rsync vs fs.readStream - How to handle special characters - node.js

Question

I'm using two approaches to backing up a database file, rsync and a server-based API approach.
I'm getting slightly different results because of some particular certain high-numbered unicode characters, so the two backups are just a little bit different.
The characters in question are, in one case ⸭ (a unicode 2E2D), 猄 (unicode 7304), 璣 (74a3) which makes the voyage just fine through rsync, but which all become �� (two unicode FFFD characters) using the server / API approach.
Interestingly not all higher-numbered unicode characters get transformed to FFFDs. A 䋲 (42F2), a 0698, and thousands of others that are not converted and make it through just fine.
In fact there are only about 7 characters in the entire file that get transformed in transit.
I'm trying to get this to the point where there no difference whatsoever.
Basically there's an occasional discrepancy handling high-numbered unicode.
In both cases the backup files created are utf8 with char(10) line feeds.
Here's the basic difference between the two backup approaches:
RSYNC APPROACH
rsync -avuP path/to/server/ActiveDb.sql path/to/Backup.sql
SERVER APPROACH
ThisStream=Fs.createReadStream("/path/to/ActiveDb.sql");
ThisStream.on("open", ()=>{
ThisStream.pipe(Rr.Res);
});
On the backup machine
spawn=require("child_process").spawn("curl", ["-d", "apicall=bkup", "https://dbserver.server"]);
spawn.stdout.on("data", thisChunk=>{
Fs.appendFileSync("path/to/Backup.sql", thisChunk);
});```