any code in the process is etched, laser-marked or dot-peened, that decides the variant outright: you need DPM, because no other engine in the range will read those marks reliably. Check the parts themselves and not just the packaging, because DPM marks tend to live on components while the boxes they arrive in carry ordinary printed labels.

Where different parts of one site genuinely have different worst cases, the answer is usually two variants rather than a compromise on one. A goods-in bench on SR and a forklift fleet on ER is a sane estate. A single variant chosen to be adequate everywhere is normally a single variant that is inadequate somewhere. If you are buying across multiple sites, managing a multi-vendor hardware estate covers how to keep that from becoming unmanageable, and comparing hardware quotes line by line is worth reading before you compare two suppliers offering "a DS3678" at different numbers.

Zebra DS3678 specifications

Tell us the variant you need when you enquire, or tell us the three measurements from the survey above and we will tell you which variant those measurements point at. The base part number is the same across all four, so a purchase order that says only DS3678 is not a complete specification.

IN STOCK

DS3678 – Zebra DS3678 2D Cordless Rugged Barcode Scanner

SKU: DS3678 | 2D area imaging | Bluetooth 4.0, up to 100 m | IP67, MIL-STD-810G | 8 ft drop | 30-day warranty

View Details & Request a Quote →

What the ruggedisation numbers actually mean

Rugged is a marketing word. The numbers behind it are not, and each one describes a specific failure this scanner is built to survive.

8 ft (2.4 m) drop to concrete. Read that as a height, not as a general toughness claim. Eight feet is roughly the drop from a forklift cab or over the edge of a mezzanine onto the floor below. It is not a desk. An office scanner rated for a desk-height drop will survive being knocked off a counter for years and will break the first time it goes over a mezzanine rail. If your operators work at height, on plant or on a truck, that is the number that decides whether you are replacing scanners quarterly.

IP67. This is the sealing rating: the standardised statement of how well the enclosure keeps out dust and water. It is what lets a scanner be washed down, rained on at a dock door or used in a wet process area without the ingress that kills unsealed electronics. The rating is a property of an intact enclosure, which is why the physical condition of the seals and the battery door matters so much on any used unit, and why a cracked door quietly voids the protection you paid for.

MIL-STD-810G. This is the environmental test standard the unit has been qualified against. It is not a single test but a framework covering things like shock, vibration, temperature extremes and humidity. Its practical value to you as a buyer is that it means the environmental claims were tested to a published method rather than asserted.

Operating temperature from -30C to 50C. This is the specification that matters most and gets read least. A -30C lower limit means cold store and freezer work, and freezer work is the single most common reason a warehouse scanner dies early. An ordinary retail or office scanner taken repeatedly into a freezer does not fail from the cold itself. It fails from condensation: warm humid air enters the housing, the unit is carried into the freezer, moisture condenses and freezes inside, and the cycle repeats every shift until the internals corrode, the optics fog and the scanner stops reading. Sealing and a rated temperature range are what break that cycle. If any part of your process is a chilled or frozen environment, this specification is not optional and no amount of careful handling substitutes for it.

Cordless discipline: weight, battery and charging routine

Cordless is a genuine operational gain and it comes with two costs that need managing rather than ignoring.

The first is weight. At 0.85 lbs this scanner is noticeably heavy in the hand compared with a light retail gun, because the mass is the ruggedisation: sealed housing, over-moulding and a large battery. For an operator scanning a few hundred items across a shift with movement between them, that weight is unremarkable. For an operator scanning continuously in one position all day, it is something to check with the team before you commit to a fleet, and it is a good reason to put a lighter presentation or corded scanner on any fixed bench in the same building.

The second is the battery, and here it is worth being precise about what the numbers mean. The battery is recorded at 3,100 mAh Li-Ion with up to 71,000 scans. A scan count is a laboratory figure: a controlled duty cycle, a controlled radio load, a controlled temperature, and a fresh cell. It is useful for comparing one product against another and it is close to useless for predicting your Tuesday. Real consumption depends on how hard the illumination and decode engine work on your particular codes, how far the radio is reaching, how much idle time there is, and how cold the environment is, because cells deliver less in the cold.

The only question that matters operationally is whether the shift ends before the battery does, and the way you answer it is to run one unit for a full shift in the real environment and look at what is left. Then plan around the answer rather than the specification. That means a charging routine written into the shift handover, a base for every scanner rather than a shared one at the end of the aisle, and at least one charged spare battery per shift so a flat unit is a thirty second swap rather than an hour of downtime. Fleets that skip the spares are the fleets where someone is always scanning with a unit on its last bar. Our guide to hardware acceptance testing before production covers how to run that shift trial properly, and hardware refresh timing on risk not age is relevant once the fleet has a few years on it, because batteries age out long before the scanners do.

The base station is part of the product, not an accessory

This is the buying check that catches out more people than anything else on this page, so it is worth stating plainly: a cordless scanner without its base station is not a working scanner.

The base is recorded here with USB and RS-232 connectivity, and it is doing three jobs at once. It is the host connection, so it is the thing that actually plugs into the till, the terminal or the warehouse system and delivers the scanned data. It is the radio endpoint the handheld pairs to. And it is the charger the handheld sits in between uses. Remove it and you have a handheld that cannot talk to anything, cannot be paired and cannot be recharged.

