If you are looking for a good example of code that handles multiple files, see here. You can then view the source code.
This is the key source of the demo (modified just slightly):
var obj = this;
var model = (function() {
    var zipFileEntry, zipWriter, writer, creationMethod, URL = obj.webkitURL || obj.mozURL || obj.URL;
    return {
        setCreationMethod : function(method) {
            creationMethod = method;
        },
        addFiles : function addFiles(files, oninit, onadd, onprogress, onend) {
            var addIndex = 0;
            function nextFile() {
                var file = files[addIndex];
                onadd(file);
                // Modified here to use the Data64URIReader instead of BlobReader
                zipWriter.add(file.name, new zip.Data64URIReader(file.data), function() {
                    addIndex++;
                    if (addIndex < files.length)
                        nextFile();
                    else
                        onend();
                }, onprogress);
            }
            function createZipWriter() {
                zip.createWriter(writer, function(writer) {
                    zipWriter = writer;
                    oninit();
                    nextFile();
                }, onerror);
            }
            if (zipWriter)
                nextFile();
            else if (creationMethod == "Blob") {
                writer = new zip.BlobWriter();
                createZipWriter();
            } else {
                createTempFile(function(fileEntry) {
                    zipFileEntry = fileEntry;
                    writer = new zip.FileWriter(zipFileEntry);
                    createZipWriter();
                });
            }
        },
        getBlobURL : function(callback) {
            zipWriter.close(function(blob) {
                var blobURL = creationMethod == "Blob" ? URL.createObjectURL(blob) : zipFileEntry.toURL();
                callback(blobURL);
                zipWriter = null;
            });
        },
        getBlob : function(callback) {
            zipWriter.close(callback);
        }
    };
})();
Usage:
Assumes a <a id="downloadLink">Download</a> element exists to provide the download once ready.
// Prepare your images
var files = [];
for (i = 0; i < len; i++) {
    // Get the image URL from a SQLite request
    var url = results.rows.item(i).url;
    (function(url){
        var img = new Image();
        img.onload = function() {
            // Add to file array [{name, data}]
            var a = document.createElement('a');
            a.href = this.src;
            var filename= a.pathname.split('/').pop();
            console.log("Loaded file " + filename);
            files.push({name: filename, data: getBase64Image(img) });
        }
        img.src = url;
    })(url);
}
// Wait for the image to load
var check = setInterval(function(){
    if(files.length==images.length) {
        clearInterval(check);
        // Set the mode
        model.setCreationMethod("Blob");
        // Add the files to the zip
        model.addFiles(files, 
            function() {
                // Initialise Method
                console.log("Initialise");
            }, function(file) {
                // OnAdd
                console.log("Added file");
            }, function(current, total) {
                // OnProgress
                console.log("%s %s", current, total);
            }, function() {
                // OnEnd
                // The zip is ready prepare download link
                // <a id="downloadLink" href="blob:url">Download Zip</a>
                model.getBlobURL(function(url) {
                    document.getElementById("downloadLink").href = url;
                    document.getElementById("downloadLink").style.display = "block";
                    document.getElementById("downloadLink").download = "filename.zip";
                });
            });
    }
}, 500);
You can use the example source code to add in progress indicators.
Hope this helps, the nice thing about this method is the zip model is easily reusable if you make it it's own JS file.
Another thought: I presume you are using the getBase64Image function from here, if so and you still experience corruption issues, perhaps try modifying the return to simply return dataURL; and comment out the .replace(..., as the Data64URIReader may expect the prefix.