tigger:
It seemed a strange thing to do to a reasonably well designed product. No need for your comments like that
OK, well thank you for introducing me to a fascinating topic; I have (just) now done due research on the matter.
The TM1637 datasheet recommends 10k pull-ups toVcc and 100 pF - that is picofarad capacitors to ground for RF suppression. Precisely how they determined those capacitors to be advisable will be lost in the annals of time, but the value is reasonable.
On to "Robotdyn". We must assume they are the originators of the more expensive version of the modules as these - the ones with pinout on both ends and a "pilot"(!) LED on the rear - all bear essentially identical markings except for the "Robotdyn" logo. And they have published a schematic which clearly illustrates the work of some complete moron!
OK, that may sound rude, but is I think a reasonable summary. My guess is that they did not have 100 pF capacitors in inventory and did not feel the need to do so essentially for this one design, so substituted the "nearest" value to hand - 10 nF. Actually, as they also produce actual UNO clones (as well as non-"UNO"s), they do have inventory of 22 pF which would have been a quite appropriate substitute.
In the event, the modules have been produced with a completely unworkable capacitor value - happens to be the same as one of the correct supply bypass capacitors so yes, easier to inventory and produce. And that 10 nF value has been copied into the "no name" clones which I confirmed by measuring the value of one removed (the second one - I not surprisingly lost the first when testing) by my Ardutester as "9888 pF".
tigger:
There is at least one published tutorial that advises removing the capacitors on the clock line. I am quite certain that this is an unnecessary modification.
Well, given that the capacitors are indeed 10 nF, I am quite certain that removing them is in fact, an essential repair to a faulty product. Presumably the smaller $1 Catalex/ tabao modules (with no "pilot" and only one end pinout) are likewise impaired.
I removed the capacitors easily with (my head magnifiers and) the soldering iron tip pressed against the side of the capacitor, then tapping the iron against the table to eject it. (That was not where I lost the capacitor, rather as I pressed it against the SMD test pads on the Ardutester. 
)
tigger:
What I would interested to know is if anyone has thoughts on using two modules to increase the count range to 9999,9999. I guess it's use two instances of TM1637 and a simple carry over of 1000's to 10,000's?
The problem here is not so much of wiring the two displays with the clock pin commoned and separate DIO pins, but mounting the displays adjacent.
Much more sensible to just use a MAX7219 display.
Now noted bodmer's addendum. 