The interface question is worth settling at the same time. USB is the default for anything modern. RS-232 exists on this base because a great deal of working equipment in warehouses, weighbridges, industrial controllers and older point-of-sale estates still speaks serial, and replacing that equipment is a far larger project than buying a scanner that can talk to it. If you are connecting into something long-lived, confirm which interface the host expects before you order, because the two behave quite differently in how data arrives at the application. Barcode scanner interfaces: keyboard wedge, USB and serial covers that decision, and barcode check digits and data validation covers what to do about mis-scans once the data is flowing.

When you are pricing a used or ex-fleet unit, ask explicitly whether the base is included and which base it is. A handheld-only listing is a legitimate thing to sell to someone replacing a broken gun on an existing base. It is not a scanner for someone starting from nothing, and the difference is easy to miss on a line item that just says DS3678.

Bluetooth range in a real building

The wireless standard here is Bluetooth 4.0, with a range recorded as up to 100 meters line-of-sight. Line-of-sight is the important qualifier and it describes a condition that essentially does not exist in a warehouse.

What actually sits between a handheld and its base in a live aisle is steel racking, stacked pallets, metal-skinned goods, refrigeration plant, roller shutters and people, all of which absorb or reflect the signal. The usable figure indoors is always shorter than the published one, sometimes dramatically so, and it varies with how full the racking is. A range test done in an empty aisle on a Sunday is a test of an empty aisle on a Sunday.

The practical approach is to test in the real place under real conditions before you commit a fleet. Take one scanner and one base, put the base where it would actually live, and walk the full route an operator walks while scanning. Find the point where reads stop arriving, and note it, because that point is the real coverage boundary and it tells you how many bases you need and where they go. For a larger deployment, wireless coverage versus capacity covers why adding more radios does not always add more usable coverage, and if the answer turns out to be that operators need to roam the whole site rather than return to a base, a mobile computer such as the Honeywell CN80 is a better shape of device than a cordless handheld.

Where this is the wrong scanner

Being straight about this saves everyone time. The DS3678 is a poor choice for a supermarket till, a pharmacy counter, a reception desk or a light office. On all of those it costs more than the alternative, weighs more in the hand, and introduces a charging obligation that a counter environment does not need and will not maintain.

A till or a counter wants a scanner that is always on, always connected and never flat. That is a corded unit, either a handheld or a presentation scanner sitting in the counter reading items passed over it. A corded scanner draws power from the host, so there is no battery, no charging routine and no dead unit at the start of a shift. For that work, look at a presentation scanner such as the Zebra DS9908 or a corded handheld such as the Honeywell Hyperion 1300g, both of which do counter work better than this scanner does and cost less doing it.

Buy the DS3678 when the environment is the problem: cold, wet, dusty, dropped, at height, or at a distance from the host. Buy something lighter and corded when the environment is a counter.

Buying it used or refurbished: what to check

Rugged cordless scanners hold up well over years of use, which is exactly why there is a healthy used market in them. The checks that matter are specific to this class of device and none of them are about cosmetics.

Is the base station present, and which one is it? Covered above, and it is the first question because a missing base is the most expensive omission on the list. Confirm the base, the interface it provides and that the host cable is included.

Are the charging contacts clean and undamaged? Look at the contacts on both the handheld and the base. Corrosion, pitting or bent contacts cause intermittent charging, which presents as a scanner that is mysteriously flat some mornings and fine on others. That symptom gets blamed on the battery for weeks before anyone looks at the contacts.

Is the battery door intact, with its seal in place? This is the one that people skip and it is the one that matters most, because the IP67 rating depends on the enclosure being closed properly. A cracked door, a missing catch or a perished or missing gasket means the unit is no longer sealed, and an unsealed rugged scanner in a wet or cold environment will fail like an office scanner while everyone assumes it is protected. Inspect it, and treat a damaged door as a fault rather than wear.

Is the scan window free of deep scratches? Light surface marking is normal and mostly harmless. Deep scratches scatter the illumination and degrade read performance, and they do it worst on exactly the hardest reads: small dense codes, low-contrast codes and DPM marks. On a DPM unit in particular, the window is part of the optical path and its condition is not cosmetic.

Also confirm the variant on the unit in front of you rather than on the listing, since all four look broadly alike from the outside and used stock gets mixed. For the wider vocabulary of the used market, see refurbished hardware grading explained and new pull versus new versus refurbished, and checking warranty and entitlement status for what cover an ex-fleet unit may or may not carry.

Ordering

We supply the DS3678 with condition confirmed before despatch and a 30-day Tech Seller USA warranty, singly or in quantity. Tell us the smallest code you scan, the furthest distance you scan from, whether any code is marked into a part rather than printed, and what the base station has to plug into, and we will confirm the variant and the interface before you order rather than after. Browse the Zebra range or the full printers and scanners catalogue. If the scanning project also involves producing the labels being read, RFID label printing and encoding and an IT asset labelling scheme that works are the companion pieces, and understanding part numbers is worth a read if you are matching this against an existing estate.

Common questions

Send us the smallest code you scan, the furthest distance you scan from, whether any codes are marked into parts, and what the base has to connect to, and we will confirm the variant before quoting.