Further to my earlier post I have had problems in the past with totally spurious cards caused by blockages or leakages of the indicator cock, or leakages on the screw thread on the cock connection to the indicator machine.
If everything has been overhauled, all the fuel rack standouts on the engine are within +/- 1mm, injection timing is same as testbed, and there is no large variation in exhaust temperature, the exhaust sounds smooth when you stand by the funnel and there is no surging of the turbocharger, I don't think you have a problem with the engine. I would also check turbo RPM and charge air pressure from the Test Bed trials to see if they correspond to the fuel rack standout. though they can be hard to compare if you have a fixed pitch prop.
The problem is most likely with the cards you are taking. If you are getting good cards from the other cylinders, the indicator machine is OK and so is the method used.
The only variables left that are specific to one unit are the drive mechanism, which you say you have already overhauled and the indicator cock.
Q.E.D. as they used to say.
With regard to cracks:-
I had a cracked fuel injector on a MAN engine many years ago, this showed up as an intermittent water flow at the return hopper, there was no loss of water from the system, but when the engine fired, exhaust gas got pushed into the cooling space.
I also joined a ship with 2 medium speed MaK engines, and found that every time the load on the engine was increased to about 60% we lost cooling water pressure. We were not losing cooling water and no significant quantities of water appeared when we blew the engine over on air prior to starting.
Eventually I ran the engines steadily just below the the point at which we lost cooling pressure and manually increased the fuel rack standouts on one cylinder at a time until we lost cooling pressure or the cylinder was on full power.
Having determined which cylinders were causing the problem we changed the cylinder heads. This cured the problem. We hydraulically pressure tested the heads immediatly after they came off the engine (still hot) and found small cracks in the combustion face. However, when the heads were sent to a workshop ashore (cold) they could not find any leaks. However, after changing about 4 heads we were able to run both engines at full load. So eventually we had a success.
I suspect someone, in the past had run the engines dry and not recorded it.
It is possible that if exhaust gas is dissolving in the cooling water it may turn the water acid so chemical testing of the ph may indicate this is the problem. But I did not find the cooling water acid when I had problems.
It is always better to ask a stupid question than to do a stupid thing.