You need to understand the concept of "voltage".
Voltage is a "level", or more exactly a electric potential difference.
Take two common batteries and a multimeter as an example:
0 1 2
if you measure voltage between 0 and 2 you will read 3 V, and if you measure voltage between 0 and 1 you will read 1.5 V, but if you measure between 1 and 2 you will read 1.5 V as well:
- between 2 and 0 there are 3V as (1.5+1.5)-0=3 V
- between 1 and 0 there are +1.5-0=1.5 V
- between 2 and 1 there are 3-1.5=1.5 V
If you prefer point 2 is at +3 V over ground level (i.e. relative to point 0) and at level +1.5 V relative to point 1.
The TTL levels used in this hard disk and converter use very "narrow" intervals, and signals are transmitted at a relatively high speed, so an even minimal difference in the "base level" can lead to mis-communication.
The idea is that if ALL the ground or 0 level points of ALL devices involved are connected together, they are ALL at the same level, a so-called equipotential connection or isopotential locus:
this allows the "peaks" (which are the actual 0's and 1's transmitted) to be at the correct level, because they "share a common base".
More in detail, according to specs a "peak" measured between the Tx and Ground between 0.35 V and 0.8 V is a "0", and a "peak" measured between the Tx and Ground between 2.0 V and 3.3 V is a "1".
If the TTL converter sends a "0" by sending an impulse at 0.6 V, but the ground (for whatever reason) is locally at 0.21 volts, a connected device with ground at the proper level of 0 V may "receive" it at 0.6+0.21=0.81 V and fail to "understand" it as being a "0", since these peaks are transmitted at a 38400 baud/rate there is the added complication of transition times (from 0 to the peak and back).
When doing a loopback test this is not an issue, as there is no other device connected and of course the Tx and Rx of the converter share a "common base".
If the loopback tests do not succeed, you have in your hands a *somehow* defective hardware, if you do the loopback test a the TTL level you have no way to know if the defective part is the USB to RS-232 or the RS-232 to TTL converter, so you should additionally do a loopback test on the RS-232 terminals, pinout and howto can be found here:
Be careful, the RS-232 has a relatively high voltage level, and if you misplace some connection you can "fry" something.
All in all, if you feel not familiar enough with the involved matters, are you sure you don't have a (local of course) friend with some more experience in the field?
There is nothing actually complex *anyone* with a very minimal electric/electronic experience or education can do properly the testing (and the grounding).
Edited by jaclaz, 24 June 2015 - 06:05 AM.