Scanner not reading some barcodes

[poot36]

I have a Aquila s500 Bematech scanner that will not read a 23 charicter long barcode.  I have attached a picture to this post of the barcode in question.  Got any ideas as to why it won't scan?

[cliffyk]

It looks like a Code 39 label, but the wide bars appear too wide vs. the narrow--especially the white bars. In Code 39 the minimum narrow bar width is 7.5 mils (.0075"), and if the bars are narrower than 20 mils the wide bars should be no wider than 2.0 to 3.2 x the narrow bar width. if narrow is > 20 mils the wide bars should be in the range of 2.0 to 3.0 x wider than the little guys.

More about Code 39 here...

The LOGMARS code used by the U.S. Department of Defense is Code 39 with better defined specifications for symbol height and bar width, and the HRI (Human Readable Interpretation) height.

Had to deal with them all the tine when i was in Logistics many years back...

[poot36]

Sorry about the low res photo, I used the webcam on all in one that I had to open up to change the focal length to get a clear photo.  It is possible that it is distorted due to that.  Is it possible that it is a custom bar-code?  It is on a Chinese android car infotainment system as a serial number if that helps with identification.

[cliffyk]

Sorry about the low res photo, I used the webcam on all in one that I had to open up to change the focal length to get a clear photo.  It is possible that it is distorted due to that.  Is it possible that it is a custom bar-code?  It is on a Chinese android car infotainment system as a serial number if that helps with identification.

If by "custom", you mean printed from software driving a graphic label printer rather than a 'real" bar code label printer with bar code fonts in EPROM; you may be correct. When we first had to use LOGMARS in the mid/late 60s I wrote LOGMARS drivers using assembly and Turbo Pascal on a Radio Shack Model III running LDOS--a CPM-like aftermarket o/s for the TRS-80 Z-80 based machines. We printed them on an Epson MX-80 dot-matrix printer...

[poot36]

I have checked that Code 39 is enabled as well as the extended ASCII set for it as well and it still wont scan.  Got any ideas what other type of barcode it could be?

[Domagoj T]

If by "custom", you mean printed from software driving a graphic label printer rather than a 'real" bar code label printer with bar code fonts in EPROM; you may be correct.

Is that really such a big deal? I've been having my own fight with a bar code system for a few years and suspected as much, but still haven't found a solution. In my case bar codes are being printed on a receipt by a thermal printer. The readings are not particularly reliable with two different readers, however, they are much more reliable when being read by an app on a mobile phone.
To be honest, I'm not exactly sure how the bar codes are generated in the software, but they do appear to have some mild, yet still visible differences when compared to the ones generated by various online generators (some lines are a bit thicker, some are thinner).

[poot36]

It would be nice to be able to scan the bar-code as I have 100+ of these things that I need to put into inventory and inputting them by hand is going to be a lot of work.  I have tried the PC webcam based applications and none of the free ones I have found yet have been able to read the bar-code.  Is there one you would recommend?  The webcam is only 640 by 480 so that may not work.

[Ranayna]

Could there be a maximum length to the scanner? I know of some Datalogic scanners where you can program in a maximum and minimum barcode length. The "datasheet" does not mention anything though.
I was not able to find a programming book for that scanner, only a short quickstart manual, so this is (educated, i hope  ) guesswork.

[CatalinaWOW]

The following link provides some clues, but doesn't give an exact match.

https://www.barcodefaq.com/barcode-match/

Looks similar to code 128

[poot36]

I have checked the default settings for the length and they fit for the bar-code that I am using.  The start and end of the bar-code do not match anything I can find online. The entire sticker is 60mm long and 15mm wide, barcode is 50mm long and 6mm high.

[nali]

It's definitely not code 39, it does look like some weird propriety encoding to me. E.g. if you look between the 3rd & 8th characters there's nothing except a bunch of single thin lines with variable spacing.

Attached is the same number in 39

[apomerleau]

I agree that it's definitely not Code 39.  From the left it looks like Code 93, but about halfway across the structure breaks down.  So through the 'D' looks like a valid Code 93, then it gets odd as some patterns thereafter are valid and some are not.  There are also not enough elements for a Code 93 symbol to represent the characters shown.  I can give this a better analysis after work in my "lab".

[apomerleau]

Here is what the data looks like as Code 93:

[CatalinaWOW]

Don't know how well you can link your inventory to a camera, but it might be easier to OCR the text than to decipher this code which is likely a Chinese local standard.

[poot36]

I have yet to find a good free webcam based OCR software, all the ones out there want a picture first.  Here is another barcode to help with identification.

[apomerleau]

Seeing a second symbol allows some additional guessing.  I agree with those who suggested this is a proprietary symbology.  My guess is that it's based on Code 93.  Both symbols follow a character structure of 3 bars and 3 spaces spanning 9 modules (plus a final bar to mark the stop character).  Both symbols begin with a Code 93 start character and follow standard encoding up to a point (happens to be the 'D' in both).  From there the next character is invalid as proper Code 93 but maybe marks the start of a numeric encoding that compacts 'n' digits (13 in these symbols) into fewer than 'n' symbols characters.  The final 4 bars and 3 spaces are a custom stop character that makes the symbology unique yet readable with simple modifications to existing decoder software.  I think the 13 digits after the 'D' might be encoded into 8 symbol characters, which would leave 2 symbol characters for check values similar to standard Code 93, but I haven't been able to work that out completely (yeah, this is a fun puzzle for me).  All guesses, but reasonable guesses based on experience.

Not sure how helpful this is to the OP.  A custom symbology probably means someone wanted to be the sole source of readers for a customer.  Without that original reader, the obvious workarounds are build your own or use a different technology to read it, like OCR as someone suggested, or manual entry if it's just a small number of them.

[CatalinaWOW]

So the simplest (though not the most elegant ) solution is to take pictures of all of your boxes.  OCR the symbols.  Print stickers in a known bar code.  Apply stickers to boxes.  Inventory as usual (and any time in the future).

[apomerleau]

Boy, I feel dumb.  These bar codes are not custom after all.  They are a symbology called "Code 93i".  This code is based on the structure of Code 93 but enhanced to offer additional features.  So to read these one needs a reader that includes 93i.

[poot36]

I have tried scanning it on a Honeywell Orbit 7120 scanner that apparently supports 93i and it did not work.  Do I have to set it up further then enabling the code 93 option?

[apomerleau]

From what I see looking at the MetroSelect Configuration Guide the 7120 supports 93 but not 93i.  I'm not sure who really uses 93i but maybe a more industrial scanner than a retail scanner is more likely to support it.  I think Intermec may have helped make 93i a reality so perhaps one of their legacy products (before being absorbed by Honeywell) may have it.