My initial thoughts are the DMA you had in before has some fault.. Probably best leave the current working one in place and wait for the new 1772 to come...
If you can try both chips in a STFM then it will help test if they are faulty or not.. I guess it could be possible that without the STE buffer circuit, the DMA vs 1772 data bus conflict could possibly damage one or both of the chips.. Will be useful if you could test both chips on the working machine to verify they are not damaged..
Currently I still suspect we need the STE buffer circuit in place... But this isn't going to be so easy to set up on the remake board..
Probably the way to go would be to run HDCS via a flip-flop to delay by 1x 8Mhz cycles... I think this would pretty much do the same job as what the STE circuit does... As if HDCS is delayed, then the hard drive would not output data until later there would not be any need to isolate the databus ... At at least in theory the hard drive should not be hogging the databus when it is not selected... But I think we proved this as you type the line yesterday the floppy drive was working at that point.. So I think there is definitely a bus conflict going on..
But again, lifting the 1772 databus out of the circuit should mean the hard drive should work but not floppy drive of course...As now the hard drive is working, after changing the DMA, I can only assume that chip is damaged for whatever reason, so definitely worthwhile trying this on a working machine..
If either of the DMA or 1772 turns out to be faulty (ones you have been currently using) it is likely data conflict which is damaging them.. In which case we will pretty much know at that point what the issue is and how to solve it..
If neither of the chips are faulty, the we still have some figuring out to do